- Email: [email protected]
Diagnosis of deep vein thrombosis
Progress in
Cardiovascular Diseases VOL XXXVII, NO 1
JULY/AUGUST 1994
Diagnosis of Deep Vein Thrombosis M.M.W. Koopman, E.J.R. van Beek, and J.W. ten Cate
OR SEVERAL decades, the diagnosis deep
vein thrombosis (DVT) had been based on F clinical grounds alone, because the inaccuracy of the clinical diagnosis was not well appreciated. With the introduction of contrast venography, it became apparent that clinical diagnosis was insensitive and nonspecific. Many studies have now shown that only 20% to 30% of patients with the classical signs and symptoms of DVT indeed have venous thrombosis on venography) -4 The reason for this inaccuracy of clinical diagnosis is that none of the signs and symptoms of DVT are unique for DVT, and many other disorders can mimic venous thrombosis. These disorders include muscle strain, direct twisting injury to the leg, swelling in a paralyzed leg, venous stasis, postthrombotic syndrome, lymphangitis, muscle tear, ruptured Baker's cyst, cellulitis, and orthopedic lesions of the knee and ankle joint. In a consecutive series of 160 patients with signs and symptoms of DVT, it was possible to make an alternative diagnosis in the majority of patients after ruling out venous thrombosis by venography. 5 The clinical diagnosis is also insensitive because thrombi may occasionally be nonobstructive with minimal symptoms, such as is observed in postoperative DVT. Thus, objective testing is mandatory. Moreover, treatment of all patients with clinically suspected DVT will result in unnecessary exposure of patients without the disease to anticoagulants, which are associated with the potential hazard of treatment-induced bleeding. 6-s Venography has been the reference method for the diagnosis of DVT (Fig 1). 1,9,10However, this procedure is invasive and may be associated with discomfort and side effects, such as pain, allergic reactions, nausea, and vomiting. 11 Fur-
thermore, venography has a limited feasibility, especially in severely ill patients. These disadvantages have led to the development of several noninvasive tests over the past two decades. A new diagnostic method requires extensive evaluation before it can be safely introduced into medical practice. The test should be standardized and should be compared with venography in consecutive patients to determine the diagnostic accuracy. Only a few of the available noninvasive tests have been evaluated adequately and will be discussed in more detail. The accuracy of the test depends on the nature of the patient group under consideration; therefore, it is important to distinguish three different patient categories, (1) patients with a suspected first episode of DVT, (2) patients with recurrent thrombosis, and (3) asymptomatic patients at high risk for DVT. Thrombi in these patient groups differ in their origin, localization, and extension and, therefore, influence the test outcome. Most of the thrombi in patients with a first episode of DVT have extended proximally. 2,12 These patients are at high risk for pulmonary embolism (PE). In a study of 101 patients with venography-proven DVT, 51% already had scintigraphic proof of (silent) PE. 13 Conversely, isolated calf vein thrombi appear at lower risk of embolization; however, approximately 25%
From the Centrefor Hemostasis, Thrombosis,Atherosclerosis, and Inflammation Research, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands. Address reprint requests to M.M.W. Koopman, MD, Centre for Hemostasis, Thrombosis, Atherosclerosis, and Inflammation Research, F4-133,Academic Medical Centre, Universityof Amsterdam, Amsterdam, the Netherlands. Copyright 9 1994 by W.B. Saunders Company 0033-0620/94/3701-000155.00/0
Progress in Cardiovascular Diseases, Vol XXXVII, No 1 (July/August), 1994: pp 1-12
1
2
Fig 1. Contrast venogram shows a filling defect in the femoral vein.
will extend proximally into the popliteal vein and may subsequently lead to pulmonary emboli. 14,15 Therefore, the diagnostic approach must be sensitive and specific to detect proximal DVT as well as calf vein thrombosis extending into the popliteal vein. After a first episode of DVT, approximately 10% to 20% of the patients will return with recurrent signs and symptoms.16,17 Only about 30% to 40% of these patients actually have
KOOPMAN, VAN BEEK, AND TEN CATE
recurrent DVT. Difficulties in the diagnosis of recurrent DVT exist for all available tests. Venography has its limitations because it cannot readily be repeated and because the interpretation is often difficult because of anatomical changes of the venous system, vessel damage, and resultant irregularities of the vessel wall caused by previous thrombi. Impedance plethysmography (IPG) is a reliable method in the diagnosis of recurrent DVT, as long as a normal test result was obtained after a previous thrombotic event.16 For compression ultrasonography, similar rules as for IPG apply. However, a follow-up study with this technique showed that the ultrasound investigations normalized within 1 year after a first event of DVT in only 50% of the patients. 18 The clinical utility of a newly developed compression ultrasound method to measure residual thrombus and its application for the diagnosis of recurrent DVT seems promising and is currently under investigation.19 Patients who have undergone major surgery or who were immobilized are at high risk for developing asymptomatic thrombosis. Thrombi in these patients are often small and nonocclusive.2~ However, these small thrombi may extend proximally and lead to PE. Therefore, the diagnostic approach in these patients must be able to detect these small thrombi. Furthermore, the test must be feasible in a large group of often immobilized and sometimes critically ill patients. At the moment, only venography has been extensively evaluated and proven sufficiently accurate in this patient category. The diagnosis of DVT with the use of various diagnostic techniques in acute and recurrent symptomatic patients as well as in asymptomatic high risk patients will be subsequently discussed. The emphasis will be on those tests that have undergone adequate evaluation. However, some brief remarks with regards to other, less well-investigated methods will be included. CONTRAST VENOGRAPHY
Lower extremity contrast venography is generally regarded as the reference method ("gold standard") for the diagnosis of DVT and enables the visualization of the entire deep venous system of the lower extremity and the large pelvic veins to the vena cava. 1,9,21 Contrast venography uses radio-opaque (iodinated) contrast medium, which is usually
DIAGNOSIS OF DVT
injected into a dorsal foot vein.9 Multiple injections may be required, and views of the entire deep system in at least two projections are required to establish a definite diagnosis. For this purpose, both separate films of the calf, knee, thigh, and pelvic region 9 or long-leg films may be used. 21 Venographic criteria for acute DVT have been defined previously.9,21 The most reliable criterion is the presence of an intraluminal filling defect that remains constant in shape and location in at least two different projections (Fig 1). Other, less reliable criteria are nonfilling of a segment with abrupt termination and reappearance of contrast media below and above the segment and nonfilling of the deep venous system in the trajectory of the popliteal and femoral vein, respectively. As with any diagnostic technique, contrast venography may lead to inadequate investigations or observer disagreement. Inadequate venograms may occur in as many as 21% of all examinations;5,14-16,21-23 however, this may be even more frequent in the hands of less experienced radiologists. Observer disagreement on the absence or presence of thrombi occurs in approximately 10% of adequate venograms. 21,24 A randomized study in 249 symptomatic patients with a first episode of clinically suspected DVT compared the use of separate films (Rabinov-Paulin technique) versus long-leg films (Fig 2). 2~ The inadequacy rate was 20% and 2% for the separate films and the long-leg films groups, respectively. At the same time, the observer disagreement decreased from 21% in the group with separate exposures to 4% in the long-leg film group. 2a Therefore, this latter technique may be recommended. The importance of contrast venography for the clinical management of symptomatic patients was shown by a study in 106 consecutive patients with clinical signs of DVT in whom anticoagulants were withheld after a normal venogram.5 Only 2 patients (1.3%) returned within 5 days and were shown to have developed DVT. No patients developed PE or died at follow-up for 3 months, which shows the safety of this approach. In patients with recurrent signs and symptoms of DVT, venography may be of diagnostic value, especially if a previous venogram is available for comparison. In view of the limited
3
Fig 2, Normal long-leg venogram outlines the complete deep venous system of the left leg.
diagnostic accuracy of noninvasive tests in patients with asymptomatic thrombi (ie, high-risk [post surgery] patients), venography is the diagnostic test of choice. The disadvantages of venography are related to the fact that it is invasive and requires the use of contrast media. The main adverse effects are discomfort and contrast-media-related complications, u,25 The latter are more frequently encountered with conventional hyperosmolar agents and are less frequently encountered with recently developed, low-osmolar contrast media. 11,26-29 In a study of 463 DVT patients who underwent venography with low-osmolar contrast media, the following side effects were
4
KOOPMAN, VAN BEEK, AND TEN CATE
recorded. Two patients (0.4%) developed severe adverse reactions (bronchospasm), and 103 patients (22.2%) suffered from minor side effects such as: local pain, 12.1%; nausea, 4.5%; dizziness, 1.5%; skin reactions, 1.1%; vomiting, 0.4%; edema, 1.9%; and superficial thrombopflebitis, 0.6%. 11 The frequency of postvenography DVT, as shown by 1251-fibrinogen leg scanning in patients with initial normal venograms, may be as high as 9%. 5'11'26 However, this relatively high frequency was only observed with leg scanning and could not be confirmed by repeated contrast venography. Finally, contrast venography cannot be performed in up to 10% of patients because of the inability to acquire venous access, the presence of an obvious local infection of the leg, a history of allergic reactions to contrast media, or renal insufficiency. 11,14,~5,23 Although pregnancy is a relative contraindication for performing any radiologic investigation, limited contrast venography with abdominal shielding reduces radiation to the fetus to acceptable doses. 3~ IMPEDANCE PLETHYSMOGRAPHY IPG, a noninvasive method, is based on the principle that blood volume changes of the leg lead to changes in electrical resistance (impedance). During the investigation, the patient is in the supine position with the lower limb elevated 25 ~ to 30~ and with the knee flexed 10~ to 20~ A pneumatic cuff around the thigh is inflated to a pressure of 50 cm HzO, thereby occluding the venous outflow of the leg. Circumferential calf electrodes detect the changes in impedance and are connected to a recorder. After a predetermined period, the cuff is deflated and both the total increase and decrease during cuff inflation and deflation during the first 3 seconds are plotted on a two-way IPG graph. The graph includes a discriminant line separating normal and abnormal test results. 23,32,33 The IPG test can be falsely normal if thrombi are nonocclusive or if proximal DVT is associated with well-developed collaterals. Furthermore, the IPG approach cannot distinguish between thrombotic and nonthrombotic obstruction of venous flow. Several other disorders, such as venous compression by an extravascular mass or raised central venous pressure, can produce a falsely abnormal test result. How-
ever, accuracy studies using contrast venography as the reference method have shown an overall mean sensitivity and specificity of 92% and 95%, respectively, for patients with a first episode of proximal DVT. 23,34-37The diagnostic accuracy for calf vein thrombosis is much lower, with a sensitivity of about 20%. The specificity is about 95%; this means that an abnormal test result can be used to make therapeutic decisions, but a normal test result does not rule out calf vein thrombosis and/or nonocclusive proximal thrombosis. From follow-up studies, it is known that most of the calf vein thrombi resolve spontaneously and that about 20% of these thrombi may extend proximally. To detect these extending thrombi it is necessary to repeat the noninvasive tests over a period of 1 week. Therefore, management studies have been designed using repeated (serial) IPG testing to detect these thrombi. Several studies have evaluated the use of serial IPG in patients with a first episode of signs and symptoms suggesting DVT and have showed a high positive predictive value of 83% to 95%. 14-16,38 m summary of these studies is shown in Table 1. The safety of withholding anticoagulant therapy in patients with repeated normal IPG findings was investigated in all these studies. However, over the years, duration and number of IPG assesments at follow-up has steadily decreased without compromising the safety. Initially, IPG was performed at days 1, 2, 5, 7, 10, and 14.15 With this regimen, 6 of 311 patients (1.9%) with normal repeated IPG resuits returned with documented DVT within 1 year, and no fatalities occured. 15 A subsequent study omitted the 14-day IPG test, and only i of 289 patients (0.3%) with normal IPG findings returned with PE during a 6-month follow-up Table 1. Summary of Various Studies Evaluating Serial IPG
Reference
Positive Predictive Value (n/n)
Hull et al, 1985 TM Huisman et al, 1986 TM Huisman et al, 1989 TM Heyboer et al, 19933s
90/95 113/123 38/42 89/107
(95) (92) (90) (83)
Total
330/367 (90)*
RecurrentVTE (n/n) in Normal Serial IPG 6/311 1/289 1/131 9/369
(1.9) (0.3) (0.8) (2.5)
17/1100 (1.5)1
The values shown in parentheses are percentages. *95% CI, 86% to 94%. 1"95% CI, 0.9% to 2.1%.
DIAGNOSIS OF DVT
period. 14 The next two studies performed IPG at days 1, 2, and 7 only. 16'38 In a community hospital study, 1 of 131 patients (0.8%) with repeated normal IPG returned with DVT, 16 whereas in a similar study in symptomatic outpatients, 9 of 369 patients (2.5%) with normal serial IPG results returned with objectively documented venous thromboembolism (VTE).38 Despite these satisfactory results, two recent studies reported a much lower reliability of IPG. 39,4~ In one study, a recently developed, computerized version of IPG was used. 39 In 311 patients with a normal IPG test result, no anticoagulant therapy was administered. During follow-up, 10 patients (3.2%) developed VTE, including 4 fatal episodes of PE (1.3%) at day 3, 5, and 8 and at 5 months. 39 The other (retrospective cohort) study reevaluated the accuracy of a newer conventional, noncomputerized, IPG machine for detection of DVT. In this investigation, IPG showed a sensitivity of only 66% for proximal DVT. 4~Possible explanations for this discrepancy with previous studies is the use of newer, less sensitive equipment and an increased awareness of DVT among referring physicians, resulting in less extensive thrombi at onset. These results emphasize the great importance of proper evaluation of diagnostic tests and equipment before they can be introduced safely into medical practice. Although less extensively evaluated, IPG seems to be a safe and effective method to detect DVT in pregnant woman. 41 In a study in 139 pregnant women in whom anticoagulants were withheld on the basis of normal serial IPG findings, no thromboembolism (95% confidence interval [CI], 0% to 2.6%) was encountered during long-term follow-up.41 In patients with suspected recurrent DVT, IPG can only be used if a normal baseline test result is available. Prospective cohort studies have shown that, 12 months after the acute thrombotic episode, 95% of the IPG tests return to normal. 42 If the IPG test has returned to normal, it can be used alone or in combination with fibrinogen leg scanning to diagnose recurrent DVT in symptomatic patients. 17,42 In one study, 270 patients were referred with clinical signs and symptoms of recurrent DVT. IPG was negative in 200 of these patients. In all IPGnegative patients, leg scanning was performed, and normal test results were obtained in 181 of
5
these 200 IPG-negative patients, and anticoagulant therapy was withheld. During 3 months of follow-up, 3 of 181 patients (1.7%) were shown to have recurrent DVT, despite normal IPG and legscan test results. In another study of patients with suspected recurrent DVT, anticoagulants were withheld safely in all 18 patients with repeated normal IPG tests. IPG is not useful in the management of asymptomatic postoperative patients because of its low overall sensitivity of 22%. The additional use of 125I-fibrinogen leg scanning improves sensitivity to approximately only 50%. 43-47 ULTRASONOGRAPHY
Real- Time B-Mode Ultrasonography Real-time B-mode ultrasonography (US) is a noninvasive method that allows direct visualization of the deep veins of the leg. During the examination, the patient is in the supine position. The ultrasound probe is placed in the groin to identify the common femoral vein (medial to the common femoral artery). The whole venous trajectory is then visualized throughout its course until the superficial femoral vein is lost in the adductor canal. For examination of the popliteal vein, the patient is in the prone or lateral position, with flexed knees. The visualization of the calf veins is more difficult because of the variable anatomical positions of the veins and the small size of the vessels. Using a 5- to 10-mHz high-resolution transducer, it is not only possible to localize deep veins but also to visualize the thrombus in the vessel, one of the suggested diagnostic criteria for DVT. Further diagnostic criteria comprise noncompressibility of the vein and the absence of vein extension upon a Valsalva manoevre. However, a large prospective study showed that the single criterion of noncompressibility of the vein on gentle pressure with the probe is the most accurate for the detection of DVT (Fig 3a and b). 48 Most of the compression ultrasonography (CUS) accuracy studies have limited the examination to the common femoral vein in the groin and the vein in the popliteal fossa. The overall mean sensitivity and specificity of CUS is high, 97% (range, 83% to 100%) and 97% (range, 86 to 100%), respectively.48-55Using this two-point CUS, isolated thrombi in the superficial femoral vein and in isolated calf veins are missed.
6
KOOPMAN, VAN BEEK, AND TEN CATE
Fig 3. (A) B-mode US shows the popliteal vein (PV} and artery (PA) without compression by the probe. (B) Normal compression B*mode US leaves the popliteal artery (PA) open but completely occludes the popliteal vein.
However, a recent venogram study showed the relative low frequencies of these localizations (ie, none of the patients [0%; 95% CI, 0% to 1.9%] had an isolated, superficial femoral thrombus, and isolated calf vein thrombosis was detected in only 12% of the patients. ~-2Therefore, an abnormal two-point CUS result can be used to make therapeutic decisions; however, a norreal test result does not rule out calf vein thrombosis. Although most of the calf vein thrombi resolve spontaneously, approximately 20% extend proximally and may lead to PE. Therefore, CUS should be repeated over a 1-week period to detect these extending thrombi, which is similar to the approach used for I P G ? s Falsely normal test results can be obtained in patients with duplication of the popliteal vein and thrombus in one of these popiteal venous systems. Alternatively, a dilated collateral vein may be misinterpreted as a normal, open venous system in the presence of an acute, completely occluding thrombus in the popliteal or femoral vein. Compression of the vein by extravascular masses, such as a tumor or a gravid uterus, may cause noncompressibility of the vein at a lower level as a result of increased venous pressure and may be the reason for a falsely abnormal test result. In patients with a previous history of DVT, residual thrombus mass in the vein can mimick acute thrombosis. In a recent study, the safety of withholding
anticoagulant therapy in symptomatic patients with serial normal CUS and serial normal IPG results was compared? 8 From a total of 985 consecutive symptomatic outpatients, 491 were randomized to undergo serial CUS. Of this group, 89 patients with an abnormal test result underwent venography that confirmed D V T in 84 patients (positive predictive value, 94%). No anticoagulant therapy was administered to the patients with a repeated normal CUS test result. During a 6-month follow-up, 6 of those patients (1.5%; 95% CI, 0.5% to 3.3%) had recurrent V T E that was confirmed by venogram. 38 Hence, the use of serial CUS appears to be a safe method in the management of patients with signs and symptoms of DVT. The usefulness of B-mode CUS in patients with recurrent thrombosis seems limited and depends on the rate of normalization of the test during follow-up. 18,56 In a prospective cohort study, the CUS results became normal in only 50% of the patients after 1 year. is A newly developed quantitative method, which is based on measuring thrombus size, is now under investigation. 19 In asymptomatic patients, CUS seems more accurate than IPG; nevertheless, the overall mean sensitivity is unfortunately as low as 59% (range, 43% to 100%), with an overall mean specificity of 98% (range, 91% to 95%) for proximal D V T . 57-60
DIAGNOSIS OF DVT
In conclusion, CUS is useful in symptomatic patients with a first episode of acute DVT but, as yet, can not be recommended for the diagnosis of recurrent DVT. Further studies are required to determine the diagnostic accuracy in this condition. For asymptomatic DVT, CUS (ie, as a screening method) is not sensitive enough to be used in the routine management of high-risk patients.
Doppler Ultrasonography Doppler US is based on the principle that an ultrasound beam is reflected by blood cells in a vessel. If the blood is stationary, no sound is recorded. In the case of moving blood cells, a difference in frequency between the initial and reflecting beam can be converted into a flow sound?'1 During the examination, the patient is in the supine position, avoiding tight clothes that can interfere with venous flow. The common femoral artery, located in the groin, can be identified by moving the probe in a medial direction. For examination of the popliteal vein, the patient is in the prone position with the knee flexed and the feet rested on a pillow. The popliteal vein is located lateral to the popliteal artery. The venous flow sound is typically lowpitched, changes with respiration, and can be augmented by squeezing the thigh and calf. Obstruction to venous outflow may result in loss of phasicity of the venous signal and loss of the relationship with respiratory fluctuation. Furthermore, augmentation of the venous flow by squeezing the thigh and calf may be diminished o r a b s e n t . 61
Several investigators have compared the accuracy of Doppler US with contrast venography in patients with symptoms of DVT. 62-67The accuracy of the Doppler US depends on the experience of the investigator and can be disturbed by subjective interpretation of the test result. Therefore, the sensitivity and specificity for proximal DVT ranges from 65% to 100% and 72% to 98%, respectively.62-67The introduction of a standardized protocol and the use of strip-chart ultrasound recordings have improved the accuracy of the test. 6s Using the objective criteria indicative for proximal DVT (ie, a continuous or absent venous Doppler signal), the sensitivity and specificity can be increased to 91% and 99%, respectively.6s In a recently published study, real-time CUS
7
and Doppler US were compared with venography in a consecutive series of 161 outpatients referred for clinically suspected DVT. 69 For proximal DVT, the sensitivity for CUS was 100% and for Doppler US was 95% (P = .056). However, for all thrombi (including isolated calf vein thrombi) the sensitivity of CUS was significantly better than that of Doppler US (95% and 76%, respectively; P < .04). The specificity of both tests was not significantly different (100% for compression US and 98% for Doppler U S ) , 69
In conclusion, this study showed the superiority of real-time CUS, when compared with Doppler US, for detection of all venous thrombi and equal accuracy in detecting proximal DVT. However, prospective studies on the safety of withholding anticoagulant therapy in patients with normal Doppler US results are lacking. The value of Doppler US in patients with recurrent DVT and in postoperative asymptomatic patients at high risk have not been evaluated.
ColorDoppler Ultrasonography The addition of color Doppler utilities to the conventional US equipment allows direct visualization of the flow within vessels, which might imply a more accurate and easier performance of the test. 7~However, the sensitivity and specificity (both 97%) of color Doppler US for proximal DVT in symptomatic patients was similar to that of C U S . 71-74 The potential advantage of color Doppler for the detection of thrombi in asymptomatic postoperative patients, which are often small and nonocclusive, was not confirmed in clinical studies, which showed a mean sensitivity of only 43% with a specificity of 9 7 % . 75-77 In conclusion, the extra costs that are required to perform color Doppler US do not result in improved diagnostic accuracy compared with CUS. Additional comparative studies are currently in progress. SCINTIGRAPHIC METHODS
125I-FibrinogenLeg Scanning Fibrin makes up an essential part of thrombi, and radiolabeled fibrinogen, after cleavage by thrombin, becomes incorporated in thrombi. Subsequently, the radioactivity may be detected from the surface of the limb. The method was introduced over 30 years ago 78-8~and was subse-
8
KOOPMAN, VAN BEEK, AND TEN CATE
quently used in the diagnosis of asymptomatic DVT in high-risk patients as an investigative tool to evaluate the effectiveness of thromboprophylaxis. 81-84 Although initial studies showed very high sensitivity for distal thrombi, this was later disputed by several other studies. 43,44,85In a recent review of the literature, accuracy studies and positive predictive value studies were assessed to address this apparent change in sensitivity of the 125I-fibrinogen-uptake test in later years. 86 Bias in the earlier studies was identified as the most probable cause of the apparent decrease in sensitivity. 86 In addition, the sensitivity may also be related to the size of the thrombus and may be as low as 20% for small calf vein thrombi. 43 The virus safety of the product is another point of concern. Fibrinogen is derived from human plasma and can not be pasteurized. Thus, it is possible for blood-borne virus transmission to take place. This fear was the main reason for the product's removal from the market. At present, new attempts to produce a virus safe product are being made. In conclusion, 125I-fibrinogen leg scanning does not appear to be useful for the diagnosis of DVT at present. If the test were to become available again, new studies may be required to define its role.
Other Scintigraphic Methods Some techniques, such as 99mTc-labeled macroaggregated albumin or 99mTc-labeled erythrocytes, are based on the fact that perfusion is absent in the presence of obstructing thrombi. 87,88 Other techniques are based on incorporation of 111In-labeled platelets into the thrombus. 89 However, these techniques appeared to be inaccurate and were never introduced in clinical practice. BLOOD TESTS
For several decades, investigators have attempted to develop a blood test that could adequately diagnose DVT. Fibrinopeptide A, fibrin(ogen) degradation products (FDP), degradation products of cross-linked fibrin (Ddimer) prothrombin fragments 1 + 2 (F1 + 2), and thrombin-antithrombin III (TAT) complexes have been investigated. 9~ For DVT to be adequately excluded by a blood test, the sensitivity (and, thus, the nega-
tive predictive value) would need to approximate 100%. Using this principle, a cut-off value for D-dimer would consist of the lowest Ddimer value detected in patients with DVT. This would ensure that no patients with DVT could remain unnoticed, and this would allow safe exclusion to take place. Several investigators have shown that plasma D-dimer may be able to adequately identify those patients without DVT (Table 2). For example, an enzyme immunoassay (EIA) for D-dimer reached a sensitivity of 100% in patients with suspected DVT; however, the specificity of the test was much lower in inpatients (4%) than in outpatients (57%). 93 In another study, the value of latex and EIA D-dimer assays were determined in 62 patients with suspected DVT, using venography as a reference method. 94 For latex D-dimer, an inferior sensitivity of 73% was found, whereas EIA D-dimer reached a sensitivity of 100%. This was confirmed by other investigators in 56 consecutive outpatients and inpatients with suspected DVT. 96 Of the tested D-dimer enzyme-linked immunosorbant assay (ELISA), two D-dimer latex assays, and a latex FDP test, only D-dimer ELISA showed a sensitivity of 100%. In a series of 116 consecutive outpatients with clinically suspected DVT, the diagnostic value of F1 + 2, TAT complexes, and D-dimer were compared. 98Only D-dimer ELISA showed sufficient sensitivity and specificity to be of potential clinical use. In a large study in outpatients with suspected DVT, the value of D-dimer was evaluated as a replacement of serial noninvasive tests for DVT (IPG or C U S ) . 99 In this study, D-dimer was performed in addition to noninvasive tests for DVT on day 1, and the results were compared with the findings of serial IPG or CUS outcome. After determination of a cut-off value to obtain a sensitivity of 100% when compared with Table 2. Summary of Studies in the Evaluation of the Value of Plasma D-Dimer in the Diagnosis of DVT Reference Rowbotham et al, 198793 Heaton et al, 19879== Bounarneaux et al, 198996 Speiser et al, 199097 Boneu et al, 199198 Heyboer et al, 199299
No. of Cut Off Sensitivity Specificity Patients (p,g/L) (%) (%) 104 62 53 97 116 309
250 200 200 120 500 300
100 100 100 100 94 100
36 47 31 48 51 29
DIAGNOSIS OF DVT
9
contrast venography, the assay was prospectively validated. In 309 consecutive patients, a sensitivity of 100% was obtained, but the specificity was relatively low (28%; 95% CI, 23% to 34%). A subsequently performed cost-effectiveness analysis did not show decreasing costs per diagnosed DVT when D-dimer was used in addition to either IPG or CUS. However, 10 of 70 patients with proven DVT were identified earlier using D-dimer, which could reduce the number of patients that would be required to return to the hospital for serial tests to be performed. In conclusion, D-dimer may be of potential clinical use in the exclusion of DVT. Unfortunately, management studies have yet to be performed. Furthermore, the reliability of fast D-dimer tests remains to be proven. Therefore, at present, D-dimer can not be recommended for the routine management of patients with clinically suspected DVT. OTHER NONINVASlVE SCREENING METHODS
Liquid crystal (contact) thermography is based on the fact that the temperature of a thrombotic leg will increase as compared with that of the unaffected limb. Several reports have shown a high sensitivity, which would make this test clinically useful in the exclusion of DVT. 1~176176 However, a subsequent study in 185 patients could not show a sufficiently high, negative predictive value for it to be useful as a screening test.102 Light reflection rheography uses the observation that exercise results in a decrease of blood in the venous plexus of the calf. Light that is absorbed by red blood cells will be increasingly reflected with the decreasing blood pool, whereas this returns to normal during rest. In DVT, the blood pool remains in the calf, which yields a different pattern of light reflection. This test was shown to be sensitive, but thorough evaluation has yet to take place2 ~ SUMMARY
In this chapter, various tests have been discussed in the diagnosis of DVT and have been
classified according to various patient categories. To summarize, the following guidelines may be of clinical use in the management of patients with suspected DVT.
Acute, First Event of Suspected DVT These patients often suffer from occluding, proximal thrombi. Therefore, noninvasive tests such as CUS or IPG are most suitable for these patients. If an abnormal CUS or IPG result is found, the diagnosis is virtually proven, and this may serve as a basis to treat the patient with anticoagulants. If a normal CUS or IPG result is obtained, serial testing is indicated to detect extending calf vein thrombi or nonoccluding DVT, which becomes occlusive at follow up. Anticoagulants may be withheld safely if the test remains normal within 1 week.
Acute, Recurrent Suspected DVT These patients may have residual thrombi present, which makes the noninvasive tests (CUS/IPG) less useful. However, if a normal noninvasive test was documented previous to the acute recurrent event, this test may be used. If an abnormal test result is found in the presence of a documented, normal previoustest outcome, this may serve as a basis for anticoagulant therapy. Although no formal studies have been performed to evaluate the safety of withholding anticoagulants if a normal CUS or IPG result is obtained, serial testing is likely to be adequate in these circumstances. Phlebography is the only truly evaluated approach, and this could be considered in all suspected recurrent DVT. Furthermore, contrast venography is the test of choice to discern acute from old thrombi.
Asymptomatic D VT in High-Risk Patients The majority of these thrombi are mostly localized in the calf veins only and are often nonocclusive. This makes noninvasive tests unreliable for their detection. Therefore, only contrast venography should be used in this patient category.
REFERENCES
1. Bauer G: A venographic study of thromboembolic problems. Acta ChirScandSupp161:1-75, 1940 2. HaegerK: Problemsof acute deepvenousthrombosis.
I. The interpretations of signs and symptoms.Angiology 20:219-223, 1969 3. CranleyJJ, CanosAJ, SullWJ: The diagnosisof deep
10
venous thrombosis. Fallibility of clinical symptoms and signs. Arch Surg 111:34-36, 1976 4. O'Donnel TF, Abbott WM, Athanasoulis CA, et al: Diagnosis of deep vein thrombosis in the outpatient by venography. Surg Gynecol Obstet 150:69-74, 1980 5. Hull RD, Hirsh J, Sackett DL, et al: Clinical validity of a negative venogram in patients with clinically suspected venous thrombosis. Circulation 64:622-625, 1981 6. Oudkerk M, van Beek EJR, van Putten WLJ, et al: Cost-effectiveness analysis of various strategies in the diagnostic management of pulmonary embolism. Arch Intern Med 153:947-954, 1993 7. Levine MN, Hirsh J, Landefeld S, et al: Hemorrhagic complications of anticoagulant treatment. Chest 102:352S363S, 1992 (suppl) 8. Van der Meer FJM, Rosendaal FR, Vandenbroucke JP, et al: Bleeding complications in oral anticoagulant therapy. An analysis of risk factors. Arch Intern Med 153:1557-1562, 1993 9. Rabinov K, Paulin S: Roentgen diagnosis of venous thrombosis in the leg. Arch Surg 104:134-144, 1972 10. Lensing AWA, BOiler HR, Prandoni P, et al: Contrast venography, the gold standard for the diagnosis of deep vein thrombosis: Improvement in observer agreement. Thromb Haemost 67:8-12, 1992 11. Lensing AWA, Prandoni P, BOller HR, et al: Lower extremity venography with iohexol: Results and complications. Radiology 177:503-505, 1990 12. Hull RD, Secker-Walker RH, Hirsh J: Diagnosis of deep vein thrombosis, in Colman RW, Hirsh J, Marder V J, Salzman EW (eds): Hemostasis and Thrombosis. Philadelphia, PA, Lippincott, 1987, pp 1220-1239 13. Huisman MV, BOller HR, ten Cate JW, et al: Unexpected high prevalence of silent pulmonary embolism in patients with deep venous thrombosis. Chest 95:498-502, 1989 14. Huisman MV, BOller HR, ten Cate JW, et al: Serial impedance plethysmography for suspected deep venous thrombosis in outpatients. The Amsterdam General Practitioner Study. N Engl J Med 314:823-828, 1986 15. Hull RD, Hirsh J, Carter C J, et al: Diagnostic efficacy of impedance plethysmography for clinically suspected deepvein thrombosis: a randomized trial. Ann Intern Med 102:21-26, 1985 16. Huisman MV, BOller HR, ten Cate JW, et al: Management of clinically suspected acute venous thrombosis in outpatients with serial impedance plethysmography in a community hospital setting. Arch Intern Med 149:5ll-513, 1989 17. Hull RD, Carter CJ, Jay RM: The diagnosis of acute recurrent deep-vein thrombosis: A diagnostic challenge. Circulation 67:901-906, 1983 18. Heyboer H, Jongbloets LMM, BOiler HR, et al: The clinical utility of real-time compression ultrasonography in the diagnostic management of patients with recurrent venous thrombosis. Acta Radiol Scand 33:297-300, 1992 19. Koopman MMW, Jongbloets LMM, Lensing AWA, et al: Clinical utility of a quantitative B-mode ultrasonography method in patients with suspected recurrent deep vein thrombosis (DVT). Thromb Haemost 69:623, 1993 (abstr) 20. Salzman EW, Hirsh J: Prevention of venous thrombo-
KOOPMAN, VAN BEEK, AND TEN CATE
embolism, in Colman RW, Hirsh J, Marder VJ, Salzman EW (editors): Hemostasis and Thrombosis. Philadelphia, PA, Lippincott, 1987, pp 1252-1265 21. Lensing AWA, BOller HR, Prandoni P, et al: Contrast venography, the gold standard for the diagnosis of deep vein thrombosis: Improvement in observer agreement. Thromb Haemost 67:8-12, 1992 22. Lensing AWA, Prandoni P: Distribution of venous thrombi in symptomatic patients. Thromb Haemost 65: 1176, 1991 (abstr) 23. Hull RD, van Aken WG, Hirsh J, et al: Impedance plethysmography using the occlusive cuff technique in the diagnosis of venous thrombosis. Circulation 53:696-700, 1976 24. MacLachlan MSF, Thomson JG, Taylor DW, et al: Observer variation in the interpretation of lower limbs venograms. Am J Radiol 132:227-229, 1979 25. Bettmann MA, Paulin S: Leg phlebography: The incidence, nature and modification of undesirable side effects. Radiology 122:101-104, 1977 26. Bettmann MA, Robbins A, Braun SD, et al: Contrast venography of the leg: Diagnostic efficacy, tolerance, and complication rates with ionic and nonionic contrast media. Radiology 165:113-116, 1987 27. Katayama H, Yamaguchi K, Kozuka T, et al: Adverse reactions to ionic and nonionic contrast media. A report from the Japanese committee on the safety of contrast media. Radiology 175:621-628, 1990 28. Barrett BJ, Parfrey PS, McDonald JR, et al: Nonionic Iow-osmolality versus ionic high-osmolality contrast material for intravenous use in patients perceived to be at high risk: Randomized trial. Radiology 183:105-110, 1992 29. Moore RD, Steinberg EP, Powe NR, et al: Nephrotoxicity of high-osmolality versus low-osmolality contrast media: Randomized clinical trial. Radiology 182:649-655, 1992 30. Ginsberg JS, Hirsh J, Rainbow AJ, et al: Risks to the fetus of radiologic procedures used in the diagnosis of maternal venous thromboembolic disease. Thromb Haemost 61:189-196, 1989 31. Maternal and Neonatal Haemostasis Working Party of the Haemostasis and Thrombosis Task: Guidelines on the prevention, investigation and management of thrombosis associated with pregnancy. J Clin Pathol 46:489-496, 1993 32. Anderson FA Jr: Impedance plethysmography in the diagnosis of arterial and venous disease. Ann Biomed Eng 12:79-102, 1984 33. Wheeler HB, O'Donnell JA, Anderson FA Jr, et al: Occlusive impedance phlebography: A diagnostic procedure for venous thrombosis and pulmonary embolism. Prog Cardiovasc Dis 17:199-205, 1974 34. Hull RD, Taylor DW, Hirsh J, et at: Impedance plethysmography: The relationship between filling and sensitivity and specificity for proximal vein thrombosis. Circulation 58:898-902, 1978 35. Flanigan DP, Goodreau JJ, Burnham SJ, et al: Vascular laboratory diagnosis of clinically suspected acute deep vein thrombosis. Lancet 2:331-334, 1978 36. Wheeler HB, Anderson FA Jr: Can non-invasive tests be used as the basis for treatment of deep vein thrombosis,
DIAGNOSIS OF DVT
in Bernstein EF (ed): Non-Invasive Diagnostic Techniques in Vascular Disease. St Louis, MO, Mosby, 1981, pp 545-559 37. Peters SHA, Jonker JJC, de Boer AC, et al: Home diagnosis of deep vein thrombosis with impedance plethysmography. Thromb Haemost 48:297-300, 1982 38. Heyboer H, Bfiller HR, Lensing AWA, et al: A randomized comparison of the clinical utility of real-time compression ultrasonography versus impedance plethysmography in the diagnosis of deep vein thrombosis in symptomatic outpatients. N Engl J Med 329:1365-1369, 1993 39. Prandoni P, Lensing AWA, B~ller HR, et al: Failure of computerized impedance plethysmography in the diagnostic management of patients with clinically suspected deep vein thrombosis. Thromb Haemost 65:233-236, 1991 40. Anderson D, Lensing AWA, Wells PS, et al: Limitations of impedance plethysmography in the diagnosis of clinically suspected deep-vein thrombosis. Ann Intern Med 118:25-30, 1993 41. Hull RD, Raskob GE, Carter CJ, et al: Serial impedance plethysmography in pregnant patients with clinically suspected deep vein thrombosis. Ann Intern Med 112:663-667, 1990 42. Huisman MV, Bfiller HR, ten Cate JW: Utility of impedance plethysmography in the diagnosis of recurrent deep vein thrombosis. Arch Intern Med 148:681-684, 1988 43. Paiement G, Wessinger SJ, Waltman AC, et al: Surveillance of deep vein thrombosis in asymptomatic total hip replacement patients. Impedance plethysmography and fibrinogen scanning versus roentgenographic phlebography. Am J Surg 155:400-404, 1988 44. Cruickshank MK, Levine MN, Hirsh J, et al: An evaluation of impedance plethysmography and 125Ifibrinogen leg scanning in patients following hip surgery. Thromb Haemost 62:830-834, 1989 45. Ginsberg JS, Caco CC, Brill-Edwards P, et al: Venous thrombosis in patients who have undergone major hip or knee surgery: Detection with compression ultrasound and impedance plethysmography. Radiology 181:651-654, 1991 46. Harris WH, Athanasoulis C, Waltman AC, et al: Cuff impedance phlebography and |25I fibrinogen leg scanning versus roentgenographic phlebography for diagnosis of thrombophlebitis following hip surgery. J Bone Joint Surg (Am) 58:939-944, 1976 47. Hull RD, Raskob GE, Gent M, et al: Effectiveness of intermittent pneumatic leg compression for preventing deep vein thrombosis after total hip replacement. JAMA 263:2313-2317, 1990 48. Lensing AWA, Prandoni P, Brandies DPM, et al: Detection of deep vein thrombosis by real-time B-mode ultrasonography. N Engl J Med 320:342-345, 1989 49. Effeney D J, Friedman MD, Gooding GAW: Ileofemoral venous thrombosis: Real-time ultrasound diagnosis, normal criteria, and clinical application. Radiology 150:787792, 1984 50. Dauzat MM, Laroche JP, Charras C, et al: Real-time B-mode ultrasonography for better specificity in the noninvasive diagnosis of deep venous thrombosis. J Ultrasound Med 5:625-631, 1986 51. Raghavendra BN, Horii SC, Hilton S, et al: Deep venous thrombosis: Detection by probe compression of veins. J Ultrasound Med 5:89-95, 1986
11
52. Aitken AGF, Godden DJ: Real-time ultrasound diagnosis of deep vein thrombosis: A comparison with venography. Clin Radiol 38:309-312, 1987 53. Appelman PT, de Jong TE, Lampman LE, et al: Deep venous thrombosis of the leg: US findings. Radiology 163:743-746, 1987 54. Cronan J J, Dorfman GS, Scola FH, et al: Deep venous thrombosis: US assessment using vein compressibility. Radiology 162:191-194, 1987 55. Habscheid W, Hohmann M, Wilhelm T, et al: Realtime ultrasound in the diagnosis of acute deep venous thrombosis of the lower extremity. Angiology 41:599-608, 1990 56. Cronan JJ, Leen V: Recurrent deep venous thrombosis: Limitations of US. Radiology 170:739-742, 1989 57. Borris LC, Christiansen HM, Lassen MR, et al: Comparison of real-time B-mode ultrasonography and bilateral ascending phlebography for detection of post-operative deep vein thrombosis following elective hip surgery. Thromb Haemost 61:363-365, 1989 58. Froelich JA, Dorfman GS, Cronan JJ, et al: Compression ultrasonography for the detection of deep venous thrombosis in patients who have a fracture of the hip. J Bone Joint Surg (Am) 71:249-256, 1989 59. Borris LC, Christiansen HM, Lassen MR, et al: Real-time B-mode ultrasonography in the diagnosis of post-operative deep vein thrombosis in non-symptomatic high-risk patients. Eur J gasc Surg 4:473-475, 1990 60. Agnelli G, Volpato R, Radicchia S, et al: Accuracy of real-time B-mode ultrasonography in the diagnosis of asymptomatic deep vein thrombosis in hip surgery patients. Thromb Haemost 65:1175, 1991 (abstr) 61. Sigel B, Felix WR, Popky GL, et al: Diagnosis of lower limb venous thrombosis by Doppler ultrasound technique. Arch Surg 104:174-179, 1972 62. Holmes MCG: Deep venous thrombosis of the lower limbs diagnosed by ultrasound. Med J Aust 4:210-211, 1973 63. Medway J, Nicolaides AN, Walker CJ, et al: Value of Doppler ultrasound in diagnosis of clinically suspected deep vein thrombosis. Br Med J 4:552-554, 1975 64. Dosick SM, Blakemore WS: The role of Doppler ultrasound in acute deep vein thrombosis. Am J Surg 136:265-268, 1978 65. Flanigan DP, Goodreau JJ, Burnham SJ, et al: Vascular laboratory diagnosis of clinically suspected acute deep vein thrombosis. Lancet 2:331-334, 1978 66. Bendick PJ: Pitfalls of the Doppler examination for venous thrombosis. Am J Surg 49:320-323, 1983 67. Zielinsky A, Hull R, Carter C, et al: Doppler ultrasonography in patients with clinically suspected deep vein thrombosis: improved sensitivity by inclusion of posterior tibial vein examination site. Thromb Haemost 50:153, 1983 (abstr) 68. Lensing AWA, Levi MM, B~ller HR, et al: An objective Doppler method for the diagnosis of deep vein thrombosis. Ann Intern Med 113:9-14, 1990 69. Cogo A, Lensing AWA, Prandoni P, et al: Comparison of real-time B-mode ultrasonography and Doppler ultrasound with contrast venography in the diagnosis of venous thrombosis in symptomatic outpatients. Thromb Haemost 70:404-407, 1993
12
70. Polak JF, Culter SS, O'Leary DH: Deep veins of the calf: Assessment with color Doppler flow imaging. Radiology 171:481-485, 1989 71. Rose SC, Zwiebel WJ, Nelson BD, et al: Symptomatic lower extremity deep venous thrombosis: Accuracy, limitations, and role of color duplex flow imaging in diagnosis. Radiology 175:639-644, 1990 72. Schindler JM, Kaiser M, Gerber A, et al: Colour coded duplex sonography in suspected deep vein thrombosis of the leg. Br Med J 301:1369-1370, 1991 73. Baxter GM, McKechnie S, Duffy P: Colour Doppler ultrasound in deep venous thrombosis: A comparison with venography. Clin Radio142:32-36, 1990 74. Mattos MA, Londrey GL, Leutz DW, et al: Colorflow duplex scanning for the surveillance and diagnosis of acute deep venous thrombosis. J Vasc Surg 15:366-376, 1992 75. Sumner DS, Londrey GL, Spadone DP, et al: Study of deep venous thrombosis in high-risk patients using color flow Doppler, in Bergan JJ, Yao JST (eds): Venous Disorders. Philadelphia, PA, Saunders, 1991, pp 63-76 76. Rose SC, Zwiebel WJ, Murdock LE, et al: Insensitivity of color Doppler flow imaging for detection of acute calf deep venous thrombosis in asymptomatic postoperative patients. J Vasc Intervent Radiol 4:111-117, 1993 77. Davidson BL, Elliott CG, Lensing AWA: Low accuracy of color Doppler ultrasound in the detection of proximal leg vein thrombosis in asymptomatic high risk patients. Ann Intern Med 117:735-738, 1992 78. Atkins P, Hawkins LA: Detection of venous thrombosis in the legs. Lancet 1:1217-1219, 1965 79. Nanson EM, Palko PD, Dick AA, et al: Early detection of deep vein thrombosis of the leg using 1311tagged human fibrinogen: A clinical study. Ann Surg 162:438445, 1965 80. Flanc C, Kakkar VV, Clarke MB: Detection of venous thrombosis of the legs using 12SI-labelled flbrinogen. Br J Surg 55:742-747, 1968 81. Kakkar VV: The diagnosis of deep vein thrombosis using the 125I-flbrinogen test. Arch Surg 104:152-159, 1972 82. Kakkar VV: Deep vein thrombosis. Detection and prevention. Circulation 51:8-19, 1975 83. Tsapogas MJ, Goussus H, Peabody RA, et al: Postoperative venous thrombosis and the effectiveness of prophylactic measures. Arch Surg 103:561-567, 1971 84. Myrvold HE, Persson JE, Svensson B, et al: Prevention of thromboembolism with Dextran 70 and heparin in patients with femoral neck fractures. Acta Chir Scand 139:609-616, 1973 85. Sautter RD, Larson DE, Bhattaracharyya SK, et al: The limited utility of fibrinogen 1251 leg scanning. Arch Intern Med 139:148-153, 1979 86. Lensing AWA, Hirsh J: 12sI-fibrinogen leg scanning: Reassessment of its role for the diagnosis of venous thrombosis in post-operative patients. Thromb Haemost 69:2-7, 1993 87. Webber MM, Bennett LR, Cragin M, et al: Thrombophlebitis--Demonstration by scintiscanning. Radiology 96: 620-623, 1969 88. Lisbona R, Stern J, Derbekyan V: 99mTc red blood
KOOPMAN, VAN BEEK, AND TEN CATE
cell venography in deep vein thrombosis of the leg: A correlation with contrast venography. Radiology 143:771773, 1982 89. Ezekowitz MD, Pope CF, Sostman HD, et al: Indium111 platelet scintigraphy for the diagnosis of acute venous thrombosis. Circulation 73:668-674, 1986 90. Tibbutt DA, Chesterman CN, Allington M J, et al: Measurement of fibrinogen-fibrin-related antigen in serum as aid to diagnosis of deep vein thrombosis in outpatients. Br Med J 1:367-369, 1975 91. Yudelman IM, Nossel HL, Kaplan KL, et al: Plasma fibrinopeptide A levels in symptomatic venous thromboembolism. Blood 51:1189-1195, 1978 92. Elms M J, Bunce IK, Bundesen PG, et al: Measurement of crosslinked fibrin degradation products--An immunoassay using monoclonal antibodies. Thromb Haemost 50:591-594, 1983 93. Rowbotham BJ, Carroll P, Whitaker AN, et al: Measurement of crosslinked fibrin derivatives--use in the diagnosis of venous thrombosis. Thromb Haemost 57:59-61, 1987 94. Heaton DC, Billings JD, Hickton CM: Assessment of D-dimer assays for the diagnosis of deep vein thrombosis. J Lab Clin Med 1110:588-591, 1987 95. Ott P, Astrup L, Jensen RH, et al: Assessment of D-dimer in plasma: Diagnostic value in suspected deep venous thrombosis of the leg. Acta Med Scand 224:263-267, 1988 96. Bounameaux H, Schneider PA, Reber G, et al: Measurement of plasma D-dimer for diagnosis of deep venous thrombosis. Am J Clin Pathol 91:82-85, 1989 97. Speiser W, Mallek R, Koppensteiner R, et al: Ddimer and TAT measurements in patients with deep venous thrombosis: Utility in diagnosis and judgement of anticoagulant treatment effectiveness. Thromb Haemost 64:196-201, 1990 98. Boneu B, Bes G, Pelzer H, et al: D-dimers, thrombinantithrombin III complexes and prothrombin fragments 1 + 2: Diagnostic value in clinically suspected deep vein thrombosis. Thromb Haemost 65:28-31, 1991 99. Heyboer H, Ginsberg JS, BOiler HR, et al: The use of the D-dimer test in combination with non-invasive testing versus serial non-invasive testing alone for the diagnosis of deep-vein thrombosis. Thromb Haemost 67:510-513, 1992 100. Jensen C, Lomboldt Knudsen L, Hegediis V: The role of contact thermography in the diagnosis of deep venous thrombosis. Eur J Radiol 3:99-102, 1983 101. Sandier DA, Martin JF: Liquid crystal thermography as a screening test for deep-vein thrombosis. Lancet 1:665-667, 1985 102. Bounameaux H, Khabiri E, Huber O, et al: Value of liquid crystal contact thermography and plasma level of D-dimer for screening of deep venous thrombosis following general abdominal surgery. Thromb Haemost 67:603-606, 1992 103. Thomas PRS, Butler CM, Bowman J, et al: Light reflection rheography: An effective non-invasive technique for screening patients with suspected deep venous thrombosis. Br J Surg 78:207-209, 1991
Cardiovascular Diseases VOL XXXVII, NO 1
JULY/AUGUST 1994
Diagnosis of Deep Vein Thrombosis M.M.W. Koopman, E.J.R. van Beek, and J.W. ten Cate
OR SEVERAL decades, the diagnosis deep
vein thrombosis (DVT) had been based on F clinical grounds alone, because the inaccuracy of the clinical diagnosis was not well appreciated. With the introduction of contrast venography, it became apparent that clinical diagnosis was insensitive and nonspecific. Many studies have now shown that only 20% to 30% of patients with the classical signs and symptoms of DVT indeed have venous thrombosis on venography) -4 The reason for this inaccuracy of clinical diagnosis is that none of the signs and symptoms of DVT are unique for DVT, and many other disorders can mimic venous thrombosis. These disorders include muscle strain, direct twisting injury to the leg, swelling in a paralyzed leg, venous stasis, postthrombotic syndrome, lymphangitis, muscle tear, ruptured Baker's cyst, cellulitis, and orthopedic lesions of the knee and ankle joint. In a consecutive series of 160 patients with signs and symptoms of DVT, it was possible to make an alternative diagnosis in the majority of patients after ruling out venous thrombosis by venography. 5 The clinical diagnosis is also insensitive because thrombi may occasionally be nonobstructive with minimal symptoms, such as is observed in postoperative DVT. Thus, objective testing is mandatory. Moreover, treatment of all patients with clinically suspected DVT will result in unnecessary exposure of patients without the disease to anticoagulants, which are associated with the potential hazard of treatment-induced bleeding. 6-s Venography has been the reference method for the diagnosis of DVT (Fig 1). 1,9,10However, this procedure is invasive and may be associated with discomfort and side effects, such as pain, allergic reactions, nausea, and vomiting. 11 Fur-
thermore, venography has a limited feasibility, especially in severely ill patients. These disadvantages have led to the development of several noninvasive tests over the past two decades. A new diagnostic method requires extensive evaluation before it can be safely introduced into medical practice. The test should be standardized and should be compared with venography in consecutive patients to determine the diagnostic accuracy. Only a few of the available noninvasive tests have been evaluated adequately and will be discussed in more detail. The accuracy of the test depends on the nature of the patient group under consideration; therefore, it is important to distinguish three different patient categories, (1) patients with a suspected first episode of DVT, (2) patients with recurrent thrombosis, and (3) asymptomatic patients at high risk for DVT. Thrombi in these patient groups differ in their origin, localization, and extension and, therefore, influence the test outcome. Most of the thrombi in patients with a first episode of DVT have extended proximally. 2,12 These patients are at high risk for pulmonary embolism (PE). In a study of 101 patients with venography-proven DVT, 51% already had scintigraphic proof of (silent) PE. 13 Conversely, isolated calf vein thrombi appear at lower risk of embolization; however, approximately 25%
From the Centrefor Hemostasis, Thrombosis,Atherosclerosis, and Inflammation Research, Academic Medical Centre, University of Amsterdam, Amsterdam, the Netherlands. Address reprint requests to M.M.W. Koopman, MD, Centre for Hemostasis, Thrombosis, Atherosclerosis, and Inflammation Research, F4-133,Academic Medical Centre, Universityof Amsterdam, Amsterdam, the Netherlands. Copyright 9 1994 by W.B. Saunders Company 0033-0620/94/3701-000155.00/0
Progress in Cardiovascular Diseases, Vol XXXVII, No 1 (July/August), 1994: pp 1-12
1
2
Fig 1. Contrast venogram shows a filling defect in the femoral vein.
will extend proximally into the popliteal vein and may subsequently lead to pulmonary emboli. 14,15 Therefore, the diagnostic approach must be sensitive and specific to detect proximal DVT as well as calf vein thrombosis extending into the popliteal vein. After a first episode of DVT, approximately 10% to 20% of the patients will return with recurrent signs and symptoms.16,17 Only about 30% to 40% of these patients actually have
KOOPMAN, VAN BEEK, AND TEN CATE
recurrent DVT. Difficulties in the diagnosis of recurrent DVT exist for all available tests. Venography has its limitations because it cannot readily be repeated and because the interpretation is often difficult because of anatomical changes of the venous system, vessel damage, and resultant irregularities of the vessel wall caused by previous thrombi. Impedance plethysmography (IPG) is a reliable method in the diagnosis of recurrent DVT, as long as a normal test result was obtained after a previous thrombotic event.16 For compression ultrasonography, similar rules as for IPG apply. However, a follow-up study with this technique showed that the ultrasound investigations normalized within 1 year after a first event of DVT in only 50% of the patients. 18 The clinical utility of a newly developed compression ultrasound method to measure residual thrombus and its application for the diagnosis of recurrent DVT seems promising and is currently under investigation.19 Patients who have undergone major surgery or who were immobilized are at high risk for developing asymptomatic thrombosis. Thrombi in these patients are often small and nonocclusive.2~ However, these small thrombi may extend proximally and lead to PE. Therefore, the diagnostic approach in these patients must be able to detect these small thrombi. Furthermore, the test must be feasible in a large group of often immobilized and sometimes critically ill patients. At the moment, only venography has been extensively evaluated and proven sufficiently accurate in this patient category. The diagnosis of DVT with the use of various diagnostic techniques in acute and recurrent symptomatic patients as well as in asymptomatic high risk patients will be subsequently discussed. The emphasis will be on those tests that have undergone adequate evaluation. However, some brief remarks with regards to other, less well-investigated methods will be included. CONTRAST VENOGRAPHY
Lower extremity contrast venography is generally regarded as the reference method ("gold standard") for the diagnosis of DVT and enables the visualization of the entire deep venous system of the lower extremity and the large pelvic veins to the vena cava. 1,9,21 Contrast venography uses radio-opaque (iodinated) contrast medium, which is usually
DIAGNOSIS OF DVT
injected into a dorsal foot vein.9 Multiple injections may be required, and views of the entire deep system in at least two projections are required to establish a definite diagnosis. For this purpose, both separate films of the calf, knee, thigh, and pelvic region 9 or long-leg films may be used. 21 Venographic criteria for acute DVT have been defined previously.9,21 The most reliable criterion is the presence of an intraluminal filling defect that remains constant in shape and location in at least two different projections (Fig 1). Other, less reliable criteria are nonfilling of a segment with abrupt termination and reappearance of contrast media below and above the segment and nonfilling of the deep venous system in the trajectory of the popliteal and femoral vein, respectively. As with any diagnostic technique, contrast venography may lead to inadequate investigations or observer disagreement. Inadequate venograms may occur in as many as 21% of all examinations;5,14-16,21-23 however, this may be even more frequent in the hands of less experienced radiologists. Observer disagreement on the absence or presence of thrombi occurs in approximately 10% of adequate venograms. 21,24 A randomized study in 249 symptomatic patients with a first episode of clinically suspected DVT compared the use of separate films (Rabinov-Paulin technique) versus long-leg films (Fig 2). 2~ The inadequacy rate was 20% and 2% for the separate films and the long-leg films groups, respectively. At the same time, the observer disagreement decreased from 21% in the group with separate exposures to 4% in the long-leg film group. 2a Therefore, this latter technique may be recommended. The importance of contrast venography for the clinical management of symptomatic patients was shown by a study in 106 consecutive patients with clinical signs of DVT in whom anticoagulants were withheld after a normal venogram.5 Only 2 patients (1.3%) returned within 5 days and were shown to have developed DVT. No patients developed PE or died at follow-up for 3 months, which shows the safety of this approach. In patients with recurrent signs and symptoms of DVT, venography may be of diagnostic value, especially if a previous venogram is available for comparison. In view of the limited
3
Fig 2, Normal long-leg venogram outlines the complete deep venous system of the left leg.
diagnostic accuracy of noninvasive tests in patients with asymptomatic thrombi (ie, high-risk [post surgery] patients), venography is the diagnostic test of choice. The disadvantages of venography are related to the fact that it is invasive and requires the use of contrast media. The main adverse effects are discomfort and contrast-media-related complications, u,25 The latter are more frequently encountered with conventional hyperosmolar agents and are less frequently encountered with recently developed, low-osmolar contrast media. 11,26-29 In a study of 463 DVT patients who underwent venography with low-osmolar contrast media, the following side effects were
4
KOOPMAN, VAN BEEK, AND TEN CATE
recorded. Two patients (0.4%) developed severe adverse reactions (bronchospasm), and 103 patients (22.2%) suffered from minor side effects such as: local pain, 12.1%; nausea, 4.5%; dizziness, 1.5%; skin reactions, 1.1%; vomiting, 0.4%; edema, 1.9%; and superficial thrombopflebitis, 0.6%. 11 The frequency of postvenography DVT, as shown by 1251-fibrinogen leg scanning in patients with initial normal venograms, may be as high as 9%. 5'11'26 However, this relatively high frequency was only observed with leg scanning and could not be confirmed by repeated contrast venography. Finally, contrast venography cannot be performed in up to 10% of patients because of the inability to acquire venous access, the presence of an obvious local infection of the leg, a history of allergic reactions to contrast media, or renal insufficiency. 11,14,~5,23 Although pregnancy is a relative contraindication for performing any radiologic investigation, limited contrast venography with abdominal shielding reduces radiation to the fetus to acceptable doses. 3~ IMPEDANCE PLETHYSMOGRAPHY IPG, a noninvasive method, is based on the principle that blood volume changes of the leg lead to changes in electrical resistance (impedance). During the investigation, the patient is in the supine position with the lower limb elevated 25 ~ to 30~ and with the knee flexed 10~ to 20~ A pneumatic cuff around the thigh is inflated to a pressure of 50 cm HzO, thereby occluding the venous outflow of the leg. Circumferential calf electrodes detect the changes in impedance and are connected to a recorder. After a predetermined period, the cuff is deflated and both the total increase and decrease during cuff inflation and deflation during the first 3 seconds are plotted on a two-way IPG graph. The graph includes a discriminant line separating normal and abnormal test results. 23,32,33 The IPG test can be falsely normal if thrombi are nonocclusive or if proximal DVT is associated with well-developed collaterals. Furthermore, the IPG approach cannot distinguish between thrombotic and nonthrombotic obstruction of venous flow. Several other disorders, such as venous compression by an extravascular mass or raised central venous pressure, can produce a falsely abnormal test result. How-
ever, accuracy studies using contrast venography as the reference method have shown an overall mean sensitivity and specificity of 92% and 95%, respectively, for patients with a first episode of proximal DVT. 23,34-37The diagnostic accuracy for calf vein thrombosis is much lower, with a sensitivity of about 20%. The specificity is about 95%; this means that an abnormal test result can be used to make therapeutic decisions, but a normal test result does not rule out calf vein thrombosis and/or nonocclusive proximal thrombosis. From follow-up studies, it is known that most of the calf vein thrombi resolve spontaneously and that about 20% of these thrombi may extend proximally. To detect these extending thrombi it is necessary to repeat the noninvasive tests over a period of 1 week. Therefore, management studies have been designed using repeated (serial) IPG testing to detect these thrombi. Several studies have evaluated the use of serial IPG in patients with a first episode of signs and symptoms suggesting DVT and have showed a high positive predictive value of 83% to 95%. 14-16,38 m summary of these studies is shown in Table 1. The safety of withholding anticoagulant therapy in patients with repeated normal IPG findings was investigated in all these studies. However, over the years, duration and number of IPG assesments at follow-up has steadily decreased without compromising the safety. Initially, IPG was performed at days 1, 2, 5, 7, 10, and 14.15 With this regimen, 6 of 311 patients (1.9%) with normal repeated IPG resuits returned with documented DVT within 1 year, and no fatalities occured. 15 A subsequent study omitted the 14-day IPG test, and only i of 289 patients (0.3%) with normal IPG findings returned with PE during a 6-month follow-up Table 1. Summary of Various Studies Evaluating Serial IPG
Reference
Positive Predictive Value (n/n)
Hull et al, 1985 TM Huisman et al, 1986 TM Huisman et al, 1989 TM Heyboer et al, 19933s
90/95 113/123 38/42 89/107
(95) (92) (90) (83)
Total
330/367 (90)*
RecurrentVTE (n/n) in Normal Serial IPG 6/311 1/289 1/131 9/369
(1.9) (0.3) (0.8) (2.5)
17/1100 (1.5)1
The values shown in parentheses are percentages. *95% CI, 86% to 94%. 1"95% CI, 0.9% to 2.1%.
DIAGNOSIS OF DVT
period. 14 The next two studies performed IPG at days 1, 2, and 7 only. 16'38 In a community hospital study, 1 of 131 patients (0.8%) with repeated normal IPG returned with DVT, 16 whereas in a similar study in symptomatic outpatients, 9 of 369 patients (2.5%) with normal serial IPG results returned with objectively documented venous thromboembolism (VTE).38 Despite these satisfactory results, two recent studies reported a much lower reliability of IPG. 39,4~ In one study, a recently developed, computerized version of IPG was used. 39 In 311 patients with a normal IPG test result, no anticoagulant therapy was administered. During follow-up, 10 patients (3.2%) developed VTE, including 4 fatal episodes of PE (1.3%) at day 3, 5, and 8 and at 5 months. 39 The other (retrospective cohort) study reevaluated the accuracy of a newer conventional, noncomputerized, IPG machine for detection of DVT. In this investigation, IPG showed a sensitivity of only 66% for proximal DVT. 4~Possible explanations for this discrepancy with previous studies is the use of newer, less sensitive equipment and an increased awareness of DVT among referring physicians, resulting in less extensive thrombi at onset. These results emphasize the great importance of proper evaluation of diagnostic tests and equipment before they can be introduced safely into medical practice. Although less extensively evaluated, IPG seems to be a safe and effective method to detect DVT in pregnant woman. 41 In a study in 139 pregnant women in whom anticoagulants were withheld on the basis of normal serial IPG findings, no thromboembolism (95% confidence interval [CI], 0% to 2.6%) was encountered during long-term follow-up.41 In patients with suspected recurrent DVT, IPG can only be used if a normal baseline test result is available. Prospective cohort studies have shown that, 12 months after the acute thrombotic episode, 95% of the IPG tests return to normal. 42 If the IPG test has returned to normal, it can be used alone or in combination with fibrinogen leg scanning to diagnose recurrent DVT in symptomatic patients. 17,42 In one study, 270 patients were referred with clinical signs and symptoms of recurrent DVT. IPG was negative in 200 of these patients. In all IPGnegative patients, leg scanning was performed, and normal test results were obtained in 181 of
5
these 200 IPG-negative patients, and anticoagulant therapy was withheld. During 3 months of follow-up, 3 of 181 patients (1.7%) were shown to have recurrent DVT, despite normal IPG and legscan test results. In another study of patients with suspected recurrent DVT, anticoagulants were withheld safely in all 18 patients with repeated normal IPG tests. IPG is not useful in the management of asymptomatic postoperative patients because of its low overall sensitivity of 22%. The additional use of 125I-fibrinogen leg scanning improves sensitivity to approximately only 50%. 43-47 ULTRASONOGRAPHY
Real- Time B-Mode Ultrasonography Real-time B-mode ultrasonography (US) is a noninvasive method that allows direct visualization of the deep veins of the leg. During the examination, the patient is in the supine position. The ultrasound probe is placed in the groin to identify the common femoral vein (medial to the common femoral artery). The whole venous trajectory is then visualized throughout its course until the superficial femoral vein is lost in the adductor canal. For examination of the popliteal vein, the patient is in the prone or lateral position, with flexed knees. The visualization of the calf veins is more difficult because of the variable anatomical positions of the veins and the small size of the vessels. Using a 5- to 10-mHz high-resolution transducer, it is not only possible to localize deep veins but also to visualize the thrombus in the vessel, one of the suggested diagnostic criteria for DVT. Further diagnostic criteria comprise noncompressibility of the vein and the absence of vein extension upon a Valsalva manoevre. However, a large prospective study showed that the single criterion of noncompressibility of the vein on gentle pressure with the probe is the most accurate for the detection of DVT (Fig 3a and b). 48 Most of the compression ultrasonography (CUS) accuracy studies have limited the examination to the common femoral vein in the groin and the vein in the popliteal fossa. The overall mean sensitivity and specificity of CUS is high, 97% (range, 83% to 100%) and 97% (range, 86 to 100%), respectively.48-55Using this two-point CUS, isolated thrombi in the superficial femoral vein and in isolated calf veins are missed.
6
KOOPMAN, VAN BEEK, AND TEN CATE
Fig 3. (A) B-mode US shows the popliteal vein (PV} and artery (PA) without compression by the probe. (B) Normal compression B*mode US leaves the popliteal artery (PA) open but completely occludes the popliteal vein.
However, a recent venogram study showed the relative low frequencies of these localizations (ie, none of the patients [0%; 95% CI, 0% to 1.9%] had an isolated, superficial femoral thrombus, and isolated calf vein thrombosis was detected in only 12% of the patients. ~-2Therefore, an abnormal two-point CUS result can be used to make therapeutic decisions; however, a norreal test result does not rule out calf vein thrombosis. Although most of the calf vein thrombi resolve spontaneously, approximately 20% extend proximally and may lead to PE. Therefore, CUS should be repeated over a 1-week period to detect these extending thrombi, which is similar to the approach used for I P G ? s Falsely normal test results can be obtained in patients with duplication of the popliteal vein and thrombus in one of these popiteal venous systems. Alternatively, a dilated collateral vein may be misinterpreted as a normal, open venous system in the presence of an acute, completely occluding thrombus in the popliteal or femoral vein. Compression of the vein by extravascular masses, such as a tumor or a gravid uterus, may cause noncompressibility of the vein at a lower level as a result of increased venous pressure and may be the reason for a falsely abnormal test result. In patients with a previous history of DVT, residual thrombus mass in the vein can mimick acute thrombosis. In a recent study, the safety of withholding
anticoagulant therapy in symptomatic patients with serial normal CUS and serial normal IPG results was compared? 8 From a total of 985 consecutive symptomatic outpatients, 491 were randomized to undergo serial CUS. Of this group, 89 patients with an abnormal test result underwent venography that confirmed D V T in 84 patients (positive predictive value, 94%). No anticoagulant therapy was administered to the patients with a repeated normal CUS test result. During a 6-month follow-up, 6 of those patients (1.5%; 95% CI, 0.5% to 3.3%) had recurrent V T E that was confirmed by venogram. 38 Hence, the use of serial CUS appears to be a safe method in the management of patients with signs and symptoms of DVT. The usefulness of B-mode CUS in patients with recurrent thrombosis seems limited and depends on the rate of normalization of the test during follow-up. 18,56 In a prospective cohort study, the CUS results became normal in only 50% of the patients after 1 year. is A newly developed quantitative method, which is based on measuring thrombus size, is now under investigation. 19 In asymptomatic patients, CUS seems more accurate than IPG; nevertheless, the overall mean sensitivity is unfortunately as low as 59% (range, 43% to 100%), with an overall mean specificity of 98% (range, 91% to 95%) for proximal D V T . 57-60
DIAGNOSIS OF DVT
In conclusion, CUS is useful in symptomatic patients with a first episode of acute DVT but, as yet, can not be recommended for the diagnosis of recurrent DVT. Further studies are required to determine the diagnostic accuracy in this condition. For asymptomatic DVT, CUS (ie, as a screening method) is not sensitive enough to be used in the routine management of high-risk patients.
Doppler Ultrasonography Doppler US is based on the principle that an ultrasound beam is reflected by blood cells in a vessel. If the blood is stationary, no sound is recorded. In the case of moving blood cells, a difference in frequency between the initial and reflecting beam can be converted into a flow sound?'1 During the examination, the patient is in the supine position, avoiding tight clothes that can interfere with venous flow. The common femoral artery, located in the groin, can be identified by moving the probe in a medial direction. For examination of the popliteal vein, the patient is in the prone position with the knee flexed and the feet rested on a pillow. The popliteal vein is located lateral to the popliteal artery. The venous flow sound is typically lowpitched, changes with respiration, and can be augmented by squeezing the thigh and calf. Obstruction to venous outflow may result in loss of phasicity of the venous signal and loss of the relationship with respiratory fluctuation. Furthermore, augmentation of the venous flow by squeezing the thigh and calf may be diminished o r a b s e n t . 61
Several investigators have compared the accuracy of Doppler US with contrast venography in patients with symptoms of DVT. 62-67The accuracy of the Doppler US depends on the experience of the investigator and can be disturbed by subjective interpretation of the test result. Therefore, the sensitivity and specificity for proximal DVT ranges from 65% to 100% and 72% to 98%, respectively.62-67The introduction of a standardized protocol and the use of strip-chart ultrasound recordings have improved the accuracy of the test. 6s Using the objective criteria indicative for proximal DVT (ie, a continuous or absent venous Doppler signal), the sensitivity and specificity can be increased to 91% and 99%, respectively.6s In a recently published study, real-time CUS
7
and Doppler US were compared with venography in a consecutive series of 161 outpatients referred for clinically suspected DVT. 69 For proximal DVT, the sensitivity for CUS was 100% and for Doppler US was 95% (P = .056). However, for all thrombi (including isolated calf vein thrombi) the sensitivity of CUS was significantly better than that of Doppler US (95% and 76%, respectively; P < .04). The specificity of both tests was not significantly different (100% for compression US and 98% for Doppler U S ) , 69
In conclusion, this study showed the superiority of real-time CUS, when compared with Doppler US, for detection of all venous thrombi and equal accuracy in detecting proximal DVT. However, prospective studies on the safety of withholding anticoagulant therapy in patients with normal Doppler US results are lacking. The value of Doppler US in patients with recurrent DVT and in postoperative asymptomatic patients at high risk have not been evaluated.
ColorDoppler Ultrasonography The addition of color Doppler utilities to the conventional US equipment allows direct visualization of the flow within vessels, which might imply a more accurate and easier performance of the test. 7~However, the sensitivity and specificity (both 97%) of color Doppler US for proximal DVT in symptomatic patients was similar to that of C U S . 71-74 The potential advantage of color Doppler for the detection of thrombi in asymptomatic postoperative patients, which are often small and nonocclusive, was not confirmed in clinical studies, which showed a mean sensitivity of only 43% with a specificity of 9 7 % . 75-77 In conclusion, the extra costs that are required to perform color Doppler US do not result in improved diagnostic accuracy compared with CUS. Additional comparative studies are currently in progress. SCINTIGRAPHIC METHODS
125I-FibrinogenLeg Scanning Fibrin makes up an essential part of thrombi, and radiolabeled fibrinogen, after cleavage by thrombin, becomes incorporated in thrombi. Subsequently, the radioactivity may be detected from the surface of the limb. The method was introduced over 30 years ago 78-8~and was subse-
8
KOOPMAN, VAN BEEK, AND TEN CATE
quently used in the diagnosis of asymptomatic DVT in high-risk patients as an investigative tool to evaluate the effectiveness of thromboprophylaxis. 81-84 Although initial studies showed very high sensitivity for distal thrombi, this was later disputed by several other studies. 43,44,85In a recent review of the literature, accuracy studies and positive predictive value studies were assessed to address this apparent change in sensitivity of the 125I-fibrinogen-uptake test in later years. 86 Bias in the earlier studies was identified as the most probable cause of the apparent decrease in sensitivity. 86 In addition, the sensitivity may also be related to the size of the thrombus and may be as low as 20% for small calf vein thrombi. 43 The virus safety of the product is another point of concern. Fibrinogen is derived from human plasma and can not be pasteurized. Thus, it is possible for blood-borne virus transmission to take place. This fear was the main reason for the product's removal from the market. At present, new attempts to produce a virus safe product are being made. In conclusion, 125I-fibrinogen leg scanning does not appear to be useful for the diagnosis of DVT at present. If the test were to become available again, new studies may be required to define its role.
Other Scintigraphic Methods Some techniques, such as 99mTc-labeled macroaggregated albumin or 99mTc-labeled erythrocytes, are based on the fact that perfusion is absent in the presence of obstructing thrombi. 87,88 Other techniques are based on incorporation of 111In-labeled platelets into the thrombus. 89 However, these techniques appeared to be inaccurate and were never introduced in clinical practice. BLOOD TESTS
For several decades, investigators have attempted to develop a blood test that could adequately diagnose DVT. Fibrinopeptide A, fibrin(ogen) degradation products (FDP), degradation products of cross-linked fibrin (Ddimer) prothrombin fragments 1 + 2 (F1 + 2), and thrombin-antithrombin III (TAT) complexes have been investigated. 9~ For DVT to be adequately excluded by a blood test, the sensitivity (and, thus, the nega-
tive predictive value) would need to approximate 100%. Using this principle, a cut-off value for D-dimer would consist of the lowest Ddimer value detected in patients with DVT. This would ensure that no patients with DVT could remain unnoticed, and this would allow safe exclusion to take place. Several investigators have shown that plasma D-dimer may be able to adequately identify those patients without DVT (Table 2). For example, an enzyme immunoassay (EIA) for D-dimer reached a sensitivity of 100% in patients with suspected DVT; however, the specificity of the test was much lower in inpatients (4%) than in outpatients (57%). 93 In another study, the value of latex and EIA D-dimer assays were determined in 62 patients with suspected DVT, using venography as a reference method. 94 For latex D-dimer, an inferior sensitivity of 73% was found, whereas EIA D-dimer reached a sensitivity of 100%. This was confirmed by other investigators in 56 consecutive outpatients and inpatients with suspected DVT. 96 Of the tested D-dimer enzyme-linked immunosorbant assay (ELISA), two D-dimer latex assays, and a latex FDP test, only D-dimer ELISA showed a sensitivity of 100%. In a series of 116 consecutive outpatients with clinically suspected DVT, the diagnostic value of F1 + 2, TAT complexes, and D-dimer were compared. 98Only D-dimer ELISA showed sufficient sensitivity and specificity to be of potential clinical use. In a large study in outpatients with suspected DVT, the value of D-dimer was evaluated as a replacement of serial noninvasive tests for DVT (IPG or C U S ) . 99 In this study, D-dimer was performed in addition to noninvasive tests for DVT on day 1, and the results were compared with the findings of serial IPG or CUS outcome. After determination of a cut-off value to obtain a sensitivity of 100% when compared with Table 2. Summary of Studies in the Evaluation of the Value of Plasma D-Dimer in the Diagnosis of DVT Reference Rowbotham et al, 198793 Heaton et al, 19879== Bounarneaux et al, 198996 Speiser et al, 199097 Boneu et al, 199198 Heyboer et al, 199299
No. of Cut Off Sensitivity Specificity Patients (p,g/L) (%) (%) 104 62 53 97 116 309
250 200 200 120 500 300
100 100 100 100 94 100
36 47 31 48 51 29
DIAGNOSIS OF DVT
9
contrast venography, the assay was prospectively validated. In 309 consecutive patients, a sensitivity of 100% was obtained, but the specificity was relatively low (28%; 95% CI, 23% to 34%). A subsequently performed cost-effectiveness analysis did not show decreasing costs per diagnosed DVT when D-dimer was used in addition to either IPG or CUS. However, 10 of 70 patients with proven DVT were identified earlier using D-dimer, which could reduce the number of patients that would be required to return to the hospital for serial tests to be performed. In conclusion, D-dimer may be of potential clinical use in the exclusion of DVT. Unfortunately, management studies have yet to be performed. Furthermore, the reliability of fast D-dimer tests remains to be proven. Therefore, at present, D-dimer can not be recommended for the routine management of patients with clinically suspected DVT. OTHER NONINVASlVE SCREENING METHODS
Liquid crystal (contact) thermography is based on the fact that the temperature of a thrombotic leg will increase as compared with that of the unaffected limb. Several reports have shown a high sensitivity, which would make this test clinically useful in the exclusion of DVT. 1~176176 However, a subsequent study in 185 patients could not show a sufficiently high, negative predictive value for it to be useful as a screening test.102 Light reflection rheography uses the observation that exercise results in a decrease of blood in the venous plexus of the calf. Light that is absorbed by red blood cells will be increasingly reflected with the decreasing blood pool, whereas this returns to normal during rest. In DVT, the blood pool remains in the calf, which yields a different pattern of light reflection. This test was shown to be sensitive, but thorough evaluation has yet to take place2 ~ SUMMARY
In this chapter, various tests have been discussed in the diagnosis of DVT and have been
classified according to various patient categories. To summarize, the following guidelines may be of clinical use in the management of patients with suspected DVT.
Acute, First Event of Suspected DVT These patients often suffer from occluding, proximal thrombi. Therefore, noninvasive tests such as CUS or IPG are most suitable for these patients. If an abnormal CUS or IPG result is found, the diagnosis is virtually proven, and this may serve as a basis to treat the patient with anticoagulants. If a normal CUS or IPG result is obtained, serial testing is indicated to detect extending calf vein thrombi or nonoccluding DVT, which becomes occlusive at follow up. Anticoagulants may be withheld safely if the test remains normal within 1 week.
Acute, Recurrent Suspected DVT These patients may have residual thrombi present, which makes the noninvasive tests (CUS/IPG) less useful. However, if a normal noninvasive test was documented previous to the acute recurrent event, this test may be used. If an abnormal test result is found in the presence of a documented, normal previoustest outcome, this may serve as a basis for anticoagulant therapy. Although no formal studies have been performed to evaluate the safety of withholding anticoagulants if a normal CUS or IPG result is obtained, serial testing is likely to be adequate in these circumstances. Phlebography is the only truly evaluated approach, and this could be considered in all suspected recurrent DVT. Furthermore, contrast venography is the test of choice to discern acute from old thrombi.
Asymptomatic D VT in High-Risk Patients The majority of these thrombi are mostly localized in the calf veins only and are often nonocclusive. This makes noninvasive tests unreliable for their detection. Therefore, only contrast venography should be used in this patient category.
REFERENCES
1. Bauer G: A venographic study of thromboembolic problems. Acta ChirScandSupp161:1-75, 1940 2. HaegerK: Problemsof acute deepvenousthrombosis.
I. The interpretations of signs and symptoms.Angiology 20:219-223, 1969 3. CranleyJJ, CanosAJ, SullWJ: The diagnosisof deep
10
venous thrombosis. Fallibility of clinical symptoms and signs. Arch Surg 111:34-36, 1976 4. O'Donnel TF, Abbott WM, Athanasoulis CA, et al: Diagnosis of deep vein thrombosis in the outpatient by venography. Surg Gynecol Obstet 150:69-74, 1980 5. Hull RD, Hirsh J, Sackett DL, et al: Clinical validity of a negative venogram in patients with clinically suspected venous thrombosis. Circulation 64:622-625, 1981 6. Oudkerk M, van Beek EJR, van Putten WLJ, et al: Cost-effectiveness analysis of various strategies in the diagnostic management of pulmonary embolism. Arch Intern Med 153:947-954, 1993 7. Levine MN, Hirsh J, Landefeld S, et al: Hemorrhagic complications of anticoagulant treatment. Chest 102:352S363S, 1992 (suppl) 8. Van der Meer FJM, Rosendaal FR, Vandenbroucke JP, et al: Bleeding complications in oral anticoagulant therapy. An analysis of risk factors. Arch Intern Med 153:1557-1562, 1993 9. Rabinov K, Paulin S: Roentgen diagnosis of venous thrombosis in the leg. Arch Surg 104:134-144, 1972 10. Lensing AWA, BOiler HR, Prandoni P, et al: Contrast venography, the gold standard for the diagnosis of deep vein thrombosis: Improvement in observer agreement. Thromb Haemost 67:8-12, 1992 11. Lensing AWA, Prandoni P, BOller HR, et al: Lower extremity venography with iohexol: Results and complications. Radiology 177:503-505, 1990 12. Hull RD, Secker-Walker RH, Hirsh J: Diagnosis of deep vein thrombosis, in Colman RW, Hirsh J, Marder V J, Salzman EW (eds): Hemostasis and Thrombosis. Philadelphia, PA, Lippincott, 1987, pp 1220-1239 13. Huisman MV, BOller HR, ten Cate JW, et al: Unexpected high prevalence of silent pulmonary embolism in patients with deep venous thrombosis. Chest 95:498-502, 1989 14. Huisman MV, BOller HR, ten Cate JW, et al: Serial impedance plethysmography for suspected deep venous thrombosis in outpatients. The Amsterdam General Practitioner Study. N Engl J Med 314:823-828, 1986 15. Hull RD, Hirsh J, Carter C J, et al: Diagnostic efficacy of impedance plethysmography for clinically suspected deepvein thrombosis: a randomized trial. Ann Intern Med 102:21-26, 1985 16. Huisman MV, BOller HR, ten Cate JW, et al: Management of clinically suspected acute venous thrombosis in outpatients with serial impedance plethysmography in a community hospital setting. Arch Intern Med 149:5ll-513, 1989 17. Hull RD, Carter CJ, Jay RM: The diagnosis of acute recurrent deep-vein thrombosis: A diagnostic challenge. Circulation 67:901-906, 1983 18. Heyboer H, Jongbloets LMM, BOiler HR, et al: The clinical utility of real-time compression ultrasonography in the diagnostic management of patients with recurrent venous thrombosis. Acta Radiol Scand 33:297-300, 1992 19. Koopman MMW, Jongbloets LMM, Lensing AWA, et al: Clinical utility of a quantitative B-mode ultrasonography method in patients with suspected recurrent deep vein thrombosis (DVT). Thromb Haemost 69:623, 1993 (abstr) 20. Salzman EW, Hirsh J: Prevention of venous thrombo-
KOOPMAN, VAN BEEK, AND TEN CATE
embolism, in Colman RW, Hirsh J, Marder VJ, Salzman EW (editors): Hemostasis and Thrombosis. Philadelphia, PA, Lippincott, 1987, pp 1252-1265 21. Lensing AWA, BOller HR, Prandoni P, et al: Contrast venography, the gold standard for the diagnosis of deep vein thrombosis: Improvement in observer agreement. Thromb Haemost 67:8-12, 1992 22. Lensing AWA, Prandoni P: Distribution of venous thrombi in symptomatic patients. Thromb Haemost 65: 1176, 1991 (abstr) 23. Hull RD, van Aken WG, Hirsh J, et al: Impedance plethysmography using the occlusive cuff technique in the diagnosis of venous thrombosis. Circulation 53:696-700, 1976 24. MacLachlan MSF, Thomson JG, Taylor DW, et al: Observer variation in the interpretation of lower limbs venograms. Am J Radiol 132:227-229, 1979 25. Bettmann MA, Paulin S: Leg phlebography: The incidence, nature and modification of undesirable side effects. Radiology 122:101-104, 1977 26. Bettmann MA, Robbins A, Braun SD, et al: Contrast venography of the leg: Diagnostic efficacy, tolerance, and complication rates with ionic and nonionic contrast media. Radiology 165:113-116, 1987 27. Katayama H, Yamaguchi K, Kozuka T, et al: Adverse reactions to ionic and nonionic contrast media. A report from the Japanese committee on the safety of contrast media. Radiology 175:621-628, 1990 28. Barrett BJ, Parfrey PS, McDonald JR, et al: Nonionic Iow-osmolality versus ionic high-osmolality contrast material for intravenous use in patients perceived to be at high risk: Randomized trial. Radiology 183:105-110, 1992 29. Moore RD, Steinberg EP, Powe NR, et al: Nephrotoxicity of high-osmolality versus low-osmolality contrast media: Randomized clinical trial. Radiology 182:649-655, 1992 30. Ginsberg JS, Hirsh J, Rainbow AJ, et al: Risks to the fetus of radiologic procedures used in the diagnosis of maternal venous thromboembolic disease. Thromb Haemost 61:189-196, 1989 31. Maternal and Neonatal Haemostasis Working Party of the Haemostasis and Thrombosis Task: Guidelines on the prevention, investigation and management of thrombosis associated with pregnancy. J Clin Pathol 46:489-496, 1993 32. Anderson FA Jr: Impedance plethysmography in the diagnosis of arterial and venous disease. Ann Biomed Eng 12:79-102, 1984 33. Wheeler HB, O'Donnell JA, Anderson FA Jr, et al: Occlusive impedance phlebography: A diagnostic procedure for venous thrombosis and pulmonary embolism. Prog Cardiovasc Dis 17:199-205, 1974 34. Hull RD, Taylor DW, Hirsh J, et at: Impedance plethysmography: The relationship between filling and sensitivity and specificity for proximal vein thrombosis. Circulation 58:898-902, 1978 35. Flanigan DP, Goodreau JJ, Burnham SJ, et al: Vascular laboratory diagnosis of clinically suspected acute deep vein thrombosis. Lancet 2:331-334, 1978 36. Wheeler HB, Anderson FA Jr: Can non-invasive tests be used as the basis for treatment of deep vein thrombosis,
DIAGNOSIS OF DVT
in Bernstein EF (ed): Non-Invasive Diagnostic Techniques in Vascular Disease. St Louis, MO, Mosby, 1981, pp 545-559 37. Peters SHA, Jonker JJC, de Boer AC, et al: Home diagnosis of deep vein thrombosis with impedance plethysmography. Thromb Haemost 48:297-300, 1982 38. Heyboer H, Bfiller HR, Lensing AWA, et al: A randomized comparison of the clinical utility of real-time compression ultrasonography versus impedance plethysmography in the diagnosis of deep vein thrombosis in symptomatic outpatients. N Engl J Med 329:1365-1369, 1993 39. Prandoni P, Lensing AWA, B~ller HR, et al: Failure of computerized impedance plethysmography in the diagnostic management of patients with clinically suspected deep vein thrombosis. Thromb Haemost 65:233-236, 1991 40. Anderson D, Lensing AWA, Wells PS, et al: Limitations of impedance plethysmography in the diagnosis of clinically suspected deep-vein thrombosis. Ann Intern Med 118:25-30, 1993 41. Hull RD, Raskob GE, Carter CJ, et al: Serial impedance plethysmography in pregnant patients with clinically suspected deep vein thrombosis. Ann Intern Med 112:663-667, 1990 42. Huisman MV, Bfiller HR, ten Cate JW: Utility of impedance plethysmography in the diagnosis of recurrent deep vein thrombosis. Arch Intern Med 148:681-684, 1988 43. Paiement G, Wessinger SJ, Waltman AC, et al: Surveillance of deep vein thrombosis in asymptomatic total hip replacement patients. Impedance plethysmography and fibrinogen scanning versus roentgenographic phlebography. Am J Surg 155:400-404, 1988 44. Cruickshank MK, Levine MN, Hirsh J, et al: An evaluation of impedance plethysmography and 125Ifibrinogen leg scanning in patients following hip surgery. Thromb Haemost 62:830-834, 1989 45. Ginsberg JS, Caco CC, Brill-Edwards P, et al: Venous thrombosis in patients who have undergone major hip or knee surgery: Detection with compression ultrasound and impedance plethysmography. Radiology 181:651-654, 1991 46. Harris WH, Athanasoulis C, Waltman AC, et al: Cuff impedance phlebography and |25I fibrinogen leg scanning versus roentgenographic phlebography for diagnosis of thrombophlebitis following hip surgery. J Bone Joint Surg (Am) 58:939-944, 1976 47. Hull RD, Raskob GE, Gent M, et al: Effectiveness of intermittent pneumatic leg compression for preventing deep vein thrombosis after total hip replacement. JAMA 263:2313-2317, 1990 48. Lensing AWA, Prandoni P, Brandies DPM, et al: Detection of deep vein thrombosis by real-time B-mode ultrasonography. N Engl J Med 320:342-345, 1989 49. Effeney D J, Friedman MD, Gooding GAW: Ileofemoral venous thrombosis: Real-time ultrasound diagnosis, normal criteria, and clinical application. Radiology 150:787792, 1984 50. Dauzat MM, Laroche JP, Charras C, et al: Real-time B-mode ultrasonography for better specificity in the noninvasive diagnosis of deep venous thrombosis. J Ultrasound Med 5:625-631, 1986 51. Raghavendra BN, Horii SC, Hilton S, et al: Deep venous thrombosis: Detection by probe compression of veins. J Ultrasound Med 5:89-95, 1986
11
52. Aitken AGF, Godden DJ: Real-time ultrasound diagnosis of deep vein thrombosis: A comparison with venography. Clin Radiol 38:309-312, 1987 53. Appelman PT, de Jong TE, Lampman LE, et al: Deep venous thrombosis of the leg: US findings. Radiology 163:743-746, 1987 54. Cronan J J, Dorfman GS, Scola FH, et al: Deep venous thrombosis: US assessment using vein compressibility. Radiology 162:191-194, 1987 55. Habscheid W, Hohmann M, Wilhelm T, et al: Realtime ultrasound in the diagnosis of acute deep venous thrombosis of the lower extremity. Angiology 41:599-608, 1990 56. Cronan JJ, Leen V: Recurrent deep venous thrombosis: Limitations of US. Radiology 170:739-742, 1989 57. Borris LC, Christiansen HM, Lassen MR, et al: Comparison of real-time B-mode ultrasonography and bilateral ascending phlebography for detection of post-operative deep vein thrombosis following elective hip surgery. Thromb Haemost 61:363-365, 1989 58. Froelich JA, Dorfman GS, Cronan JJ, et al: Compression ultrasonography for the detection of deep venous thrombosis in patients who have a fracture of the hip. J Bone Joint Surg (Am) 71:249-256, 1989 59. Borris LC, Christiansen HM, Lassen MR, et al: Real-time B-mode ultrasonography in the diagnosis of post-operative deep vein thrombosis in non-symptomatic high-risk patients. Eur J gasc Surg 4:473-475, 1990 60. Agnelli G, Volpato R, Radicchia S, et al: Accuracy of real-time B-mode ultrasonography in the diagnosis of asymptomatic deep vein thrombosis in hip surgery patients. Thromb Haemost 65:1175, 1991 (abstr) 61. Sigel B, Felix WR, Popky GL, et al: Diagnosis of lower limb venous thrombosis by Doppler ultrasound technique. Arch Surg 104:174-179, 1972 62. Holmes MCG: Deep venous thrombosis of the lower limbs diagnosed by ultrasound. Med J Aust 4:210-211, 1973 63. Medway J, Nicolaides AN, Walker CJ, et al: Value of Doppler ultrasound in diagnosis of clinically suspected deep vein thrombosis. Br Med J 4:552-554, 1975 64. Dosick SM, Blakemore WS: The role of Doppler ultrasound in acute deep vein thrombosis. Am J Surg 136:265-268, 1978 65. Flanigan DP, Goodreau JJ, Burnham SJ, et al: Vascular laboratory diagnosis of clinically suspected acute deep vein thrombosis. Lancet 2:331-334, 1978 66. Bendick PJ: Pitfalls of the Doppler examination for venous thrombosis. Am J Surg 49:320-323, 1983 67. Zielinsky A, Hull R, Carter C, et al: Doppler ultrasonography in patients with clinically suspected deep vein thrombosis: improved sensitivity by inclusion of posterior tibial vein examination site. Thromb Haemost 50:153, 1983 (abstr) 68. Lensing AWA, Levi MM, B~ller HR, et al: An objective Doppler method for the diagnosis of deep vein thrombosis. Ann Intern Med 113:9-14, 1990 69. Cogo A, Lensing AWA, Prandoni P, et al: Comparison of real-time B-mode ultrasonography and Doppler ultrasound with contrast venography in the diagnosis of venous thrombosis in symptomatic outpatients. Thromb Haemost 70:404-407, 1993
12
70. Polak JF, Culter SS, O'Leary DH: Deep veins of the calf: Assessment with color Doppler flow imaging. Radiology 171:481-485, 1989 71. Rose SC, Zwiebel WJ, Nelson BD, et al: Symptomatic lower extremity deep venous thrombosis: Accuracy, limitations, and role of color duplex flow imaging in diagnosis. Radiology 175:639-644, 1990 72. Schindler JM, Kaiser M, Gerber A, et al: Colour coded duplex sonography in suspected deep vein thrombosis of the leg. Br Med J 301:1369-1370, 1991 73. Baxter GM, McKechnie S, Duffy P: Colour Doppler ultrasound in deep venous thrombosis: A comparison with venography. Clin Radio142:32-36, 1990 74. Mattos MA, Londrey GL, Leutz DW, et al: Colorflow duplex scanning for the surveillance and diagnosis of acute deep venous thrombosis. J Vasc Surg 15:366-376, 1992 75. Sumner DS, Londrey GL, Spadone DP, et al: Study of deep venous thrombosis in high-risk patients using color flow Doppler, in Bergan JJ, Yao JST (eds): Venous Disorders. Philadelphia, PA, Saunders, 1991, pp 63-76 76. Rose SC, Zwiebel WJ, Murdock LE, et al: Insensitivity of color Doppler flow imaging for detection of acute calf deep venous thrombosis in asymptomatic postoperative patients. J Vasc Intervent Radiol 4:111-117, 1993 77. Davidson BL, Elliott CG, Lensing AWA: Low accuracy of color Doppler ultrasound in the detection of proximal leg vein thrombosis in asymptomatic high risk patients. Ann Intern Med 117:735-738, 1992 78. Atkins P, Hawkins LA: Detection of venous thrombosis in the legs. Lancet 1:1217-1219, 1965 79. Nanson EM, Palko PD, Dick AA, et al: Early detection of deep vein thrombosis of the leg using 1311tagged human fibrinogen: A clinical study. Ann Surg 162:438445, 1965 80. Flanc C, Kakkar VV, Clarke MB: Detection of venous thrombosis of the legs using 12SI-labelled flbrinogen. Br J Surg 55:742-747, 1968 81. Kakkar VV: The diagnosis of deep vein thrombosis using the 125I-flbrinogen test. Arch Surg 104:152-159, 1972 82. Kakkar VV: Deep vein thrombosis. Detection and prevention. Circulation 51:8-19, 1975 83. Tsapogas MJ, Goussus H, Peabody RA, et al: Postoperative venous thrombosis and the effectiveness of prophylactic measures. Arch Surg 103:561-567, 1971 84. Myrvold HE, Persson JE, Svensson B, et al: Prevention of thromboembolism with Dextran 70 and heparin in patients with femoral neck fractures. Acta Chir Scand 139:609-616, 1973 85. Sautter RD, Larson DE, Bhattaracharyya SK, et al: The limited utility of fibrinogen 1251 leg scanning. Arch Intern Med 139:148-153, 1979 86. Lensing AWA, Hirsh J: 12sI-fibrinogen leg scanning: Reassessment of its role for the diagnosis of venous thrombosis in post-operative patients. Thromb Haemost 69:2-7, 1993 87. Webber MM, Bennett LR, Cragin M, et al: Thrombophlebitis--Demonstration by scintiscanning. Radiology 96: 620-623, 1969 88. Lisbona R, Stern J, Derbekyan V: 99mTc red blood
KOOPMAN, VAN BEEK, AND TEN CATE
cell venography in deep vein thrombosis of the leg: A correlation with contrast venography. Radiology 143:771773, 1982 89. Ezekowitz MD, Pope CF, Sostman HD, et al: Indium111 platelet scintigraphy for the diagnosis of acute venous thrombosis. Circulation 73:668-674, 1986 90. Tibbutt DA, Chesterman CN, Allington M J, et al: Measurement of fibrinogen-fibrin-related antigen in serum as aid to diagnosis of deep vein thrombosis in outpatients. Br Med J 1:367-369, 1975 91. Yudelman IM, Nossel HL, Kaplan KL, et al: Plasma fibrinopeptide A levels in symptomatic venous thromboembolism. Blood 51:1189-1195, 1978 92. Elms M J, Bunce IK, Bundesen PG, et al: Measurement of crosslinked fibrin degradation products--An immunoassay using monoclonal antibodies. Thromb Haemost 50:591-594, 1983 93. Rowbotham BJ, Carroll P, Whitaker AN, et al: Measurement of crosslinked fibrin derivatives--use in the diagnosis of venous thrombosis. Thromb Haemost 57:59-61, 1987 94. Heaton DC, Billings JD, Hickton CM: Assessment of D-dimer assays for the diagnosis of deep vein thrombosis. J Lab Clin Med 1110:588-591, 1987 95. Ott P, Astrup L, Jensen RH, et al: Assessment of D-dimer in plasma: Diagnostic value in suspected deep venous thrombosis of the leg. Acta Med Scand 224:263-267, 1988 96. Bounameaux H, Schneider PA, Reber G, et al: Measurement of plasma D-dimer for diagnosis of deep venous thrombosis. Am J Clin Pathol 91:82-85, 1989 97. Speiser W, Mallek R, Koppensteiner R, et al: Ddimer and TAT measurements in patients with deep venous thrombosis: Utility in diagnosis and judgement of anticoagulant treatment effectiveness. Thromb Haemost 64:196-201, 1990 98. Boneu B, Bes G, Pelzer H, et al: D-dimers, thrombinantithrombin III complexes and prothrombin fragments 1 + 2: Diagnostic value in clinically suspected deep vein thrombosis. Thromb Haemost 65:28-31, 1991 99. Heyboer H, Ginsberg JS, BOiler HR, et al: The use of the D-dimer test in combination with non-invasive testing versus serial non-invasive testing alone for the diagnosis of deep-vein thrombosis. Thromb Haemost 67:510-513, 1992 100. Jensen C, Lomboldt Knudsen L, Hegediis V: The role of contact thermography in the diagnosis of deep venous thrombosis. Eur J Radiol 3:99-102, 1983 101. Sandier DA, Martin JF: Liquid crystal thermography as a screening test for deep-vein thrombosis. Lancet 1:665-667, 1985 102. Bounameaux H, Khabiri E, Huber O, et al: Value of liquid crystal contact thermography and plasma level of D-dimer for screening of deep venous thrombosis following general abdominal surgery. Thromb Haemost 67:603-606, 1992 103. Thomas PRS, Butler CM, Bowman J, et al: Light reflection rheography: An effective non-invasive technique for screening patients with suspected deep venous thrombosis. Br J Surg 78:207-209, 1991