EPIDEMIOLOGY AND USUAL CAUSES OF VENOUS THROMBOSIS
Deep venous thrombosis (DVT) has been estimated to affect more than 250,000 patients per year (
1,
2). It has also been estimated that DVT and pulmonary embolism (PE) together are responsible for 300,000 to 600,000 hospitalizations and as many as 50,000 deaths per year; other estimates suggest an even higher yearly death rate (
3). DVT is responsible for a 21% yearly rate of mortality in the elderly, and the cost of treatment for venous thromboembolism has been estimated to be between $1.0 billion and $2.5 billion per year. Thus venous thromboembolism remains a significant problem today. Usual risk factors associated with DVT include older age, malignancy, obesity, varicose veins, prior DVT, surgery, vascular injury, immobility, oral contraceptive use, heart failure, and various hypercoagulable states (
Table 32.1).
HELPFUL TESTS FOR DIAGNOSIS OF DEEP VENOUS THROMBOSIS AND PULMONARY EMBOLISM
Tests for the diagnosis of DVT involve indirect tests of historic interest and morecurrent tests that visualize thrombus. Venous duplex ultrasound imaging is now the standard for DVT diagnosis and has virtually replaced contrast phlebography.
Duplex ultrasound imaging includes analysis of both image and flow. Acute thrombosis is diagnosed from noncompressibility of the vein, vein enlargement, and the lack of collateral veins. Chronic thrombosis is indicated by increased echogenicity of thrombi, a small and shrunken vein, and the preference of collateral vessels. Sensitivity, specificity, positive predictive value, and negative predictor value for the diagnosis of acute DVT with color-flow duplex imaging in symptomatic patients are greater than 95% (
4). Even for calf vein thrombi, the sensitivity in symptomatic patients is greater than 90%, although the sensitivity in the below-knee position may be much lower in asymptomatic patients being screened.
The excellent specificity of venous duplex imaging allows therapeutic decisions. Withholding anticoagulation on the basis of a negative scan is safe and reasonable. In a study of 431 negative duplex scans and 66 corresponding phlebograms, only three peroneal thrombi were found on phlebography, whereas more-proximal thrombi were not missed (
5). Follow-up over an 8-month period revealed no PE and no recurrent DVT. Some clinicians have also combined clinical characteristics with duplex ultrasound imaging in an attempt to improve on the results of imaging (
6).
Thus venous duplex imaging is now the “gold standard” for the diagnosis of DVT and has replaced contrast phlebography. It is safe, painless, and accurate; requires no contrast material; and can be performed during pregnancy. It is noninvasive and repeatable, it can follow the progression or resolution of DVT, and it detects other abnormalities, such as pseudoaneurysms, venous aneurysms, Baker cysts, superficial thrombophlebitis, and cellulitis. The incidence of positive studies in a busy vascular laboratory should be approximately 30%.
Magnetic resonance venography (MRV) has demonstrated promise as a diagnostic modality for both DVT and PE. The sensitivity and specificity are 100% and 96% both for DVT (
7) and for PE (
8). MRV with gadolinium has been found to define thrombus age. During acute DVT, an inflammatory response is found in the vein wall and perivenous tissue, and gadolinium extravasates into the inflammation (
9). As the DVT organizes and matures, gadolinium enhancement fades as the vein shrinks. In many locations, the inaccessibility of the magnetic resonance imaging machines and the cost limit the use of MRV for DVT diagnosis.
The diagnosis of PE involves ventilation-perfusion (V/Q) scanning or pulmonary angiography; newer techniques include spiral computed tomographic scanning and magnetic resonance imaging. The sensitivity of V/Q scanning is excellent, at 98%, but specificity is low, at 10% (
10). However, by combining clinical risk factors with the V/Q scan, sensitivity and specificity greater than 95% have been reported. With a high-probability V/Q scan and two risk factors for PE, the sensitivity for PE diagnosis was 97%; with one risk factor, 84%; and with no risk factors, 82%. Similarly, with a normal V/Q scan, the chance of PE was 0, no matter what the risk factor status (
11). These results suggest that a normal V/Q scan or a high-probability scan provide good diagnostic information. However, only approximately one third of V/Q scans are in one of these two categories, and so the majority of patients need further testing. Such further testing includes lower extremity venous duplex ultrasound imaging (venous duplex imaging is positive in approximately 10% of cases in these patients) and, more important, spiral CT scanning. Indications for pulmonary arteriography include acute massive PE, inferior vena cava (IVC) interruption, and the planning of pulmonary interventional therapy, such as thrombolysis or pulmonary embolectomy.
Spiral computed tomographic scanning, a relatively new technique for PE diagnosis, has excellent specificity but relatively low sensitivity (50% to 65%), despite promising initial results. However, as the technology has improved, the sensitivity and specificity have also improved, and now emboli at the subsegmental level can be identified (
12). A recent study (PIOPED II) has now been completed. The sensitivity for isolated chest CT imaging was 83%, but increased to more than 90% when clinical analysis was added. Additionally, sensitivity improved when adding a leg study (either CT or ultrasound) to the chest CT scan (level 1 evidence) (
13). Magnetic resonance imaging has demonstrated excellent promise for PE diagnosis and is currently being studied in PIOPED III.
The use of D-dimer assays has been investigated in the diagnosis of both DVT and PE, and sensitivity of 96% to 98% has been reported (
14). However, specificity of only 40% to 50% has been found. It is likely that D-dimer testing will supplement other tests such as venous duplex ultrasound imaging and clinical assessment (
15).