Table 105-1 Model for Determining Clinical Suspicion of Deep Vein Thrombosis ||Download (.pdf)
Table 105-1 Model for Determining Clinical Suspicion of Deep Vein Thrombosis
Active cancer (treatment ongoing or within previous 6 mo or palliative)
Paralysis, paresis, or recent plaster immobilization of the lower extremities
Recently bedridden for more than 3 d, or major surgery within the past 4 wk
Localized tenderness along the distribution of the deep venous system
Entire leg swollen
Affected calf 3 cm greater than asymptomatic calf (measured 10 cm below tibial tuberosity)
Pitting edema confined to the symptomatic leg
Dilated superficial veins (nonvaricose)
Alternative diagnosis is at least as likely as that of deep vein thrombosis
Table 105-2 Test Results That Effectively Confirm or Exclude Deep Vein Thrombosis ||Download (.pdf)
Table 105-2 Test Results That Effectively Confirm or Exclude Deep Vein Thrombosis
Diagnostic for first DVT
Intraluminal filling defect
Noncompressible proximal veins at two or more of the common femoral, popliteal, and calf trifurcation sites*
Excludes first DVT
All deep veins seen, and no intraluminal filling defects
Negative result on a test that has at least a moderately high sensitivity (≥85%) and specificity (≥70%) and (1) normal results on venous ultrasonography of the proximal veins or (2) low clinical suspicion of DVT at presentation
Negative result on a test that has a high sensitivity (≥98%)
Normal proximal veins and (1) low clinical suspicion for DVT at presentation, or (2) normal D-dimer test at presentation, or (3) normal second test after 7 d
Normal proximal and distal veins*
Diagnostic for recurrent DVT
Intraluminal filling defect
(1) A new noncompressible common femoral or popliteal vein segment or (2) a ≥ 4.0 mm increase in diameter of the common or popliteal vein since a previous test†
Excludes recurrent DVT
All deep veins seen and no intraluminal filling defects
Normal or ≤ 1 mm increase in diameter of the common femoral or popliteal veins on venous ultrasound since a previous test and continuing normal results (no progression of venous ultrasound) at 2 and 7 d
Results as described as for a first episode of DVT; however, these criteria are less well evaluated for diagnosis of recurrence
The clinical features of DVT include localized swelling, redness, tenderness, and distal edema. As these symptoms are nonspecific, the diagnosis should always be confirmed by objective investigations. However, clinical assessment does allow division of patients into low, moderate, and high probabilities of DVT, corresponding to prevalence of 15%, 25%, and 60%, respectively. Clinical prediction rules, such as the Well's Score, are based on four factors; (1) the presence or absence of risk factors (e.g., recent immobilization, hospitalization within the past month, or malignancy), (2) symptoms and signs at presentation are considered typical or atypical, and their severity, (3) severity of symptoms and signs, and (4) whether there is an alternative explanation for the symptoms and signs considered at least as likely as DVT (Table 105-1). The conditions that are most likely to simulate DVT are ruptured Baker cyst, cellulitis, muscle tear, muscle cramp, muscle hematoma, external venous compression, superficial thrombophlebitis, and the postthrombotic syndrome. The prevalence of VTE is higher in the elderly who are investigated, compared to younger patients.
Four objective tests—venography, impedance plethysmography, and venous ultrasonography, and D-dimer testing—have been rigorously evaluated for the diagnosis of DVT. Impedence plethysmography is now used infrequently, and will not be considered in the following review. Magnetic resonance venography and computed tomography (CT) venography appear to be promising new modalities but are less well evaluated.
Venography provides the reference standard for diagnosis of DVT. It involves the injection of a radiocontrast agent into a distal vein. Venography detects both proximal vein thrombosis and calf vein thrombosis. However, it is technically difficult, expensive, requires injection of contrast dye and can be painful. Since contrast dye can cause allergic reactions or exacerbate renal impairment, venography is usually reserved to resolve discrepancies between findings on venous ultrasonography and clinical assessment of probability of DVT, or when venous ultrasonography is nondiagnostic (often in patients with previous DVT). The increased prevalence of renal impairment makes venography an even less attractive investigation option in the elderly.
Venous ultrasonography is the noninvasive imaging method of choice for diagnosing DVT. It is not painful and it is easier to perform than venography. The common femoral vein, femoral vein, popliteal vein, and calf vein trifurcation (i.e., very proximal deep calf veins) are imaged in real time and compressed with the transducer probe. Inability to fully compress or obliterate the vein is diagnostic of DVT. Duplex ultrasonography, which combines real-time imaging with pulsed Doppler and color-coded Doppler technology, facilitates imaging of the deep veins of the calf.
Venous ultrasonography is highly accurate for the detection of proximal vein thrombosis in symptomatic patients, with reported sensitivity and specificity approaching 95%. The sensitivity for symptomatic calf vein thrombosis is considerably lower and appears to be operator dependent. For this reason, many centers do not examine the deep veins of the calf with ultrasonography. Instead, if the initial test excludes proximal DVT, anticoagulants are withheld and the test is repeated in 7 days to exclude progression of a calf vein thrombosis not identified at the initial presentation. If the test remains negative after 7 days, the risk that thrombus is present and will subsequently extend to the proximal veins is negligible, and it is safe to continue withholding treatment.
While ultrasonography is an accurate test, if the results are inconsistent with the pretest probability, further investigations maybe warranted. For example, if the pretest clinical suspicion for DVT is low and the ultrasound shows a localized abnormality (i.e., less convincing findings), or if clinical suspicion is high and the ultrasound is normal, venography should be considered. In about one quarter of such cases, the results of venography differ from those of the ultrasound. Because the prevalence of DVT is only about 2% (most of which is distal), a follow-up test is not necessary when the clinical suspicion of thrombosis is low and the result of an initial proximal venous ultrasound is normal.
D-dimer is formed when cross-linked fibrin in thrombi is broken down by plasmin; thus, elevated levels of D-dimer can be used to detect DVT and PE. A variety of D-dimer assays are available, and they vary markedly in their accuracy as diagnostic tests for VTE.
All D-dimer assays have a low specificity for DVT and, therefore, an abnormal result is associated with a low positive predictive value and cannot be used to diagnose DVT. D-dimer assays that are used for diagnosis of VTE can be divided into two groups based on their sensitivity and specificity. Very highly sensitive D-dimer assays (e.g., sensitivity ≥ 98%; specificity ~40%) have a sufficiently high negative predictive value (≥98%) that a normal result can be used to exclude VTE without the need to perform additional diagnostic testing. Moderate to highly sensitive D-dimer assays (sensitivity 85% to 97%; specificity 50% to 70%) need to have a negative result combined with another assessment that identifies patients as having a lower prevalence of VTE in order to exclude DVT or PE. Management studies have shown that it is safe to withhold anticoagulant therapy in patients who have a normal result on a moderately sensitive D-dimer test in combination with (1) a low clinical suspicion for DVT or (2) a normal result on venous ultrasonography of the proximal veins. Baseline D-dimer levels are known to increase with age. Therefore, D-dimer testing are much less specific and, therefore, of less clinical utility (fewer negative tests among those without venous thrombosis) in the elderly. Also, D-dimer testing has less clinical utility among patients with a high clinical suspicion of VTE as negative results are rarely obtained, and the predictive value of a negative test is lower in this group, because of a higher prevalence of disease.
Recurrent Deep Vein Thrombosis
The diagnosis of acute recurrent DVT can be difficult. A negative D-dimer test can exclude recurrent DVT, although the safety of this approach has been less well evaluated than for first episodes of DVT, and D-dimer test is less often negative compared with patients without a history of venous thrombosis. If D-dimer testing is positive, or has not been performed, venous ultrasonography is performed. If the result is normal, the test should be repeated twice over the next 7 to 10 days. If the result is positive in the popliteal or common femoral vein segments, and the result of the previous test was negative at the same site, a recurrence is diagnosed. This diagnosis can also be made if venous ultrasonography shows other convincing evidence of more extensive thrombosis than was seen on a previous examination (e.g., an increase in thrombus diameter of ≥4 mm at the inguinal ligament or the mid-popliteal fossa; unequivocal extension within the femoral vein of the thigh). If findings on venous ultrasonography are equivocal, as compared with a previous scan, or a previous scan is not available for comparison, venography should be performed or the test can be repeated twice over the next 7 to 10 days to detect extension of thrombosis. If the venogram shows a new intraluminal filling defect or evidence of thrombus extension since a previous venogram, recurrent DVT is diagnosed. If the venogram outlines all of the deep veins and does not show an intraluminal filling defect, recurrent DVT is excluded. If the venogram is nondiagnostic (i.e., nonfilling of segments of the deep veins), the patient can be followed with repeat venous ultrasonography (as described above) or recurrent DVT can be diagnosed based on the results of all assessments, including clinical features.
Table 105-3 Model for Determining a Clinical Suspicion of Pulmonary Embolism ||Download (.pdf)
Table 105-3 Model for Determining a Clinical Suspicion of Pulmonary Embolism
Clinical signs and symptoms of deep vein thrombosis (minimum leg swelling and pain with palpation of the deep veins)
An alternative diagnosis is less likely than pulmonary embolism
Heart rate > 100 beats/min
Immobilization or surgery in the previous 4 wk
Previous deep vein thrombosis/pulmonary embolism
Malignancy (treatment ongoing or within previous 6 mo or palliative)
Table 105-4 Test Results that Effectively Confirm or Exclude Pulmonary Embolism ||Download (.pdf)
Table 105-4 Test Results that Effectively Confirm or Exclude Pulmonary Embolism
Diagnostic for PE
Intraluminal filling defect
CT pulmonary angiography (CTPA)
Intraluminal filling defect in a main or lobar or segmental pulmonary artery and moderate/high clinical probability
High-probability scan and moderate/high clinical probability
Tests for DVT*
Evidence of acute DVT with nondiagnostic ventilation-perfusion scan or CTPA
CT pulmonary angiography
Lung perfusion scan
High-sensitivity D-dimer test‡
Moderate-sensitivity D-dimer test§
Negative, plus (1) low clinical suspicion of PE or (2) normal alveolar dead space fraction or (3) nondiagnostic lung scan and negative ultrasonography of proximal leg veins
Combination of clinical assessment, ventilation-perfusion scan, ultrasonography of proximal leg vein
Low clinical suspicion, nondiagnostic scan, and negative ultrasonography
Dyspnea is the most common symptom of PE. Chest pain is also common; it is usually pleuritic but can be substernal and compressive. Tachycardia is relatively common and hemoptysis is less frequent. Although most patients with PE also have DVT, fewer than 25% have associated clinical features. However, the clinical features of PE, like those of DVT, are nonspecific, and in only about one quarter of patients suspected with PE is the diagnosis confirmed by objective tests. Furthermore, elderly patients are more likely to present with atypical symptoms and signs such as fatigue, dizziness, and syncope.
In the past, clinical assessment of the probability of PE was not standardized; physicians made the assessment informally on the basis of their experience and the results of initial routine tests (e.g., chest x-ray and electrocardiogram). Two groups have published explicit criteria for determining the clinical probability of PE. The model created by Wells and colleagues incorporates an assessment of symptoms and signs, the presence of an alternative diagnosis to account for the patient's condition, and the presence of risk factors for VTE. With this model, a patient's clinical probability of PE can be categorized as low or unlikely (prevalence of PE <10%), moderate (prevalence ~25%), or high (prevalence of 60%) (Table 105-2).
Chest Radiography and Electrocardiography
In patients with PE, chest x-rays show either normal or nonspecific findings. Chest radiography, however, is useful for exclusion of pneumothorax and other conditions that can simulate PE. The electrocardiogram also frequently shows normal or nonspecific findings, but it is valuable for excluding acute cardiac conditions (e.g., myocardial infarction, acute pericarditis). In the appropriate clinical setting, electrocardiographic (ECG) evidence of right ventricular strain suggests PE.
Pulmonary angiography is considered the reference standard for PE, however, it is now rarely performed as it is invasive and can usually be replaced by computed tomographic pulmonary angiography (CTPA). Pulmonary angiography can be complicated by arrhythmias, cardiac perforation, cardiac arrest, and hypersensitivity to the contrast medium. Complications occur in 3% to 4% of patients undergoing pulmonary angiography.
Ventilation–Perfusion Lung Scanning
In the past, ventilation–perfusion lung scanning was the most important test for diagnosing PE. More recently, CTPA has supplanted lung scanning, although lung scanning is still used, particularly when CTPA is contraindicated because of renal failure or associated radiation exposure to the chest (e.g., in young women). A normal perfusion scan excludes a diagnosis of PE, however, a normal result is only obtained in about 25% of consecutive patients with suspected PE and an even smaller proportion of elderly patients. An abnormal perfusion scan is nonspecific. Ventilation imaging improves the specificity of perfusion scanning for the diagnosis of PE; when the ventilation scan is normal at the site of a segmental or larger perfusion defect, the prevalence of PE is 85% or higher (termed a “high probability” lung scan, which justifies anticoagulant therapy). About half of patients who have PE have a “high probability” lung scan. Therefore, among consecutive patients who are investigated for PE, about 25% have a normal perfusion scan and can have the diagnosis excluded, about 15% have a “high probability scan” and can (provided clinical probability is not low) have PE diagnosed, and about 60% have an abnormal but nondiagnostic lung scan that requires further diagnostic testing.
Computed Tomographic Pulmonary Angiography
CTPA, performed using helical CT (also known as spiral or continuous volume CT), is able to directly visualize the pulmonary arteries. Helical CT technology has rapidly advanced from use of single detector scanners to use of progressively larger numbers of detectors (termed multidetector CT) and enables more detailed examination of the pulmonary arteries.
Current evidence from the PIOPED II study suggests that CTPA is nondiagnostic in 6% of patients, and that among adequate examinations, sensitivity is 83%, specificity is 96%, positive predictive value is 86%, and negative predictive value is 95%. Accuracy varies according to the size of the largest pulmonary artery involved: positive predictive value was 97% for abnormalities in the main or lobar artery, 68% for those in segmental arteries, and 25% for subsegmental abnormalities (4% of pulmonary emboli in this study). Predictive values were also influenced by clinical assessment of PE probability; positive predictive value was 96% with high, 92% with intermediate, and 58% with low clinical probability (8% of patients); negative predictive value was 96% with low, 89% with intermediate, and 60% with high clinical probability (3% of patients).
The ability of CTPA to exclude PE has also been evaluated in management studies in which anticoagulant therapy was withheld in patients with negative CTPA. More recent studies suggest that less than 2% of patients with a negative CTPA for PE will return with symptomatic VTE during follow-up.
Magnetic resonance imaging (MRI) is less well evaluated than helical CT for the diagnosis of PE and is expected to be less accurate. Both helical CT and MRI have the advantage of being able to identify alternative pulmonary diagnoses. MRI does not expose the patient to radiation. Both MRI and helical CT can be extended to look for concomitant DVT.
As was discussed for evaluation of suspected DVT, D-dimer testing is also a valuable test for the exclusion of PE, either used alone (very sensitive D-dimer assay) or in combination with other assessments that are associated with a reduced prevalence of PE (e.g., low clinical probability for PE; nondiagnostic ventilation–perfusion scan in combination with a negative ultrasound of proximal lower limb veins [see below]).
Compression ultrasonography, usually evaluating the proximal deep veins of the legs, can aid in the diagnosis of PE. Demonstration of DVT, which occurs in about 5% of patients with nondiagnostic ventilation–perfusion lung scans, can serve as indirect evidence of PE. Exclusion of proximal DVT does not rule out PE in a patient with a nondiagnostic ventilation–perfusion scan, although it does reduce that probability somewhat. However, if there are no proximal DVT on the day of presentation, and if no proximal DVT are detected on two subsequent examination 1 and 2 weeks later (DVT is diagnosed during serial testing in ~2% of patients), anticoagulant therapy can be withheld with a very low risk that patients will return with VTE (less than 2% during 3 months of follow-up). Earlier studies that evaluated CTPA suggested that a negative result did not exclude PE and, therefore, should be followed by bilateral ultrasonography of the proximal veins. However, more recent studies of CTPA, which used mostly multidector scanners, do not support the need for routine ultrasonography of the proximal deep veins in patients with a negative CTPA. Instead, it appears to be reasonable to only perform ultrasonography of the proximal deep veins in patients with a negative CTPA if clinical suspicion for PE is high. As for patients who have nondiagnostic ventilation–perfusion lung scans, withholding of anticoagulant therapy and performance of serial ultrasonography is a reasonable approach to management of patients who have a CTPA that is suspicious for isolated subsegmental PE.
Uncommon Thromboembolic Disorders
Subclavian or Axillary Veins
Thrombosis of the subclavian or axillary veins may be idiopathic or may occur as a complication of local vascular damage. It is now most frequently seen as a complication of chronic indwelling catheter use, but it also occurs as a complication after mastectomy and local radiotherapy for breast cancer. Idiopathic subclavian or axillary vein thrombosis may occur in young muscular individuals and maybe preceded by repetitive, strenuous activity involving the affected arm. Some of these persons have a fixed stenosis of the subclavian vein that is thought to be caused by compression of the vein between the first rib and the clavicle. Thrombosis of the subclavian vein or the superior vena cava is a rare complication of a transvenous cardiac pacemaker.
Subclavian or axillary thrombosis causes pain, edema, and cyanosis of the arm. In rare cases, the thrombosis extends into the superior vena cava and causes edema and cyanosis of the face and neck.
Definitive diagnosis is made by venography, venous ultrasonography, or CT angiography. Subclavian or axillary vein thrombosis is treated with anticoagulants using a similar approach for lower limb DVT and PE (see “Treatment of Venous Thromboembolism”). Regional or systemic thrombolytic therapy is usually reserved for select young patients without contraindications.
An uncommon disorder, mesenteric vein thrombosis usually occurs in the sixth or seventh decade of life. It generally involves segments of the small bowel, leading to hemorrhagic infarction. Affected patients often have associated disorders, such as inflammatory bowel disease, malignancy, portal hypertension, familial thrombophilia, or polycythemia vera, or they may have a history of recent abdominal surgery. In about 20% of cases, no underlying cause is found.
The clinical manifestations of mesenteric vein thrombosis include intermittent abdominal pain, abdominal distention, vomiting, diarrhea, and melena. Blunt, semiopaque indentations of the bowel lumen (“thumbprinting”) caused by mucosal edema, or gas in the wall of the bowel or the portal vein, or free peritoneal air may occur secondary to bowel infarction. CT, which shows an intraluminal filling defect in the mesenteric vein, is the diagnostic test of choice, and both Doppler ultrasonography and MRI are also helpful. Management includes acute and long-term anticoagulation, supportive care and surgery if bowel resection is being considered, followed by anticoagulant therapy. Mortality is about 30%, and up to 30% of patients experience recurrence.
Renal vein thrombosis can be idiopathic or a complication of the nephrotic syndrome. Patients maybe asymptomatic or may present with abdominal, back, or flank pain and tenderness. PE is a relatively common complication of renal vein thrombosis. Anticoagulant therapy results in a gradual improvement in renal function, but patients may have long-standing proteinuria. Thrombolytic agents have been used, but the data are inadequate for critical appraisal of this form of treatment.