Pulmonary embolism affects an estimated 117 people per 100,000 person years, resulting in about 350,000 cases yearly (probably at least 100,000 in the US), and causes up to 85,000 deaths/yr. It affects mainly adults.
Nearly all pulmonary emboli arise from thrombi in the veins of the legs or pelvis (deep venous thrombosis). Risk of embolization is higher with thrombi proximal to the calf veins. Thromboemboli can also originate in arm veins or central veins of the chest (caused by central venous catheters or resulting from thoracic outlet syndromes).
Risk factors for deep venous thrombosis and pulmonary embolism (see table Risk Factors for Deep Venous Thrombosis and Pulmonary Embolism) are similar in children and adults and include
Risk Factors for Deep Venous Thrombosis and Pulmonary Embolism
Once deep venous thrombosis develops, clots may dislodge and travel through the venous system and the right side of the heart to lodge in the pulmonary arteries, where they partially or completely occlude one or more vessels. The consequences depend on the size and number of emboli, the underlying condition of the lungs, how well the right ventricle (RV) is functioning, and the ability of the body’s intrinsic thrombolytic system to dissolve the clots. Death occurs due to right ventricular failure.
Small emboli may have no acute physiologic effects and may begin to lyse immediately and resolve within hours or days. Larger emboli can cause a reflex increase in ventilation (tachypnea), hypoxemia due to ventilation/perfusion (V/Q) mismatch and low mixed venous oxygen content as a result of low cardiac output, atelectasis due to alveolar hypocapnia and abnormalities in surfactant, and an increase in pulmonary vascular resistance caused by mechanical obstruction and vasoconstriction resulting in tachycardia and hypotension. Endogenous lysis reduces most emboli, even those of moderate size, and physiologic alterations decrease over hours or days. Some emboli resist lysis and may organize and persist.
PE may be designated according to the physiologic effects as
Catastrophic or super-massive (high risk): Impaired right ventricular function with severe hypotension/hypoxemia that requires aggressive pressor therapy and high-flow oxygen
Massive (also high risk): Impaired right ventricular function with hypotension, as defined by systolic BP < 90 mm Hg or a drop in systolic BP of ≥ 40 mm Hg from baseline for a period of 15 min
Submassive (intermediate risk): Impaired right ventricular function and/or abnormal troponin and/or BNP level without hypotension. Note that European Society of Cardiology defines intermediate-risk pulmonary embolism as patients with a simplified pulmonary embolism severity index (sPESI) of > 0, thus including patients with other disorders or findings.
Low risk: Absence of right ventricular impairment and absence of hypotension (and by European Society of Cardiology, sPESI score = 0)
Saddle PE describes a pulmonary embolus that lodges in the bifurcation of the main pulmonary artery and into the right and left pulmonary arteries; saddle PEs are usually, but not always, intermediate or high-risk.
In 1 to 3% of cases, chronic residual obstruction leads to pulmonary hypertension (chronic thromboembolic pulmonary hypertension) that evolves over months to years and can result in chronic right heart failure.
When large emboli occlude major pulmonary arteries, or when many small emboli occlude > 50% of the more distal vessels, RV pressure increases, which may lead to acute RV failure, shock, or sudden death. The risk of death depends on the degree and rate of rise of right-sided pressures and on the patient’s underlying cardiopulmonary status. Patients with preexisting cardiopulmonary disease are at higher risk of death, but young and/or otherwise healthy patients may survive a PE that occludes > 50% of the pulmonary bed.
Pulmonary infarction (interruption of pulmonary artery blood flow leading to ischemia of lung tissue , sometimes represented by a pleural-based [peripherally located], often wedge-shaped, pattern on chest x-ray [Hampton hump] or other imaging modalities) occurs in < 10% of patients diagnosed with PE. This low rate has been attributed to the dual blood supply to the lung (ie, bronchial and pulmonary). Generally, pulmonary infarction is due to smaller emboli that become lodged in more distal pulmonary arteries and is nearly always completely reversible; pulmonary infarction is recognized early, using sensitive radiographic criteria, often before necrosis occurs.
PE can also arise from nonthrombotic sources.
Many pulmonary emboli are small, physiologically insignificant, and asymptomatic. Even when present, symptoms are nonspecific and vary in frequency and intensity, depending on the extent of pulmonary vascular occlusion and preexisting cardiopulmonary function.
Emboli often cause
Dyspnea may be minimal at rest and can worsen during activity.
Less common symptoms include
In elderly patients, the first symptom may be altered mental status.
Massive PE may manifest with hypotension, tachycardia, light-headedness/presyncope, syncope, or cardiac arrest.
The most common signs of PE are
Less commonly, patients have hypotension. A loud 2nd heart sound (S2) due to a loud pulmonic component (P2) is possible but uncommon in acute PE because increases in pulmonary artery pressures are only modest. Crackles or wheezing may occur but are usually due to comorbid disease. In the presence of right ventricular failure, distended internal jugular veins and a RV heave may be evident, and a RV gallop (3rd heart sound [S3]), with or without tricuspid regurgitation, may be audible.
Fever, when present, is usually low-grade unless caused by an underlying condition.
Pulmonary infarction is typically characterized by chest pain (mainly pleuritic) and, occasionally, hemoptysis. The chest wall may be tender.
Chronic thromboembolic pulmonary hypertension causes symptoms and signs of right heart failure, including exertional dyspnea, easy fatigue, and peripheral edema that develops over months to years.
Patients with acute PE may also have symptoms of deep venous thrombosis (ie, pain, swelling, and/or erythema of a leg or an arm). Such leg symptoms are often not present, however.
The diagnosis is challenging because symptoms and signs are nonspecific and diagnostic tests are not 100% sensitive and specific. It is important to include PE in the differential diagnosis when nonspecific symptoms, such as dyspnea, pleuritic chest pain, hemoptysis, light-headedness, or syncope are encountered. Thus, PE should be considered in the differential diagnosis of patients suspected of having
Acute chest syndrome (in patients with sickle cell disease)
Acute anxiety with hyperventilation
Significant, unexplained tachycardia may be a clue. PE also should be considered in any elderly patient with tachypnea and altered mental status.
Initial evaluation should include pulse oximetry and chest x-ray. ECG, ABG measurements, or both may help to exclude other diagnoses (eg, acute myocardial infarction).
The chest x-ray usually is nonspecific but may show atelectasis, focal infiltrates, an elevated hemidiaphragm, or a pleural effusion. The classic findings of focal loss of vascular markings (Westermark sign), a peripheral wedge-shaped density (Hampton hump), or enlargement of the right descending pulmonary artery are suggestive but uncommon (ie, insensitive) and have low specificity. Chest x-ray can also help exclude pneumonia. Pulmonary infarction due to PE may be mistaken for pneumonia.
Pulse oximetry provides a quick way to assess oxygenation; hypoxemia is one sign of PE, and it requires further evaluation. Arterial or venous blood gas measurement may show an increased alveolar to arterial oxygen (A-a) difference (sometimes called A-a gradient) or hypocapnia; one or both of these tests are moderately sensitive for PE, but neither is specific. Blood gas testing should be considered particularly for patients with dyspnea or tachypnea who do not have hypoxemia detected with pulse oximetry. Oxygen saturation may be normal due to a small clot burden, or to compensatory hyperventilation; a very low pCO2 detected with ABG measurement can confirm hyperventilation.
ECG most often shows tachycardia and various ST-T wave abnormalities, which are not specific for PE (see figure An ECG in pulmonary embolism). An S1Q3T3 or a new right bundle branch block may indicate the effect of abrupt rise in RV size affecting RV conduction pathways; these findings are moderately specific but insensitive, occurring in only about 5% of patients, although the findings occur in a higher percentage of patients with massive PE. Right axis deviation (R > S in V1) and P-pulmonale may be present. T-wave inversion in leads V1 to V4 also occurs.
Clinical probability of pulmonary embolism can be assessed by combining ECG and chest x-ray findings with findings from the history and physical examination. Clinical prediction scores, such as the Wells score or the revised Geneva score (1), or the Pulmonary Embolism Rule-Out Criteria (PERC) rule, may aid clinicians in assessing the chance that acute PE is present. These prediction scores assign points to a variety of clinical factors, with cumulative scores corresponding to designations of the probability of PE before testing (pretest probability). For example, the Wells score result is classified as likely or unlikely for PE. Clinical probability scoring has been best studied in patients presenting to the emergency department.
One of the important clinical criteria is a judgment of whether PE is more likely than an alternate diagnosis, and this determination is somewhat subjective. However, the clinical judgment of experienced clinicians is as sensitive as, or even more sensitive, than results from formal prediction scores. PE should probably be considered more likely if one or more of the symptoms and signs, particularly dyspnea, hemoptysis, tachycardia, or hypoxemia, cannot be explained clinically or by chest x-ray results.
Pretest probability guides testing strategy and the interpretation of test results. In patients in whom the probability of PE is unlikely, only minimal additional testing (ie, D-dimer testing in outpatients) may be needed. In such cases, a negative D-dimer test (< 0.4 mcg/mL) is highly indicative of the absence of PE. Conversely, if there is a high clinical suspicion of PE and the risk of bleeding is low, immediate anticoagulation should be considered while the diagnosis is confirmed with additional tests.
The PERC rule specifies 8 criteria. Presence of these criteria in a clinically low-risk patient specifies that testing for PE is not indicated. The criteria are:
Use of the PERC rule has been recommended as a way to decrease rates of testing for PE than with conventional testing using d-dimer, but with similar rates of sensitivity and negative predictive values.
Screening of outpatients with D-dimer testing if pre-test probability is low or of intermediate probability
If pre-test probability is likely or if D-dimer result is elevated, CT angiography, or V/Q scanning if renal insufficiency is present or when CT contrast is contraindicated .
Sometimes ultrasonography of the legs or arms (to confirm DVT when lung imaging is delayed or prohibitive)
There is no universally accepted algorithm for the approach to suspected acute pulmonary embolism. Tests most useful for diagnosing or excluding PE are
Echocardiography may be useful to identify pulmonary embolism on the way to the lung (clot-in-transit).
D-Dimer is a by-product of intrinsic fibrinolysis; thus, elevated levels occur in the presence of a recent thrombus. When pretest probability is considered low or intermediate, a negative D-dimer level (< 0.4 mcg/mL) is highly sensitive for the absence of PE with a negative predictive value of > 95%; in most cases, this result is sufficiently reliable for excluding the diagnosis of PE in routine practice. However, elevated D-dimer levels are not specific for venous thrombus because many patients without deep venous thrombosis (DVT) or PE also have elevated levels, and therefore, further testing is required when the D-dimer level is elevated or when the pretest probability for PE is likely.
CT angiography is the preferred imaging technique for diagnosing acute pulmonary embolism. It is rapid, accurate, and highly sensitive and specific. It can also give more information about other lung pathology (eg, demonstration of pneumonia rather than PE as a cause of hypoxia or pleuritic chest pain) as well as severity of PE (for example by the size of the right ventricle or the reflux into the hepatic veins). Although poor quality scans due to motion artifact or poor contrast boluses can limit the sensitivity of the examination, improvements in CT technology have shortened acquisition times to less than 2 sec, providing relatively motion-free images in patients who are dyspneic. Fast scanning times allow the use of smaller volumes of iodinated contrast, which reduces the risk of acute kidney injury.
The sensitivity of CT angiography is highest for pulmonary embolism in the main pulmonary artery and lobar and segmental vessels. Sensitivity of CT angiography is lowest for emboli in subsegmental vessels (about 30% of all PEs). However, the sensitivity and specificity of CT angiography have improved as technology has evolved.
V/Q scans in PE detect areas of lung that are ventilated but not perfused. V/Q scanning takes longer than CT angiography and is less specific. However, when chest x-ray findings are normal or near normal and no significant underlying lung disease exists, it is a highly sensitive test. V/Q scanning is particularly useful when renal insufficiency precludes the use of contrast that is otherwise required for CT angiography. In some hospitals, V/Q scanning can be done with a portable machine that provides 3 views of ventilation and perfusion, which is useful when a patient is too ill to move. Results are reported as normal, very low, low, intermediate, or high probability of PE based on patterns of V/Q mismatch. A completely normal scan excludes PE with nearly 100% accuracy, but a low probability scan still carries a 15% likelihood of PE. Perfusion defects may occur in many other lung conditions (eg, COPD, pulmonary fibrosis, pneumonia, pleural effusion). Mismatched perfusion defects that may mimic PE may occur in pulmonary vasculitis, pulmonary veno-occlusive disease, and sarcoidosis.
With an intermediate-probability scan, there is a 30 to 40% probability of PE; with a high-probability scan, there is an 80 to 90% probability of PE. The results of clinical probability testing must be used together with the scan result to determine the need for treatment or further testing.
Duplex ultrasonography is a safe, noninvasive, portable technique for detecting leg or arm thrombi. A clot can be detected by showing poor compressibility of the vein or by showing reduced flow by Doppler ultrasonography. The test has a sensitivity of > 95% and a specificity of > 95% for thrombus. Confirming DVT in the calf or iliac veins can be more difficult but can generally be accomplished. The ultrasound technician should always attempt to image below the popliteal vein into its trifurcation.
Absence of thrombi in the leg veins does not exclude the possibility of thrombus from other sources, such as the upper extremities, but patients with suspected DVT and negative results on Doppler duplex ultrasonography have > 95% event-free survival, because thrombi from other sources are so much less common.
Although ultrasonography of the legs or arms is not diagnostic for PE, a study that reveals leg or axillary-subclavian thrombus establishes the need for anticoagulation and may obviate the need for further diagnostic testing unless more aggressive therapy (eg, thrombolytic therapy) is being considered. Therefore, stopping the diagnostic evaluation after detection of DVT on ultrasonography of the legs or arms is most appropriate for stable patients with contraindications to CT contrast and in whom V/Q scanning is expected to have low specificity (eg, in patients with an abnormal chest x-ray). In suspected acute PE, a negative ultrasound does not negate the need for additional studies.
Echocardiography may show a clot in the right atrium or ventricle, but it is most commonly used for risk stratification in acute PE. The presence of right ventricular dilation and hypokinesis may suggest the need for more aggressive therapy.
Cardiac marker testing is evolving as a useful means of stratifying mortality risk in patients with acute PE. Cardiac marker testing can be used adjunctively if PE is suspected or proven. Elevated troponin levels signify right ventricular (or sometimes left ventricular) ischemia. Elevated brain natriuretic peptide (BNP) and pro-BNP levels may signify RV dysfunction; however, these tests are not specific for RV dysfunction or for PE.
Thrombotic disorder (thrombophilia) testing should be considered for patients with PE and no known risk factors, especially if they are younger, have recurrent PE, or have a positive family history. Certain thrombophilias, such as antiphospholipid antibody syndrome, require disease-specific types of anticoagulation therapy.
Pulmonary arteriography is now rarely needed to diagnose acute PE because noninvasive CT angiography has similar sensitivity and specificity. However, in patients in whom catheter-based thrombolytic therapy is being used, pulmonary angiography is useful for assessment of catheter placement and may be used as a rapid means of determining success of the procedure when the catheter is removed. Pulmonary arteriography is also still used together with right-heart catheterization in assessing whether patients with chronic thromboembolic pulmonary hypertension are candidates for pulmonary endarterectomy.
An estimated 10% of patients with PE die within the first few hours after presentation. Most patients who die as a result of acute PE are never diagnosed before death. In fact, PE is not suspected in most of these patients. The best prospects for reducing mortality involve
Very high d-dimer levels appear to predict a poor outcome.
Patients with chronic thromboembolic disease represent a small, but important fraction of patients with PE who survive. Anticoagulant therapy reduces the rate of recurrence of PE to about 5%, and some studies have found anticoagulants reduce recurrence rates even lower.
Rapid assessment for the need for supportive therapy should be undertaken. In patients with hypoxemia, oxygen should be given. In patients with hypotension due to massive PE, 0.9% saline can be cautiously given IV; overloading the right ventricle can result in deterioration. Vasopressors may also be given if IV fluids fail to sufficiently increase blood pressure. Norepinephrine is the most commonly used first-line agent. Epinephrine and dobutamine have inotropic effects, but it is not clear how much these affects the normally thin-walled RV.
Anticoagulation is the mainstay of therapy for PE, and rapid reduction of clot burden via thrombolytic therapy or embolectomy is indicated for patients with hypotension that does not resolve after fluid resuscitation, and for selected patients with impaired RV function. Placement of a removable percutaneous inferior vena cava filter (IVCF) should be considered for patients with contraindications to anticoagulation or for those with recurrent PE despite anticoagulation. For example, patients who have acute PE, residual clot in the leg, and cannot be anticoagulated, should have a filter placed because they have persistent risk of subsequent DVT.
Most patients with PE should be hospitalized for at least 24 to 48 h. Patients with abnormal vital signs or massive or submassive PE require longer periods of hospitalization.
Massive PE always requires ICU admission. ICU admission should also be considered if patients have:
Patients with incidentally discovered PE or those with very small clot burdens and minimal symptoms may be managed as outpatients if their vital signs are stable, education is undertaken, and if a reasonable plan for outpatient treatment and follow-up is in place.
Increasingly, pulmonary embolism response teams (PERTs), including clinicians in pulmonary/critical care medicine, interventional cardiology, cardiothoracic surgery, hematology, and other specialties, are forming in the US; these teams aim to rapidly evaluate and treat patients with acute PE.
Initial anticoagulation followed by maintenance anticoagulation is indicated for patients with acute pulmonary embolism to prevent clot extension and further embolization as well as new clot formation. Anticoagulant therapy for acute PE should be started whenever PE is strongly suspected, as long as the risk of bleeding is deemed low. Otherwise, anticoagulation should be started as soon as the diagnosis is made. The likelihood of benefit and harm in treatment of emboli in smaller, subsegmental vessels (particularly asymptomatic and incidentally discovered emboli) is currently unknown, and it is feasible that in certain settings harm may outweigh benefit. Still, treatment is currently recommended. The primary complication of anticoagulation therapy is bleeding, and patients should be closely observed for bleeding during hospitalization.
Initial anticoagulation choices for acute PE include
Intravenous unfractionated heparin has a short half-life (useful when the potential for bleeding is deemed higher than usual) and is reversible with protamine. An initial bolus of unfractionated heparin is given, followed by an infusion of heparin dosed by protocol to achieve an activated PTT 1.5 to 2.5 times that of normal control (see figure Weight-based heparin dosing). Therefore, unfractionated heparin requires ongoing hospitalization to administer. Further, the pharmacokinetics of unfractionated heparin are relatively unpredictable, resulting in frequent periods of over-anticoagulation and under-anticoagulation and necessitating frequent dose adjustments. Regardless, many clinicians prefer this IV unfractionated heparin regimen, particularly when thrombolytic therapy is given or contemplated or when patients are at risk of bleeding because if bleeding occurs, the short half-life means that anticoagulation is quickly reversed after the infusion is stopped.
Subcutaneous low molecular weight heparin has several advantages over unfractionated heparin including
Weight-based dosing results in a more predictable anticoagulation effect than does weight-based dosing of unfractionated heparin
Ease of administration (can be given sc once or twice daily)
Decreased incidence of bleeding
Potentially better outcomes
The potential for patients to self-inject (thereby allowing earlier discharge from the hospital)
Lower risk of heparin-induced thrombocytopenia compared with standard, unfractionated heparin
In patients with renal insufficiency, dose reductions are needed (see table Some Low Molecular Weight Heparin Options in Thromboembolic Disease), and subsequent verification of appropriate dosing should be done by checking serum factor Xa levels (target: 0.5 to 1.2 IU/mL measured at 3 to 4 h after the 4th dose). Low molecular weight heparins are generally contraindicated in patients with severe renal insufficiency (creatinine clearance < 30 mL/min). Low molecular weight heparins are partially reversible with protamine.
Some Low Molecular Weight Heparin* Options in Thromboembolic Disease
Low Molecular Weight Heparin
100 units/kg sc q 12 h or 200 units/kg once/day†, ‡
2500–5000 units once/day
1 mg/kg sc q 12 h or 1.5 mg/kg sc once/day§
After abdominal surgery: 40 mg sc once/day
After hip replacement surgery: 40 mg sc once/day or 30 mg sc q 12 h; after knee replacement: 30 mg sc q 12h
For unstable angina or non-Q wave MI: 1 mg/kg sc q 12 h
For other (medical) patients not undergoing surgery: 40 mg sc once/day
175 units/kg sc once/day (in patients with or without PE)¶
3500 units once/day
*For dosing for unfractionated heparin, see figure Weight-based heparin dosing.
Note: Although low molecular weight heparins can be given by continuous IV infusion, this form of administration is rarely necessary or indicated. Low molecular weight heparins usually are given by sc injection in the abdominal area while the patient is supine.
†In patients with cancer, dalteparin is dosed as 200 units/kg once/day for the first 30 days of treatment.
‡In patients who develop renal insufficiency, dalteparin dosing should be re-evaluated.
§In patients with renal insufficiency (creatinine clearance < 30 mL/min), the dose of enoxaparin must be reduced or the drug stopped.
¶In patients with renal insufficiency, tinzaparin is given cautiously, although there are no specific recommendations.
MI = myocardial infarction; PE = pulmonary embolism; sc = subcutaneous
Adverse effects of all heparins include
Thrombocytopenia (including heparin-induced thrombocytopenia with the potential for thromboembolism)
Bleeding caused by over-heparinization with unfractionated heparin can be treated with a maximum of 50 mg of protamine per 5000 units unfractionated heparin infused over 15 to 30 min. Over-heparinization with a low molecular weight heparin can be treated with protamine 1 mg in 20 mL normal saline infused over 10 to 20 min, although the precise dose is undefined because protamine only partially neutralizes low molecular weight heparin inactivation of factor Xa.
Fondaparinux is a newer factor Xa antagonist. It can be used in acute DVT and acute PE instead of heparin or low molecular weight heparin. It has also been shown to prevent recurrences in patients with superficial venous thrombosis. Outcomes appear to be similar to those of unfractionated heparin. Advantages include once or twice daily fixed-dose administration, no need for monitoring of the degree of anticoagulation, and lower risk of thrombocytopenia. The dose (in mg/kg once/day) is 5 mg for patients < 50 kg, 7.5 mg for patients 50 to 100 kg, and 10 mg for patients >100 kg. Fondaparinux dose is decreased by 50% if creatinine clearance is 30 to 50 mL/min. The drug is contraindicated if creatinine clearance is < 30 mL/min.
The other newer factor Xa inhibitors, apixaban, rivaroxaban, and edoxaban, have the advantages of oral fixed dosing, the ability to be used as maintenance anticoagulants with no need for laboratory monitoring of the anticoagulant effect. They also cause few adverse interactions with other drugs, although azole antifungal therapy and older HIV therapies (protease inhibitors) will increase oral factor Xa inhibitor drug levels, and certain anticonvulsants and rifampin will decrease oral factor Xa inhibitor drug levels. Although rivaroxaban and apixaban do not require overlap with a parenteral anticoagulant when used as initial therapy, edoxaban requires use of a parenteral anticoagulant for 5 to 10 days.
Dose reductions are indicated for patients with renal insufficiency. Apixaban can be used in patients with renal insufficiency and recent data suggest use is safe in patients undergoing hemodialysis.
Anticoagulation reversal of the oral Xa inhibitors (rivaroxaban, apixaban, edoxaban) is possible with andexanet, although this drug is not widely available at this time. Also, the half-lives of the newer factor Xa inhibitors are much shorter than the half-life for warfarin. If bleeding develops that requires reversal, use of 4-factor prothrombin complex concentrate can be considered, and hematology consultation is recommended.
The safety and efficacy of these drugs in patients with PE complicated by severe cardiopulmonary decompensation have not yet been studied, and parenteral drugs should be used for anticoagulation in these patients until there is significant improvement in cardiopulmonary function.
The direct thrombin inhibitor dabigatran has also proven effective for treatment of acute DVT and PE. Idarucizumab has proven effective at reversing dabigatran.
Finally, in patients with suspected or proven heparin-induced thrombocytopenia, intravenous argatroban or subcutaneous fondaparinux can be used for anticoagulation. Use of the newer oral anticoagulants is currently being studied in patients with heparin-induced thrombocytopenia, but these drugs appear safe after platelet recovery.
Maintenance anticoagulation is indicated to reduce the risk of clot extension or embolization and to reduce the risk of new clot formation. Drug choices for maintenance anticoagulation include
Warfarin is an effective long-term oral anticoagulant option that has been used for decades, but it is very inconvenient for a number of reasons. In most patients, warfarin is started on the same day as heparin (or fondaparinux) therapy used for initial anticoagulation. Heparin (or fondaparinux) therapy should be overlapped with warfarin therapy for a minimum of 5 days and until the INR has been within the therapeutic range (2.0 to 3.0) for at least 24 h.
The major disadvantages of warfarin are the need for periodic INR monitoring, with frequent dose adjustments, and drug interactions. Physicians prescribing warfarin should be wary of drug interactions; in a patient taking warfarin, virtually any new drug should be checked.
Bleeding is the most common complication of warfarin treatment; patients > 65 yr and those with comorbidities (especially diabetes, recent myocardial infarction, hematocrit < 30%, or creatinine > 1.5 mg/dL) and a history of stroke or GI bleeding seem to be at greatest risk. Bleeding can be reversed with vitamin K 2.5 to 10 mg IV or po and, in an emergency, with fresh frozen plasma or a new concentrate formulation (prothrombin complex concentrates) containing factor II (prothrombin), factor VII, factor IX, factor X, protein C, and protein S. Vitamin K may cause flushing, local pain, and, rarely, anaphylaxis.
Warfarin-induced necrosis, a devastating complication of warfarin therapy, can occur in patients with heparin-induced thrombocytopenia if warfarin is started before platelet recovery. Based on these considerations and the development of more convenient oral anticoagulants, it is likely that warfarin use will decline substantially over the coming years.
The oral factor Xa inhibitor anticoagulants apixaban and rivaroxaban can be used for both initial and maintenance anticoagulation therapy (see table Oral Anticoagulants). These drugs are more convenient than warfarin due to their fixed dosing and lack of need for laboratory monitoring, as well as having fewer drug interactions. In clinical trials, rivaroxaban (1, 2), apixaban (3), and edoxaban (4) were as effective (in non-inferiority analyses) as warfarin in preventing recurrent DVT and PE. A meta-analysis of large phase III randomized controlled trials found that rates of major bleeding, including intracranial hemorrhage, were significantly lower with oral factor Xa inhibitor anticoagulants than with warfarin (5). Another advantage of both rivaroxaban and apixaban is that dosages may be lowered (10 mg po once/day of rivaroxaban and 2.5 mg po bid of apixaban) after patients have been treated for 6 to 12 mo.
Edoxaban requires that a preceding 5 to 10 days of initial heparin or low molecular weight heparin be given.
The direct thrombin inhibitor dabigatran can also be used for maintenance anticoagulation therapy. As with edoxaban, 5 to 10 days of treatment with unfractionated heparin or low molecular weight heparin is needed before dabigatran can be initiated. Clinically relevant bleeding was lower with dabigatran than with warfarin. The use of dabigatran as maintenance therapy has the same advantages and disadvantages as the use of the factor Xa inhibitors.
The need for initial heparin treatment before edoxaban or dabigatran is given is a reflection of the way the clinical trials were conducted.
Subcutaneous low molecular weight heparin is primarily used for high-risk cancer patients or patients with recurrent PE despite other anticoagulants. Recent data (SELECT-D trial) suggest efficacy of rivaroxaban in cancer patients (6).
Aspirin has been studied for long-term maintenance therapy. It appears more effective than placebo but less effective than all other available anticoagulants. Rivaroxaban, 10 mg once/day, has proven more effective at reducing recurrent DVT/PE yet is as safe as aspirin in patients already treated with anticoagulation for 6 to 12 mo (7).
Duration of maintenance anticoagulation for PE is dependent on a variety of factors (eg, risk factors for PE, bleeding risk) and can range from 3 mo to lifelong therapy. Clearly transient risk factors (eg, immobilization, recent surgery, trauma) require only 3 mo of treatment. Patients with unprovoked PE, those with more durable risk factors for PE (eg, cancer, thrombophilic disorder), and those with recurrent PE might benefit from lifelong anticoagulation provided the bleeding risk is low or moderate. In many patients, degree of risk is less clear (eg, with a minor precipitating factor such as a 4 hour flight); for them, rather than stopping rivaroxaban or apixaban at 6 months, dosage can be decreased.
Risk factors for bleeding include
Low risk for bleeding is defined as no bleeding risk factors, moderate risk for bleeding is defined as one risk factor, and high risk for bleeding is defined as two or more risk factors.
As described above, after 6 months of treatment with rivaroxaban or apixaban, dosage decreases can be considered.
1. EINSTEIN Investigators, Bauersachs R, Berkowitz SD, et al: Oral rivaroxaban for symptomatic venous thromboembolism. N Engl J Med 363(26):2499–2510, 2010.
2. EINSTEIN-PE Investigators, Buller HR, Prins MH, et al: Oral rivaroxaban for the treatment of symptomatic pulmonary embolism. N Engl J Med 366 (14):1287–1297, 2012.
3. Agnelli G, Buller HR, Cohen A, et al: Oral apixaban for the treatment of acute venous thromboembolism.N Engl J Med 369(9):799–808, 2013.
4. Hokusai-VTE Investigators, Buller HR, Decousus H, et al: Edoxaban versus warfarin for the treatment of symptomatic venous thromboembolism. N Engl J Med 369(15): 1406–1415, 2013.
5. van Es N, Coppens M, Schulman S, et al: Direct oral anticoagulants compared with vitamin K antagonists for acute symptomatic venous thromboembolism: evidence from phase 3trials. Blood124 (12): 1968–1975, 2014.
6. Young AM, Marshall A, Thirlwall J, et al: Comparison of an oral factor Xa inhibitor with LMWH in patients with cancer with VTE Results of a randomized trial (SELECT-D). J Clin Oncol 36 (20):2017–2029, 2018.
7. Weitz JI, Lensing AWA, Prins MH, et al: Rivaroxaban or aspirin for extended treatment of venous thromboembolism. N Engl J Med 376:1211–1222, 2017. doi: 10.1056/NEJMoa1700518. Epub 2017 Mar 18.
Clot elimination by means of embolectomy or dissolution with IV or catheter-based thrombolytic therapy should be considered for acute PE associated with hypotension that does not resolve after fluid resuscitation (massive PE). Patients who are hypotensive and require vasopressor therapy are obvious candidates. Patients with a systolic BP < 90 mm Hg lasting at least 15 min are hemodynamically compromised and are also candidates.
Although only anticoagulation is generally recommended for patients with very mild RV dysfunction (based on clinical, ECG, or echocardiographic findings), thrombolytic therapy or embolectomy may be needed when RV compromise and/or hypoxemia is severe even when hypotension is not present, particularly when deterioration is likely as suggested by an increase in heart rate or decrease in oxygen saturation or blood pressure.
Systemic thrombolytic therapy with alteplase (tissue plasminogen activator [tPA]), offers a noninvasive way to rapidly restore pulmonary blood flow but is controversial because long-term benefits do not clearly outweigh the risk of hemorrhage. Regardless, most experts agree that systemic thrombolytic therapy should be given to patients with hemodynamic compromise, particularly when it is severe. Although no single prospective randomized trial of systemic thrombolytic therapy has shown improved survival in patients with submassive PE, some experts recommend thrombolytics, particularly when patients also have numerous or large clots, very severe RV dysfunction, marked tachycardia, significant hypoxemia, and other concomitant findings such as residual clot in the leg, positive troponin values, and/or elevated BNP values. Others reserve thrombolytic therapy only for patients with massive (high-risk) PE. Streptokinase and urokinase generally are no longer used.
Absolute contraindications to thrombolytics include
Relative contraindications include
Recent surgery (≤ 10 days)
Hemorrhagic diathesis (as in hepatic insufficiency)
Recent punctures of large noncompressible veins (eg, subclavian or internal jugular veins)
Recent femoral artery catheterization (eg, ≤ 10 days)
Peptic ulcer disease or other conditions that increase the risk of bleeding
Severe hypertension (systolic BP > 180 mm Hg or diastolic BP > 110 mm Hg)
Head trauma from PE-induced syncope, even if brain CT is normal
Except for concurrent intracerebral hemorrhage, thrombolytic therapy is sometimes given to patients with massive PE who have "absolute contraindications" to such therapy if death is otherwise expected. In patients with relative contraindications, the decision to give systemic thrombolytics depends on individual patient factors.
In the US, alteplase (tPA) is used for systemic thrombolysis (see table Regimens for Systemic Thrombolysis). Streptokinase and urokinase are no longer used for acute PE.
In the US, when systemic thrombolytics are given, heparin is usually stopped after the initial loading dose. However, in Europe, heparin is often continued, and there is no clear determination as to which method is preferred. The bleeding risk should be considered.
Regimens for Systemic Thrombolysis
Bleeding, if it occurs, can be reversed with cryoprecipitate or fresh frozen plasma. Accessible vascular access sites that are bleeding can be compressed. The potential for bleeding after systemic thrombolysis has led to increased implementation of catheter-based thrombolysis, because much lower doses of thrombolytic agents are used.
Catheter-directed PE therapy (thrombolytics, embolectomy) uses catheter placement in the pulmonary arteries for disruption and/or lysis of clot. It is used to treat massive PE. Indications for the treatment of submassive PE are evolving. Studies to date, including prospective randomized clinical trials, have demonstrated that this approach leads to an improved RV/LV ratio at 24 h compared with anticoagulation alone. Other outcomes and safety of catheter-based therapy compared to systemic thrombolysis are under investigation.
In catheter-based PE thrombolytic therapy, the pulmonary arteries are accessed via a typical right-heart catheterization/pulmonary arteriography procedure, and thrombolytics are delivered directly to large proximal emboli via the catheter. The most widely studied technique uses high-frequency, low-power ultrasonography to facilitate delivery of the thrombolytics. Ultrasonography accelerates the thrombolytic process by disaggregating fibrin strands and increasing permeability of lytic drug into the clot. Standard dosing has been 20 to 24 mg of tPA over 15 or more hours, but lower doses and shorter durations have recently been shown to be effective with this technique.
Other clot extraction techniques involve catheter-directed vortex suction embolectomy, sometimes in combination with extracorporeal bypass. Catheter-directed vortex suction embolectomy differs from systemic thrombolysis and catheter-based PE thrombolytic therapy in that a larger bore catheter is required and blood that is suctioned out must be redirected out back into a vein (usually femoral). Patients with venal caval, right atrial, or right ventricular thrombi-in-transit are the best candidates. The pulmonary arteries are difficult to access with the current devices. Veno-arterial extracorporeal membrane oxygenation (ECMO) may be used as a rescue procedure in severely ill patients with acute PE, regardless of what other therapies are used. Other smaller suctioning devices are currently being studied.
Surgical embolectomy is reserved for patients with PE who are hypotensive despite supportive measures (persistent systolic BP ≤ 90 mm Hg after fluid therapy and oxygen or if vasopressor therapy is required) or on the verge of cardiac or respiratory arrest. Surgical embolectomy should be considered if use of thrombolysis is contraindicated; in such cases, catheter-directed vortex embolectomy may also be considered and, depending on local resources and expertise, tried before surgical embolectomy. Surgical embolectomy appears to increase survival in patients with massive PE but is not widely available. As with catheter-based thrombosis/clot extraction, the decision to proceed with embolectomy and the choice of technique depend on local resources and expertise.
Prevention of pulmonary embolism means prevention of deep venous thrombosis (DVT); the need depends on the patient’s risks, including
Bedbound patients and patients undergoing surgical, especially orthopedic, procedures benefit, and most of these patients can be identified before a thrombus forms (see table Risk Assessment for Thrombosis). Preventive measures include low-dose unfractionated heparin, low molecular weight heparin, warfarin, fondaparinux, oral anticoagulants (eg, rivaroxaban, apixaban), compression devices, and elastic compression stockings.
Choice of drug or device depends on various factors, including the patient population, the perceived risk, contraindications (eg, bleeding risk), relative costs, and ease of use. The American College of Chest Physicians has published comprehensive evidence-based recommendations for prophylaxis of acute DVT, including the duration of prophylaxis, in surgical and nonsurgical patients and during pregnancy (The American College of Chest Physicians Guidelines on Prevention of Thrombosis). The need for prophylaxis has been studied in numerous patient populations.
The type of surgery as well as patient-specific factors determine the risk of DVT. Independent risk factors include
The Caprini score is commonly used for DVT risk stratification and determination of the need for DVT prophylaxis in surgical patients (see table Risk Assessment for Thrombosis).
Risk Assessment for Thrombosis
The need for DVT prophylaxis is based on the risk assessment score (see table Prophylaxis Based on Caprini Score). Appropriate preventive measures, ranging from early ambulation to use of heparin, depend on the total score.
Drug therapy to prevent DVT is usually begun after surgery to help prevent intraoperative bleeding. However, preoperative prophylaxis is also effective.
In general surgery patients, low dose unfractionated heparin is given in doses of 5000 units sc q 8 to 12 h for 7 to 10 days or until the patient is fully ambulatory. Immobilized patients not undergoing surgery should receive 5000 units sc q 8 to 12 h until they are ambulatory.
Low molecular weight heparin dosing for DVT prophylaxis depends on the specific drug (enoxaparin, dalteparin, tinzaparin). Low molecular weight heparins are at least as effective as low dose unfractionated heparin for preventing DVT and PE.
Fondaparinux 2.5 mg sc once/day is as effective as low molecular weight heparin for orthopedic surgery and in some other settings. It is a selective factor Xa inhibitor.
Warfarin is usually effective and safe at a dose of 2 to 5 mg po once/day or at a dose adjusted to maintain an INR of 2 to 3 in patients who have undergone total hip or knee replacement. It is still used by some orthopedic surgeons for prophylaxis in these patients but is increasingly being supplanted by the use of the newer oral anticoagulants.
Rivaroxaban, an oral factor Xa inhibitor, is used for prevention of acute DVT/PE in patients undergoing total knee or hip arthroplasty. The dose is 10 mg po once/day. Its use in other patients (surgical and nonsurgical) is currently under investigation.
Apixaban, an oral factor Xa inhibitor, is also used for prevention of acute DVT/PE in patients undergoing total knee or hip arthroplasty. The dose is 2.5 mg po bid. Like rivaroxaban, its use in other types of patients is currently under investigation.
Inferior vena cava filters, intermittent pneumatic compression (also known as sequential compression devices [SCD]), and graded elastic compression stockings may be used alone or in combination with drugs to prevent PE. Whether these devices are used alone or in combination depends on the specific indication.
An inferior vena cava filter (IVCF) may help prevent PE in patients with DVT in the leg, but IVCF placement may risk long-term complications. Benefits outweigh risk if a second PE is predicted to be life-threatening; however, few clinical trial data are available. A filter is most clearly indicated in patients who have:
Because venous collaterals can develop, providing a pathway for emboli to circumvent the IVCF, and because filters occasionally thrombose, patients with recurrent DVT or nonmodifiable risk factors for DVT may still require anticoagulation. An IVCF is placed in the inferior vena cava just below the renal veins via catheterization of an internal jugular or femoral vein. Most IVCFs are removable. Occasionally, a filter dislodges and may migrate up the venous bed, even to the heart, and needs to be removed or replaced. A filter can also become thrombosed, causing bilateral venous congestion (including acute phlegmasia cerulea dolens) in the leg, lower body ischemia, and acute kidney injury.
Intermittent pneumatic compression (IPC) with SCDs provides rhythmic external compression to the legs or to the legs and thighs. It is more effective for preventing calf than proximal DVT. It is insufficient as sole prophylaxis after hip or knee replacement but is often used in low-risk patients after other types of surgery or in medical patients who have a low-risk of DVT or who are at high risk of bleeding. IPC can theoretically trigger PE in immobilized patients who have developed occult DVT while not receiving DVT prophylaxis.
Graded elastic compression stockings are likely less effective than external pneumatic leg compression, but one systematic meta-analysis suggested that they reduced the incidence of DVT in postoperative patients from 26% in the control group to 13% in the compression stockings group.
After surgical procedures with a high incidence of DVT/PE, low dose unfractionated heparin, low molecular weight heparin, or adjusted-dose warfarin is recommended.
After orthopedic surgery of the hip or knee, additional options include the newer oral anticoagulants, rivaroxaban and apixaban. These drugs are safe and effective and do not require laboratory tests to monitor the level of anticoagulation as is needed for warfarin.
For total hip arthroplasty, patients should continue to take anticoagulants for 35 days postoperatively. In selected patients at very high risk of both DVT/PE and bleeding, temporary placement of an IVCF is an option for prophylaxis.
A high risk of DVT/PE also occurs in patients undergoing elective neurosurgery and those with acute spinal cord injury and multiple trauma. Although physical methods (SCDs and elastic stockings) have been used in neurosurgical patients because of concern about intracranial bleeding, low molecular weight heparin appears to be an acceptable alternative. The combination of SCDs and low molecular weight heparin may be more effective than either alone in high-risk patients. Limited data support the combination of SCDs, elastic compression stockings, and low molecular weight heparin in patients with spinal cord injury or in multiple trauma. For very high-risk patients, a temporary IVCF may be considered.
In acutely ill medical patients, low dose unfractionated heparin, low molecular weight heparin, or fondaparinux can be given. SCDs, elastic compression stockings, or both may be used when anticoagulants are contraindicated. For ischemic stroke patients, low dose unfractionated heparin or low molecular weight heparin can be used; SCD, elastic compression stockings, or both may be beneficial.
Acute PE is a common and potentially devastating medical condition.
Clinical suspicion and a confirmatory diagnosis are essential because in most patients who die from acute PE, PE is not even suspected.
Because anticoagulation improves survival, patients should be anticoagulated when PE is diagnosed or strongly suspected.
Patients with massive PE and certain patients with submassive PE should be considered for thrombolytic therapy or embolectomy.
Prevention of deep vein thrombosis (and thus PE) should be considered in all at-risk hospitalized patients.