Cardiol Clin 22 (2004) 353–365

Acute pulmonary embolism Victor F. Tapson, MD Division of Pulmonary and Critical Care Medicine, 353 Bell Building, Duke University Medical Center, Durham, NC 27710, USA

Background and incidence

Pathophysiology

Pulmonary embolism (PE) occurs when venous thrombosis, usually from the deep veins of the proximal legs, travels to the lungs causing a potential spectrum of consequences, including dyspnea, chest pain, hypoxemia, and sometimes death. Deep venous thrombosis (DVT) and PE represent a continuum of the disease entity known as venous thromboembolism (VTE). PE probably accounts for 100,000 to 200,000 deaths per year in the United States [1,2]. Although some patients dying from acute PE have an underlying terminal illness, this disease entity seems to be responsible for death in a considerable number of patients with an otherwise good prognosis. Autopsy studies have repeatedly documented the high frequency with which PE has gone unsuspected and thus undetected [3]. VTE occurs worldwide and is usually, but not always, associated with specific risk factors [1,4–6]. A crucial point is that DVT and, therefore, PE are often preventable. Although other substances such as malignant cells, fat droplets, air bubbles, and carbon dioxide can embolize to the lung, this article focuses on VTE. Because of the lack of specific symptoms and signs, DVT and PE are frequently clinically unsuspected, leading to substantial diagnostic and therapeutic delays and resulting in considerable morbidity and mortality [1,2]. Furthermore, prophylaxis continues to be dramatically underused [4,7]. The incidence of VTE is high in hospitalized patients, and both surgical as well as medical patients are at risk.

One or more components of Virchow’s triad (stasis, hypercoagulability, and intimal injury), described more than 150 years ago, are present in nearly all patients [8]. The risk increases with age. Idiopathic VTE is well described and probably involves an underlying prothrombotic state that is present but awaits characterization. Deep vein thrombi frequently originate in the calf veins and propagate proximally before embolizing. Although emboli may occasionally originate directly from calf vein thrombi, more than 95% of thrombi that embolize to the lungs detach from a proximal deep vein of the lower extremities (including and above the popliteal veins). Thrombosis developing in the axillary-subclavian veins caused by the presence of a central venous catheter, particularly in patients with malignant disease and also in those with effort-induced upper extremity thrombosis, may result in PE as well. The impact of a particular embolic event depends on the extent of reduction of the crosssectional area of the pulmonary arterial bed and on the presence or absence of underlying cardiopulmonary disease [9,10]. Hypoxemia stimulates sympathetic tone with resulting systemic vasoconstriction, increased venous return, and a rise in stroke volume. With massive emboli, cardiac output is diminished but may be sustained as the mean right atrial pressure increases. The increase in pulmonary vascular resistance impedes right ventricular outflow and thus reduces left ventricular preload. In the absence of underlying cardiopulmonary disease, occlusion of 25% to 30% of the vascular bed by emboli is associated with a rise in pulmonary artery pressure [10]. With further vascular obstruction, hypoxemia worsens, stimulating

Division of Pulmonary and Critical Care Medicine, Box 31175, Duke University Medical Center, Durham, NC 27710. E-mail address: [email protected]

0733-8651/04/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ccl.2004.04.002

354

V.F. Tapson / Cardiol Clin 22 (2004) 353–365

vasoconstriction and a further rise in pulmonary artery pressure. More than 50% obstruction of the pulmonary arterial bed is usually present before there is substantial elevation of the mean pulmonary artery pressure. When the extent of obstruction of the pulmonary circulation approaches 75%, the right ventricle must generate a systolic pressure in excess of 50 mm Hg and a mean pulmonary artery pressure greater than 40 mm Hg to preserve pulmonary perfusion. A normal right ventricle is rarely able to meet this demand and, hence, fails. Patients with underlying cardiopulmonary disease often experience a more substantial deterioration in cardiac output than normal individuals in the setting of massive PE. Supportive measures may sustain a patient with massive PE, but additional increments in embolic burden may be fatal.

Clinical manifestations Unfortunately, history and physical examination are notoriously insensitive and nonspecific for both DVT and PE [11–13]. Patients with lower-extremity venous thrombosis often do not exhibit erythema, warmth, pain, swelling, or tenderness. When these signs are present, they are nonspecific but may still merit further evaluation. Pain with dorsiflexion of the foot (Homans’ sign) may be present in the setting of DVT, but this finding is neither sensitive nor specific. The most common symptom of acute PE is dyspnea that is often sudden in onset. Pleuritic chest pain and hemoptysis occur more commonly with pulmonary infarction caused by smaller, peripheral emboli. Palpitations, cough, anxiety, and lightheadedness are among the nonspecific symptoms of acute PE but may result from a number of other entities, contributing to the difficulty in making the diagnosis. Syncope or sudden death may occur with massive PE. Pulmonary embolism should always be considered whenever unexplained dyspnea, syncope, hypotension, or hypoxemia is present [11–13]. Tachypnea and tachycardia are the most common signs of pulmonary embolism but are also nonspecific. Other physical findings include fever, wheezing, rales, a pleural rub, a loud pulmonic component of the second heart sound, a right-sided fourth heart sound, and a right ventricular lift. Dyspnea, tachypnea, and hypoxemia in patients with concomitant cardiopulmonary disease (such as congestive heart failure, pneumonia, or chronic obstructive pulmonary disease) may be caused

by the underlying disease or by superimposed acute PE. Symptoms and signs consistent with PE (Tables 1 and 2) should be particularly heeded in patients with risk factors for VTE such as concomitant malignancy, immobility, or the postoperative state. Diagnosis The differential diagnosis for acute DVT and for PE depends on the clinical presentation and the presence of concomitant disease. For example, when patients present with calf pain or swelling in the setting of risk factors for VTE, a diagnostic study should be pursued unless there is a clear, alternative explanation. For dyspnea or chest pain, the differential diagnosis includes a flare of asthma or chronic obstructive lung disease, pneumothorax, pneumonia, anxiety with hyperventilation, heart failure, angina or myocardial infarction, musculoskeletal pain, rib fracture, pericarditis, pleuritis from collagen vascular disease, herpes zoster, intrathoracic cancer, and, occasionally, intra-abdominal processes such as acute cholecystitis. The presence of obvious risk factors for VTE, such as prolonged immobility [14], trauma [15–17], recent surgery [18], medical illness with reduced mobility [19], cancer [20,21], pregnancy [22], myocardial infarction [23], recent prolonged travel [24– 26], or previous thromboembolism in the setting of compatible symptoms and signs should prompt consideration of this entity. Acute PE can be superimposed upon another underlying cardiopulmonary disease, on which new or worsening symptoms are sometimes blamed. Table 1 Symptoms of acute pulmonary embolism

Symptom Dyspnea Pleuritic chest pain Cough Leg pain Hemoptysis Palpitations Wheezing Angina like pain

All Patients (n = 383)

Patients without Previous Cardiopulmonary Disease (n = 117)

78%

73%

59% 43% 27% 16% 13% 14%

66% 37% 26% 13% 10% 9%

6%

4%

Data from Refs. [12,13] and Stein PD, editor. Pulmonary embolism. Baltimore (MD): Williams and Wilkins; 1996.

V.F. Tapson / Cardiol Clin 22 (2004) 353–365 Table 2 Signs of acute pulmonary embolism

Tachypnea (20/min) Crackles Tachycardia (>100/min) Leg swelling Loud P2 DVT Wheezes Diaphoresis Temperature ($38.5() Pleural rub Fourth heart sound Third heart sound Cyanosis Homans’ sign Right ventricular lift

All Patients (n = 383)

Patients without Previous Cardiopulmonary Disease (n = 117)

73%

70%

55% 30%

51% 30%

31% 23% 15% 11% 10% 7%

28% 23% 11% 5% 11% 7%

4% –

3% 24%

5%

3%

3% 3% –

1% 4% 4%

DVT, deep venous thrombosis; P2, pulmonic component of second heart sound. Data from Refs. [12,13] and Stein PD, editor. Pulmonary embolism. Baltimore (MD): Williams and Wilkins; 1996.

Blood tests Hypoxemia is common in acute PE. Some individuals, particularly young patients without underlying lung disease, may have a normal arterial oxygen tension (PaO2) and even, rarely, a normal alveolar-arterial difference [11,13]. A sudden decrease in the PaO2 or in the oxygen saturation in a patient unable to communicate an accurate history (eg, a demented or mechanically ventilated patient) suggests the possibility of acute PE. The diagnostic utility of plasma measurements of circulating D-dimer (a specific derivative of cross-linked fibrin) in patients with acute PE has been extensively evaluated [27–30]. A normal ELISA seems to be sensitive in excluding PE, particularly when the clinical suspicion is relatively low. A number of D-dimer assays are available, and their sensitivity and specificity vary [30]. A positive D-dimer test means that DVT or PE is possible, but it is by no means proof. Similarly, although a negative D-dimer may strongly suggest that VTE is absent, a high clinical suspicion should not be ignored.

355

Clinical probability scores based on simple clinical parameters have been used together with a negative D-dimer to help exclude PE. In a recent prospective clinical trial, the SimpliRed D-dimer (AGEN Biomedical Limited, Brisbane, Australia) test, a rapid red blood cell agglutination D-dimer assay, was used together with simple scoring parameters readily available in the emergency department [28]. Of the 437 patients with a negative D-dimer result and low clinical probability in this study, only one developed PE during followup (Table 3). Whether or not such scoring systems are used, D-dimer assays may prove increasingly useful in excluding acute DVT and PE, particularly when low clinical suspicion supports its absence. Both cardiac troponin T and troponin I levels have been found to be elevated in acute PE [32,33]. This enzyme is specific for cardiac myocyte damage. The right ventricle seems to be the source of the enzyme elevation in acute PE and, in particular, in more massive embolism in which myocyte injury caused by right ventricular strain might be expected. Troponin levels cannot, however, be used like D-dimer testing; that is, when clinical suspicion is relatively low, they are not sensitive enough to rule out PE without additional diagnostic testing. Electrocardiography Electrocardiographic findings, which are present in most patients with acute PE, are nonspecific, but these abnormalities, including ST-segment abnormalities, T-wave changes, and left or right axis deviation, are common. Only one third of patients with massive or submassive emboli have manifestations of acute cor pulmonale such as the S1 Q3 T3 pattern, right bundle branch block, P-wave pulmonale, or right axis deviation. The usefulness of electrocardiography in suspected acute PE lies in its ability to establish or exclude alternative diagnoses such as acute myocardial infarction [13]. Chest radiography The chest radiograph is often abnormal in patients with acute PE, but, as with electrocardiography, it is nearly always nonspecific. Common radiographic findings include atelectasis, pleural effusion, pulmonary infiltrates, and mild elevation of a hemidiaphragm [13]. Classic findings of pulmonary infarction such as Hampton’s hump or decreased vascularity (Westermark’s sign) are

356

V.F. Tapson / Cardiol Clin 22 (2004) 353–365

Table 3 Determining pretest probability of acute pulmonary embolism using point system and D-dimer result Variable DVT symptoms/signs PE as or more likelya HR >100 beats/min Immobilization/surgeryb Previous DVT or PE Hemoptysis Malignancy Total score 6.0

Points 3.0 3.0 1.5 1.5 1.5 1.0 1.0 Pretest Probabilityc Low Moderate High

DVT, deep venous thrombosis; HR, heart rate; PE, pulmonary embolism. a PE is as likely or more likely than an alternative diagnosis. Physicians were told to use clinical information along with chest radiography, electrocardiography, and laboratory tests. b If within previous 4 weeks. c Of the 437 patients with a negative D-dimer result and low clinical probability, only one developed PE during follow-up; thus, the negative predictive value for the combined strategy of using the clinical model with D-dimer testing in these patients was 99.5%. Data from Ref [28].

suggestive of the diagnosis but are infrequent. A normal chest radiograph in the setting of severe dyspnea and hypoxemia without evidence of bronchospasm or anatomic cardiac shunt is strongly suggestive of PE. Pulmonary embolism frequently coexists with underlying heart or lung disease. Symptoms, signs, radiographic findings, electrocardiography, and the plasma D-dimer measurement cannot be considered diagnostic of PE. Similarly, symptoms, signs, and blood studies cannot prove the presence of DVT. When these entities are suspected, further evaluation with noninvasive or invasive testing is necessary. Deep venous thrombosis: the radiographic approach Venography has been the time-honored standard for the diagnosis of acute DVT. With the advent of ultrasound, a diagnostic test that is more than 90% sensitive in the setting of symptomatic DVT, venography is rarely used [34]. Similarly, another sensitive test, impedance plethysmography, is almost never used. MRI has proven extremely sensitive for both acute and chronic DVT [34–36], although it is generally not necessary. It is reasonable to consider MRI in the setting of suspected DVT when ultrasound cannot

be effectively used. A major limitation of ultrasound is its reduced sensitivity in the setting of asymptomatic DVT. Thus, ultrasound is not generally used as a screening test. Pulmonary embolism: the radiographic approach Ventilation-perfusion scanning Historically, the ventilation-perfusion (VQ) scan was the most commonly used diagnostic test when PE was suspected. During the past decade, spiral (helical) CT scanning has essentially replaced it at most centers. A normal perfusion scan rules out the diagnosis with a high enough degree of certainty that further diagnostic evaluation is almost never necessary [37]. Matching areas of decreased ventilation and perfusion in the presence of a normal chest radiograph generally represent a process other than PE. Nondiagnostic scans of low or intermediate probability are commonly found with PE, however, and in such situations further evaluation with pulmonary arteriography is often appropriate. In the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED), when the clinical suspicion was considered very high PE was present in 96% of patients with high-probability scans, in 66% of patients with intermediate scans, and in 40% of patients with low-probability scans [12]. If the clinical setting suggests the diagnosis, the possibility of PE should be rigorously pursued even when the lung scan shows low or intermediate probability. Stable patients with suspected acute PE, nondiagnostic lung scans, and adequate cardiopulmonary reserve (absence of hypotension or severe hypoxemia) may undergo noninvasive lower extremity testing in an attempt to diagnose DVT [38]. A positive compression ultrasound may present the opportunity to treat without further testing. If the ultrasound is negative, pulmonary angiography is an appropriate option. Serial noninvasive lower extremity testing in the setting of suspected PE should be performed only in centers where followup is guaranteed and validated protocols are used. MRI of the lower extremities may also be useful after a nondiagnostic lung scan if the medical facility has experience with this technique. Pulmonary arteriography Pulmonary arteriography remains the accepted standard technique for the diagnosis of acute PE. It is an extremely sensitive, specific, and safe test [39]. Complications of pulmonary arteriography in 1111 patients suspected of having PE in the

V.F. Tapson / Cardiol Clin 22 (2004) 353–365

357

PIOPED included death in 0.5% and major nonfatal complications in 1% [12]. Pulmonary arteriography is used when PE must be diagnosed or excluded and preliminary testing has been nondiagnostic. In some centers, pulmonary arteriography can be performed at the bedside using a pulmonary artery catheter and fluoroscopic guidance. It is being used less frequently because CT has increasingly been employed.

MRI

Spiral (helical) CT

Echocardiography in acute pulmonary embolism

Spiral CT scanning can be used for diagnosing both acute and chronic PE and is replacing VQ at many centers. Some clinical trials have suggested good sensitivity and specificity, but others reports have been less favorable. Retrospective reconstructions can be performed. A contrast bolus is required for imaging of the pulmonary vasculature. In at least one clinical trial, spiral CT was associated with greater than 95% sensitivity and specificity [40]. More recent and larger trials have suggested a lower sensitivity [41–48]. A large, prospective Swiss study revealed a sensitivity of 70%, suggesting that a negative CT scan may not absolutely rule out smaller emboli [49]. Data from the large, multicenter PIOPED II trial in the United States and Canada comparing CT (of the chest and legs) and VQ scanning is currently being analyzed. Spiral CT has the greatest sensitivity for emboli in the main, lobar, or segmental pulmonary arteries. The specificity for clot in these vessels is excellent. For subsegmental emboli, spiral CT seems to be less accurate, although the importance of emboli of this size has been questioned. The outcome of selected patients with a negative CT scan in the setting of suspected PE seems to be good in trials published thus far [50]. The use of thinner sections and techniques such as multiplanar three-dimensional reformation may enhance the usefulness of spiral CT for diagnosing PE. An advantage of spiral CT over VQ scanning and arteriography is the ability to define nonvascular structures such as lymphadenopathy, lung tumors, emphysema, and other parenchymal abnormalities as well as pleural and pericardial disease [51]. Another advantage of spiral CT over other diagnostic methods is the rapidity with which a study can be performed. Potential disadvantages of CT are that it is not portable at present and, because intravenous contrast is necessary, patients with significant renal insufficiency cannot be scanned without risk of renal failure.

Echocardiography, which can often be obtained more rapidly than either lung scanning or pulmonary arteriography, may reveal findings that strongly support hemodynamically significant pulmonary embolism [53]. Imaging or Doppler abnormalities of right ventricular size or function may suggest the diagnosis. Unfortunately, because these patients often have underlying cardiopulmonary disease such as chronic obstructive lung disease, neither right ventricular dilation nor hypokinesis can be used reliably even as indirect evidence of PE. In patients with documented acute PE, echocardiographic evidence of right ventricular dysfunction has been suggested as a means by which to determine the need for thrombolytic therapy [54]. Such cases need to be considered individually, and severe right ventricular dysfunction should lower the threshold for thrombolytic therapy once contraindications have been considered. Transesophageal echocardiography has also been evaluated in the setting of acute PE, and although it is less convenient, it may prove to have advantages over the transthoracic approach. (The role of echocardiography in acute and chronic pulmonary embolism is further discussed by Daniels et al in this issue). Intravascular ultrasound imaging has been shown in both experimental and clinical settings to image large emboli adequately and may be performed at the bedside [55]. Published guidelines suggest that clinicians be afforded a certain degree of flexibility in the diagnostic approach to suspected acute PE [56].

MRI has been used to evaluate clinically suspected PE, but at present the main advantage of MRI in this disease process is the excellent sensitivity and specificity for the diagnosis of DVT [41,52]. Disadvantages include the potential difficulty in transporting and performing the technique in critically ill patients. Additional prospective investigations will determine the role of this modality in the evaluation of VTE.

Treatment Options for treatment of acute DVT and PE include anticoagulation with low-molecular-weight heparin (LMWH) (Box 1) or standard heparin, thrombolytic therapy, and inferior vena cava filter placement. Massive PE is occasionally treated

358

V.F. Tapson / Cardiol Clin 22 (2004) 353–365

Box 1. Initiation of LMWH for therapy of acute DVT or PE Determine appropriateness of outpatient therapya  Begin subcutaneous administration of LMWHb.  Determine whether monitoring is needed (in patients with extremes of weight, renal insufficiency, or pregnancy).  Administer warfarin from day 1 at an initial dose of 5 to 10 mg, adjusted according to INR.  Check platelet count between days 3 and 5 for heparin-induced thrombocytopenia.  Stop LMWH after 5 or more days of combined therapy and when INR is 2.0 or more for 2 consecutive days.  Anticoagulate with warfarin for 3 months or longer (goal, INR 2.0–3.0)c. DVT, deep venous thrombosis; LMWH, low-molecular-weight heparin; INR, international normalized ratio; PE, pulmonary embolism. a Potential outpatients should be medically stable without severely symptomatic DVT. They should be compliant, capable of self-administration (or have a family member or visiting nurse for administration), and at low risk of bleeding; reimbursement should be addressed. b Enoxaparin (Lovenox) and tinzaparin (Innohep) are the two LMWHs that are approved by the Food And Drug Administration (FDA) for treatment of VTE. Although LMWH preparations are sometimes used for patients presenting with PE in the United States, and although clinical trials support this use, the FDA approvals read ‘‘established DVT with or without PE.’’ c The duration of warfarin therapy should be at least 6 to 12 months in patients with idiopathic venous thromboembolism.

with surgical embolectomy. Each approach has specific indications as well as advantages and disadvantages. Heparin and low-molecular-weight heparin The primary anticoagulants used to treat acute DVT or PE are unfractionated heparin and LMWH. These substances exert a prompt antithrombotic effect by accelerating the action of antithrombin III, thus preventing thrombus extension. Although they do not directly dissolve thrombus or emboli, they allow the fibrinolytic system to proceed unopposed and more readily reduce the size of the thromboembolic burden. Although thrombus growth can be prevented, early recurrence can sometimes develop even in the setting of therapeutic anticoagulation. LMWH preparations have substantial advantages over unfractionated heparin [57,58]. Because of these advantages, use of unfractionated heparin is becoming less common. When DVT or PE is diagnosed or strongly suspected, anticoagulation should be immediately instituted unless contraindications are present. Confirmatory diagnostic testing should be arranged as soon as possible. When treatment with standard, unfractionated, intravenous heparin is initiated, the activated partial thromboplastin time

(aPTT) should be followed at 6-hour intervals until it is consistently in the therapeutic range of 1.5 to 2.0 times control values. This range corresponds to a heparin level of 0.2 to 0.4 U/mL as measured by protamine sulfate titration. Achieving a therapeutic aPTT within 24 hours after the onset of treatment of PE has been shown to reduce the recurrence rate, and it has become evident that the traditional heparin regimen consisting of a 5000-U bolus and 1000 U/h is inadequate in many patients. Heparin is administered as an intravenous bolus of 5000 U followed by a maintenance dose of at least 30,000 to 40,000 U/24 hours by continuous infusion [59]. The lower dose is administered if the patient is considered at high risk for bleeding. This aggressive approach decreases the risk of subtherapeutic anticoagulation. It is possible that early initiation of warfarin without heparin or LMWH may intensify hypercoagulability and increase the clot burden because of the short-half life of anticoagulation factors (factors C and S) that are also inhibited by warfarin. Factor VII is the primary clotting factor affecting the prothrombin time and has a half-life of about 6 hours. Definitive anticoagulation requires the depletion of factor II (thrombin), which takes approximately 5 days. Thus, treatment with intravenous heparin or LMWH for at least 5 days is generally recommended. Heparin should be main-

V.F. Tapson / Cardiol Clin 22 (2004) 353–365

tained at a therapeutic level until two consecutive therapeutic international normalized ratio (INR) values of 2.0 to 3.0 have been documented at least 24 hours apart. The LMWH preparations have numerous advantages over unfractionated heparin and have dramatically changed treatment of thromboembolic disease. Among the differences between these two substances are the greater bioavailability of the LMWHs and more predictable dosing [57,58]. LMWHs can be subcutaneously administered once or twice daily even at therapeutic doses and do not require monitoring of the aPTT. Intravenous LMWH is not required in VTE. In addition, LMWHs have a more profound effect in inhibiting clotting factor Xa relative to thrombin. The reduced frequency of heparin-induced thrombocytopenia with LMWH relative to unfractionated heparin is a compelling reason to use LMWH instead of unfractionated heparin whenever possible. Because of their efficacy, safety, and convenience compared with standard heparin, these drugs are replacing standard heparin in many settings. A number of clinical trials and meta-analyses have strongly suggested the efficacy and safety of LMWH for treatment of established acute proximal DVT using recurrent symptomatic VTE as the primary outcome measure [60–65]. The incidence of DVT and recurrent bleeding in these trials indicates that LMWH preparations are at least as effective and as safe as unfractionated heparin. Meta-analytic data suggest that in the treatment of established proximal DVT the use of LMWH reduces bleeding rates and mortality compared with unfractionated heparin [65]. Monitoring anti-factor Xa levels seems reasonable in certain settings such as in morbidly obese patients, very small patients (