Chapter 38: Deep Venous Thrombosis and Pulmonary Embolism

Venous thromboembolism (VTE) is a common preventable complication in hospitalized patients. It includes a spectrum of disease, such as deep venous thrombosis (DVT) and pulmonary embolism (PE). General risk factors for DVT were first described by Virchow in the 19th century.¹ These include stasis, vascular wall abnormalities, and a hypercoagulable state; dynamic interaction between these factors leads to thrombosis (Figure 38-1).

The incidence of VTE is thought to be in excess of 600,000 cases per year in the United States.² The incidence of VTE is even higher in critically ill patients. In a prospective ultrasound study, DVT was detected in 33% of patients admitted to a medical ICU.³ VTE carries a significant risk of morbidity and mortality. Most studies have shown that 15% to 20% of patients with PE have cardiovascular or respiratory compromise manifested by shock, syncope, respiratory distress, or hypoxemia.^{4, 5}

Most of the acute neurologic disorders often put patients in periods of prolonged immobilization, thus placing them at the highest risk for VTE. This is particularly true for patients with acute ischemic stroke and spinal cord injury. Studies that used I¹²⁵ fibrinogen as a screening method in patients with acute hemiplegic stroke have shown that the incidence of DVT is approximately 50% in the first 2 weeks in the absence of heparin prophylaxis.⁶ The majority of these deep vein thromboses are asymptomatic and affect a paralyzed leg. The peak incidence of DVT is between day 2 and day 7.⁷ The incidence of PE in patients with ischemic stroke has been estimated to be between 0.8% and 13%.^{8, 9} Mortality rate of untreated PE in unselected hospital patients is up to 30%.¹⁰ One study has suggested that mortality rate of PE is even higher in patients with acute stroke.¹¹ The incidence of VTE in other neurocritical care conditions is listed in Table 38-1.

There are variety of methods available for prophylaxis against VTE. These methods are listed in Table 38-2.

Evidence for the use of physical methods to prevent DVT in patients with acute stroke is limited. A Cochrane review identified only two randomized controlled trials in this area.¹⁶ It failed to show any conclusive benefits of physical methods (both graduated compression stockings [GCSs] and sequential compression devices [SCDs]) for DVT prophylaxis after acute stroke. As literature is scant, optimal timing and duration of the therapy remain unclear. Based on the evidence, we suggest that neither GCSs nor SCDs should be used as the only therapeutic strategy for prevention of VTE in patients with acute stroke.

Pharmacologic methods, mainly unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH), remain the mainstay of therapy for prevention of VTE. One of the major concerns with the use of UFH or LMWH is the risk of intracranial hemorrhage (ICH), including hemorrhagic transformation of infarct and extracranial hemorrhage (ECH). Several studies have assessed UFH, LMWH, and heparinoids in VTE prophylaxis in patients with stroke. A systematic review by Kamphuisen and Agnelli¹⁷ assessing the efficacy and risks of UFH and LMWH included 16 trials with 23,043 patients. They found that, compared to control, the use of high-dose UFH (> 15,000 units/d) was associated with a reduction in PE, but was associated with an increased risk of ICH and ECH. Use of low-dose UFH (≤ 15,000 units/d) was associated with a reduction in the risk of DVT, but not PE, without an increase in ICH or ECH. High-dose LMWH was associated with a reduction in both DVT and PE, but at the expense of an increased risk of ICH and ECH. Low-dose LMWH was associated with a reduction in both DVT and PE without an increase in the risk of ICH or ECH (Table 38-3). For low-dose LMWH, the number needed to treat for DVT was 7 and for PE was 38.

Several studies have compared UFH and LMWH for VTE prophylaxis in stroke patients.^{18, 19, 20} A systematic review by Sandercock et al²¹ comparing LMWH and UFH included 9 trials and 3137 patients. They suggested that compared to UFH, the use of LMWH or heparinoids is associated with a lower incidence of DVT. There was no significant difference in PE, ICH, hemorrhagic transformation of infarct, and death from all causes. However, the numbers of these events were very small. A meta-analysis by Shorr et al²² suggested that the use of LMWH was associated with a reduction in proximal VTE and PE when compared to UFH. The risk of ICH, overall bleeding rate, and mortality rate were similar in both the UFH and the LMWH groups.

In recent years, many newer options have become available for prevention of VTE in general medicine patients.²³ But to date, they have not been tested in specific neurologic conditions, including stroke.

Based on the available data, we recommend the use of LMWH as a primary method of DVT prophylaxis in patients with stroke unless contraindicated.

If acute PE is suspected, a careful assessment of the patient is necessary, including a history, a physical examination, a risk factor assessment, and laboratory/imaging studies (Table 38-4). Clinical features that may suggest DVT are leg pain, warmth, swelling, or redness.²⁴ Clinical features suggestive of acute PE are acute onset of dyspnea, pleuritic chest pain, tachycardia, and hypoxemia. Massive acute PE should be considered if the patient has acute-onset hypotension, severe hypoxemia, or cardiac arrest. The presence of clinical features of both DVT and PE is highly suggestive of the diagnosis, but is not specific.²⁴ Thus, further studies need to be done to make a final diagnosis.

Additional studies should be considered whenever acute PE is suspected. Electrocardiography (ECG) may reveal unexplained tachycardia, which is common in PE, but is nonspecific. An S₁Q₃T₃ pattern, right bundle branch block, or right-axis deviation on ECG might suggest massive PE, but these findings are nonspecific.²⁵ Arterial blood-gas analysis might show hypoxemia and/or respiratory alkalosis, although arterial Po₂ levels might be normal in patients with a PE.

D-Dimer assay can also be helpful to diagnose venous thrombosis and PE, but it is nonspecific. It can also be positive in cancer, trauma, infection, or other inflammatory states.²⁶ It is used mainly in the emergency department to rule out PE in low-risk patients.²⁶ The diagnostic value of the d-dimer assay in neurocritical care is not well established.

Other biomarkers may also provide valuable information in terms of severity of PE. Levels of cardiac troponins are often increased in massive PE and can be used for risk stratification, but they are not sensitive or specific and should not be used as the sole diagnostic tool.²⁷ Elevated levels of B-natriuretic peptide (BNP) may be seen in massive PE in response to ventricular stretching,²⁸ but these, too, are not sensitive or specific enough to be used as a diagnostic tool.

Chest radiography is routinely obtained in patients with acute shortness of breath and is typically normal in patients with PE.

Lately, multidetector CT angiography (MDCTA) of the chest has become the initial test of choice for diagnosis of acute PE (Figure 38-2). Sensitivity of helical CT angiography in the diagnosis of PE has been reported to be from 57% to 100%, and its specificity has been reported from 78% to 100%.^{29, 30} A recent study using MDCTA performed in outpatients with suspected PE has reported its sensitivity as 83% and specificity as 96%.³¹ Sensitivity and specificity of MDCTA also depend on pretest clinical probability and location of the embolus. Patients with a high clinical suspicion of PE and negative MDCTA will require further testing to rule out PE. Advantages of MDCTA over ventilation/perfusion V̇/Q̇ scanning or conventional angiography include ease of performance, speed, and characterization of the right ventricle and pulmonary parenchyma. Given its advantages and high sensitivity and specificity, MDCTA has become the gold standard to diagnose or rule out PE. MDCTA does require administration of intravenous contrast, which can lead to nephrotoxicity in patients with elevated creatinine and needs to be kept in mind. There are no randomized controlled trials conducted in specific neurologic conditions to evaluate the efficacy of MDCTA in the diagnosis of PE. But we do believe that findings from studies done in other patient populations can be extrapolated to neurologic patients.

Ventilation/perfusion scanning (V̇/Q̇ scan) had a major role in the diagnosis of PE a few years ago, but it has been largely replaced by MDCTA in recent years. It has good sensitivity and specificity. A normal V̇/Q̇ scan essentially rules out PE.³² But there are some logistical issues with V̇/Q̇ scanning: patients are required to have a normal chest radiograph, and most of the patients in the ICU have an abnormal chest radiograph, thus limiting the value of the V̇/Q̇ scan as a diagnostic tool for PE in critical care setting.

Duplex ultrasonography of the lower extremity is often done, as most of the pulmonary emboli originate from lower extremities. It is positive in 10% to 20% of patients without any clinical features of DVT, and in 50% of patients with clinical features of DVT.³³ Thus, a negative duplex ultrasonography of lower extremities with high clinical suspicion does not rule out PE.³³

Pulmonary angiography is the gold standard for diagnosis of PE, but it is an invasive procedure. Because the sensitivity and specificity of MDCTA have been similar to those of pulmonary angiography, MDCTA has largely replaced pulmonary angiography as a diagnostic tool for PE.

Echocardiography is a useful tool for risk stratification of massive and submassive PE. Some recent studies have suggested that right ventricular strain on echocardiography in normotensive patients puts these patients at higher risk of death. Signs of right ventricular strain on echocardiography include right ventricle dilation and right ventricle hypokinesis (Figure 38-3). Administration of thrombolytics has been suggested in this subgroup of patients but is debatable.^{34, 35} Echocardiography requires some training and can be done by intensivists.

Figure 38-4 shows general approach to a patient with suspected PE with CT angiography as the initial test.

Anticoagulation is the main therapy for a patient with acute PE. This must be considered on a case-by-case basis, taking into account the risk of intracranial bleed.

Points to consider prior to initiating treatment include:

• Whether anticoagulation should be initiated

• Which anticoagulant and what dose should be used based on clinical evidence supporting its use

• Common complications

• Duration of therapy

Other aspects of treatment to consider include placement of an inferior vena cava filter, use of thrombolytics, and performance of an embolectomy. The efficacy of anticoagulant therapy depends on achieving a therapeutic level within the first 24 hours of therapy.^{36, 37, 38}

Available therapeutic options include the following:³⁹

• Subcutaneous (SC) LMWH

• Intravenous (IV) UFH

• SC UFH

• SC fondaparinux

Special Considerations in Anticoagulation in the Neurologic Critical Care Patient Population

Patients with acute or prior intracerebral hemorrhage

The availability of data or expert consensus regarding optimal treatment for VTE is poor in this patient population. In a review, the estimated risk of fatal PE in the acute ICH population with untreated DVT or nonfatal PE was about 25%.⁴⁰ This risk is too high to deny treatment for VTE to patients with good neurologic prognosis.

The risk of worsening ICH with or without anticoagulation is unknown. The estimated risk of recurrent ICH with anticoagulation is estimated to be 3% to 5%.⁴¹ In the absence of anticoagulation the rate of ICH is 0.4% to 3%.⁴²

Based on the above data and given the known efficacy of heparin and LMWH in the nonstroke patient population, patients with proximal DVT and nonfatal PE should be considered for treatment with heparin and LMWH. Subsequent treatment with a reduced dose of oral anticoagulation warfarin after 1 week with a target international normalized ratio (INR) of 2.0 (range of 2-2.5) for 3 months is recommended.^{43, 44} Aggressive management of blood pressure is critical to reduce the risk of subsequent bleeding.

In patients with recent ICH with nonmassive PE who have adequate cardiopulmonary reserve and negative ultrasonography for proximal DVT, the 3-month risk of recurrent PE is very low. In this population anticoagulation may be withheld.⁴⁵

Patients with an unruptured intracranial aneurysm

Unruptured intracranial aneurysms are detected in approximately 3% of adults who undergo intra-cranial arterial imaging. Subarachnoid hemorrhage from a ruptured aneurysm is associated with 40% increase in short-term mortality and 50% increase in permanent neurologic injury. Rates of rupture vary according to size and location (posterior communicating artery); risk factors include age older than 60 years, female gender, and tobacco smoking.⁴⁶

There are no randomized clinical trials reporting data supporting higher rupture rates with the use of anticoagulation. The available clinical trials in patients with ICH so infrequently include patients with subarachnoid hemorrhage; therefore, it is difficult to exclude higher rates of hemorrhage with anticoagulation. Therefore, the decision must be individualized according to the best clinical judgment of benefits versus risks.

Patients with brain tumors

Because of a hypercoagulable state, patients with brain tumors have a substantially high risk of developing VTE. There is concern, however, that antithrombotic agents can precipitate hemorrhage into the tumor site. The risk factors for developing VTE include age older than 60 years, large tumor size, neurosurgery within 2 months, leg paresis, and treatment with chemotherapy.^{47, 48}

Recommendations

1. Anticoagulation should be used in all patients except those who have a high rate of ICH, such as patients with metastatic renal cell carcinoma, melanoma, and choriocarcinoma. Given the high rate of recurrent VTE, patients with malignant glioma should be anticoagulated long-term.

2. For patients with brain tumors who have a high risk of hemorrhage as above, place an inferior vena cava filter if there is residual metastatic disease.

Inferior Vena Cava Filters

Inferior vena cava (IVC) filters are placed percutaneously via the jugular or femoral approach under fluoroscopy or ultrasound guidance. The filter is typically placed in the infrarenal position. In patients who demonstrate renal vein thrombosis, an IVC clot above the renal vein will require the filter to be placed in the suprarenal position.

Filter placement protects the pulmonary vascular bed but does nothing to decrease the predisposition to thromboembolism or lessen thrombosis. Small thrombi can pass through patent filters or through collaterals around obstructed vessels. Two types of filters are currently available: retrievable and nonretrievable filters.

The only validated indications for IVC filter placement are:

• An absolute contraindication to anticoagulation

• Failure of anticoagulation therapy (ie, recurrent thromboembolism despite therapeutic levels of anticoagulation)

• Venous thromboembolism in a patient with a high risk of bleeding

Complications of IVC filter placement include:

• Filter erosion through the IVC wall

• Filter embolization/migration⁴⁹

• IVC obstruction by a subsequent clot⁵⁰

• Local complications such as hematoma related to time of insertion

• Insertion site thrombosis

• Venous insufficiency

Prophylactic IVC filter placement may increase the relative risk of deep venous thrombosis after acute spinal cord injury.⁵¹