Etiology and Pathogenesis

A thrombus—that is, a blood clot—is the material that travels to the pulmonary circulation in pulmonary thromboembolic disease. Other material also can travel via the vasculature to the pulmonary arteries, including tumor cells or fragments, fat, amniotic fluid, and a variety of foreign materials that can be introduced into the circulation. This text does not consider these other (much less common) types of embolism, which usually have quite different clinical presentations than thromboemboli.

In the majority of cases, the lower extremities are the source of thrombi that embolize to the lungs. Although these thrombi frequently originate in the veins of the calf, propagation of the clots to the veins of the thigh is necessary to produce sufficiently large thromboemboli that can obstruct major portions of the pulmonary vascular bed and become important clinically. Rarely do pulmonary emboli originate in the arms, pelvis, or right-sided chambers of the heart; combined, these sources probably account for less than 10% of all pulmonary emboli. However, not all thrombi resulting in embolic disease are clinically apparent. In fact, only about 50% of patients with pulmonary emboli have previous clinical evidence of venous thrombosis in the lower extremities or elsewhere.

Three factors are commonly cited as potential contributors to the genesis of venous thrombosis: (1) alteration in the mechanism of blood coagulation (i.e., hypercoagulability), (2) damage to the vessel wall endothelium, and (3) stasis or stagnation of blood flow. In practice, many specific risk factors for thromboemboli have been identified, including immobilization (e.g., bed rest, prolonged sitting during travel, immobilization of an extremity after fracture), the postoperative state, congestive heart failure, obesity, underlying carcinoma, pregnancy and the postpartum state, use of oral contraceptives, and chronic deep venous insufficiency. Patients at particularly high risk are those who had trauma or surgery related to the pelvis or lower extremities, especially hip fracture or hip or knee replacement.

A number of genetic predispositions to hypercoagulability are recognized. They include deficiency or abnormal function of proteins with antithrombotic activity (e.g., antithrombin III, protein C, protein S) or the presence of abnormal variants of some of the clotting factors that are part of the coagulation cascade, especially factor V and prothrombin (factor II). The most commonly found genetic defect that causes hypercoagulability is called factor V Leiden. It is usually due to a single base pair substitution leading to replacement of an arginine residue by glutamine, causing the factor V protein to become resistant to the action of activated protein C. Individuals who are heterozygous for factor V Leiden have a threefold to fivefold increased risk for venous thrombosis. The much less common homozygous state confers a significantly higher risk. In the genetic variant of prothrombin often called the prothrombin gene mutation, there is also a single base pair deletion that appears to affect posttranslational mRNA processing, leading to increased plasma levels of prothrombin and predisposition to venous thrombosis.

Whereas deficiencies of the antithrombotic proteins (antithrombin III, protein C, and protein S) are rare, factor V Leiden may have a prevalence of up to 5%. The fact that it is found in some 20% of patients with a first episode of venous thromboembolism suggests it is an important risk factor. Both factor V Leiden and the prothrombin gene mutation are relatively common in the Caucasian population but rare among African Americans and Asians in the United States.

Pathology

Pathologic changes that result from occlusion of a pulmonary artery depend to a large extent on the location of the occlusion and the presence of other disorders that compromise O₂ supply to the pulmonary parenchyma. There are two major consequences of vascular occlusion in the lung parenchyma distal to the site of occlusion. First, if minimal or no other O₂ supply reaches the parenchyma, either from the airways or the bronchial arterial circulation, frank necrosis of lung tissue (pulmonary infarction) will result. According to one estimate, only 10% to 15% of all pulmonary emboli result in pulmonary infarction. It is sometimes said that compromise of two of the three O₂ sources to the lung (pulmonary artery, bronchial artery, and alveolar gas) is necessary before infarction results. Second, when parenchymal integrity is maintained and infarction does not result, hemorrhage and edema often occur in lung tissue supplied by the occluded pulmonary artery. The name congestive atelectasis has sometimes been applied to this process of parenchymal hemorrhage and edema without infarction.

With either pulmonary infarction or congestive atelectasis, the pathologic process generally extends to the visceral pleural surface, so corresponding radiographic changes often are pleura-based. In some cases, pleural effusion also may result. As part of the natural history of infarction, there is generally contraction of the infarcted parenchyma and eventual formation of a scar. With congestive atelectasis but no infarction, resolution of the process and resorption of the blood may leave few or no pathologic sequelae.

In many cases, neither of these pathologic changes occurs, and relatively little alteration of the distal lung parenchyma is found, presumably because of incomplete occlusion or sufficient nutrient O₂ from other sources. Frequently, the thrombus quickly fragments or undergoes a process of lysis, with smaller fragments moving progressively distally in the pulmonary arterial circulation. Whether this rapid process of clot dissolution occurs is important in determining the pathologic consequences of pulmonary embolism.

With clots that do not fragment or lyse, generally a slower process of organization in the vessel wall and eventual recanalization are seen. Webs may form within the arterial lumen and sometimes are detected on a pulmonary arteriogram or on postmortem examination as the only evidence for prior embolic disease.

Pathophysiology

When a thrombus migrates to and lodges within a pulmonary vessel, a variety of consequences ensue. They relate not only to mechanical obstruction of one or more vessels but also to the secondary effects of various mediators released from the thrombus and ischemic tissue. The effects of mechanical occlusion of the vessels are discussed first, followed by a consideration of how chemical mediators contribute to the clinical effects.

When a vessel is occluded by an embolus and forward blood flow through the vessel stops, perfusion of pulmonary capillaries normally supplied by that vessel ceases. If ventilation to the corresponding alveoli continues, it is wasted and the region of lung serves as dead space. As discussed in Chapter 1, assuming that minute ventilation remains constant, increasing the dead space automatically decreases alveolar ventilation and hence CO₂ excretion. However, despite the potential for CO₂ retention in pulmonary embolic disease, hypercapnia is an unusual consequence of pulmonary embolism, mainly because patients routinely increase their minute ventilation after an embolism occurs and more than compensate for the increase in dead space. In fact, the usual consequence of a pulmonary embolus is hyperventilation and hypocapnia, not hypercapnia. However, if minute ventilation is fixed (e.g., in an unconscious or anesthetized patient whose ventilation is controlled by a mechanical ventilator), a PCO₂ rise may result from the increase in dead space caused by a relatively large pulmonary embolus.

In addition to creating an area of dead space, another potential consequence of mechanical occlusion of one or more vessels is an increase in pulmonary vascular resistance. As discussed in Chapter 12, the pulmonary vascular bed is capable of recruitment and distention of vessels. Not surprisingly, experimental evidence indicates no increase in resistance or pressure in the pulmonary vasculature until approximately 50% to 70% of the vascular bed is occluded. The experimental model is somewhat different from the clinical setting, however, because release of chemical mediators may cause vasoconstriction and additional compromise of the pulmonary vasculature.

With further limitation of the vascular bed by the combination of mechanical occlusion and the effects of chemical mediators, the pulmonary vascular resistance and pulmonary artery pressure may rise so high that the right ventricle cannot cope with the acute increase in afterload. As a result, the forward output of the right ventricle may diminish, blood pressure may fall, and the individual may have a syncopal (fainting) episode or go into cardiogenic shock. In addition, “backward” failure of the right ventricle may occur, which manifests acutely with elevation of systemic venous pressure and appears on physical examination as distention of jugular veins.

The hemodynamic consequences of acute pulmonary embolism depend to a large extent on the presence of preexisting emboli and whether underlying pulmonary vascular disease or cardiac disease is present. When emboli have occurred previously, the right ventricular wall has already thickened (hypertrophied), and higher pressures can be generated and maintained. On the other hand, an additional embolus in an already compromised pulmonary vascular bed may act as “the straw that broke the camel’s back” and induce decompensation of the right ventricle.

In addition to the direct mechanical effects of vessel occlusion, thrombi result in release of chemical mediators that have secondary effects on both airways and blood vessels of the lung. Platelets that adhere to the thrombus are an important source of mediators such as histamine, serotonin, and prostaglandins. Injury to the pulmonary arterial endothelium by the clot results in increased release of endothelin-1, a potent vasoconstrictor, and decreased production of nitric oxide, a vasodilator. Bronchoconstriction, largely at the level of small airways, appears to be an important consequence of mediator release and is thought to contribute to the hypoxemia that commonly accompanies pulmonary embolism. In addition, areas of low ventilation and inappropriately high perfusion appear to develop because the process of hypoxic vasoconstriction becomes compromised by mediators related to thromboemboli. However, if vasoconstriction of pulmonary arteries and arterioles predominates, this adds to the likelihood of major cardiovascular compromise.

Three additional features of the pathophysiology of pulmonary embolism are noteworthy. First, as a result of vascular compromise to one or more regions of lung, synthesis of the surface-active material surfactant in the affected alveoli is compromised. Consequently, alveoli may be more likely to collapse, and liquid may more likely leak into alveolar spaces. Second, hypocapnia appears to have the effect of inducing secondary bronchoconstriction of small airways. With the hypocapnia that occurs in pulmonary embolism, and particularly with the low alveolar PCO₂ in dead space regions of lung, secondary bronchoconstriction results. Both of these mechanisms, along with the small airway constriction induced by chemical mediators, may contribute to the volume loss or atelectasis frequently observed on chest radiographs of patients with pulmonary embolism. Shunt physiology may also contribute to hypoxemia because of either perfusion of atelectatic lung or elevated right heart pressures producing intracardiac shunting across a patent foramen ovale.

A variety of bioactive substances are inactivated in the lung (see Chapter 12). Whether pulmonary embolism disturbs some of these nonrespiratory metabolic functions of the lung is not clear, and whether clinical consequences might ensue from such a potential disturbance is unknown.

Clinical Features

Most frequently, pulmonary embolism develops in the setting of one of the risk factors previously mentioned. Commonly the embolus produces no significant symptoms, and the entire episode goes unnoticed by the patient and physician. When the patient does have symptoms, acute onset of dyspnea is the most frequent complaint. Less common is pleuritic chest pain or hemoptysis. Syncope is an occasional presentation, particularly in the setting of a massive embolism, defined as obstruction of two or more lobes (or an equivalent number of segments).

On physical examination, the most common findings are tachycardia and tachypnea. The chest examination may be entirely normal or reveal a variety of nonspecific findings, such as decreased air entry, localized crackles, or wheezing. With pulmonary infarction extending to the pleura, a pleural friction rub may be detected, as may findings of a pleural effusion. Cardiac examination may show evidence of acute right ventricular overload (i.e., acute cor pulmonale), in which case the pulmonic component of the second heart sound (P₂) is increased, a right-sided S₄ is heard, and a right ventricular heave may be present. If the right ventricle fails, a right-sided S₃ may be heard, and jugular veins may be distended. Examination of the lower extremities may reveal changes suggesting a thrombus, including tenderness, swelling, or a cord (palpable clot within a vessel). However, only a minority of patients (about 10%) with emboli arising from leg veins have clinical evidence of deep venous thrombosis, so the absence of these findings should not be surprising or overly reassuring.

Diagnostic Evaluation

Diagnosis of acute pulmonary embolism can be challenging, and the approach depends on the clinician’s level of suspicion, or pretest probability, of pulmonary embolism. For patients in whom the diagnosis is considered less likely, the clinician may start with a D-dimer assay (discussed later). Most patients who present with acute dyspnea or chest pain will have a chest radiograph and oxygen saturation checked by finger oximetry. The radiographic findings in acute pulmonary embolism are quite variable. Frequently the chest radiograph is normal. When it is not, the abnormalities often are nonspecific, including areas of atelectasis or elevation of a hemidiaphragm, indicating volume loss. Volume loss may be related to decreased ventilation to the involved area due to small airways constriction and possibly loss of surfactant. If pleuritic chest pain is present, the patient may try to prevent pain by breathing more shallowly, which contributes to atelectasis.

Occasionally the chest radiograph reveals a localized area of decreased lung vascular markings corresponding to the region where the vessel has been occluded. This finding is called the Westermark sign but often is difficult to read unless prior radiographs are available for comparison. With a large proximal embolus, enlargement of a pulmonary artery near the hilum occasionally occurs secondary to distention of the vessel by the clot itself. An apparent abrupt termination of the vessel may occur, although this usually is difficult to see on a plain chest radiograph.

Both congestive atelectasis and infarction may appear as an opacified region on the radiograph. Classically the density is shaped like a truncated cone, fanning out toward and reaching the pleural surface. This finding, called a Hampton hump, is relatively infrequent. Pleural effusions may be seen as an accompaniment of pulmonary embolic disease. Pleural effusions associated with pulmonary embolism may be either exudative or transudative and contain a variable number of red blood cells.

Arterial blood gas values typically show a widened alveolar-arterial difference in partial pressure of oxygen (PAO₂-PaO₂ [AaDO₂]), a low PaO₂, and respiratory alkalosis, with hypocapnia occurring in more than 80% of cases. Because PCO₂ is decreased, arterial PO₂ is higher than it would be if hyperventilation were not present. Occasionally, PO₂ is normal, so the presence of a normal PO₂ does not exclude the diagnosis of pulmonary embolism. Unfortunately, because of the variability in values, arterial blood gases are not very useful in determining the likelihood of pulmonary embolism.

Measurement of plasma D-dimer levels is commonly used as part of the diagnostic strategy for venous thromboembolic disease, especially in patients who are less likely to have a pulmonary embolism. D-dimer is a degradation product of cross-linked fibrin, and therefore levels are increased in the setting of thrombosis of any type. Plasma levels of D-dimer increase in the setting of venous thrombosis but are also increased in many other conditions, including trauma, surgery, cancer, and inflammation. Thus, D-dimer testing for venous thrombosis or pulmonary embolism is very sensitive, but the test is nonspecific. Previously, the interpretation of D-dimer testing was complicated by the use of different assays with varying sensitivity and specificity. However, the superior sensitivity and high negative predictive value of quantitative enzyme-linked immunosorbent assay (ELISA) have established it as the standard D-dimer test. When D-dimer measurement is used in clinical practice, normal levels are considered strong evidence against thromboembolic disease. Thus, a normal D-dimer level is helpful, especially in the patient with a low pretest probability of having a pulmonary embolus, but an elevated level is considered nonspecific and therefore nondiagnostic.

Recommendations for the best radiologic test to diagnose pulmonary embolism have been shifting in the last decade. Traditionally, the major screening test for pulmonary embolism has been the perfusion lung scan (described in Chapter 3), but contrast computed tomographic angiography (CTA) is increasingly used either instead of or in addition to perfusion lung scanning. Evaluation of the large veins in the lower extremities, typically using ultrasound techniques, is another commonly used diagnostic strategy. Identification of a clot in a vein above the popliteal fossa warrants the same treatment as a documented pulmonary embolus and often obviates the need for further evaluation.

A perfusion lung scan is performed by injecting radiolabeled macroaggregated albumin particles into a peripheral vein. In areas of normal blood flow in the lungs, the albumin particles lodge in a fraction of the small vessels that have been perfused. When blood flow is obstructed by a clot within the pulmonary arterial system, perfusion lung scanning demonstrates no labeled albumin and therefore absence of perfusion to the region of lung supplied by the occluded vessel (Fig. 13-1). If results of the scan are normal, pulmonary embolism is, for all practical purposes, excluded. However, abnormalities do not automatically indicate the presence of embolic disease. False-positive lung scans are common because local decreases in blood flow may result from primary disease of the parenchyma or the airways. A ventilation scan, which involves inhalation of a xenon radioisotope, is often added because if regions of decreased blood flow are secondary to airway disease, corresponding abnormalities should be seen on the ventilation scan. If a defect in perfusion is due to a pulmonary embolism, ventilation still will be present in the area, and the perfusion defect will be mismatched (i.e., will not have a corresponding ventilation defect). If parenchymal disease (e.g., pneumonia) is the cause of a perfusion defect, the corresponding abnormality should be seen on the chest radiograph.

Interpretation of the perfusion lung scan is a complicated process that depends on the clinical setting, results of the chest radiograph, and frequently the findings on a ventilation lung scan. Because the perfusion scan often is not definitive, a probability is placed on the likelihood of pulmonary embolism, taking into account the size and number of defects and the presence or absence of corresponding abnormalities on the radiograph and ventilation lung scan. The scan results are analyzed in conjunction with the pretest probability of pulmonary embolism, the term used to represent the clinician’s assessment of the likelihood of pulmonary embolism based on the patient’s clinical presentation.

When the lung scan is inconclusive, it is critical that additional diagnostic evaluation be performed. Different options for further workup are available, focusing either on the veins of the lower extremities or on the pulmonary vasculature itself. However, because lower extremity studies often are negative even in the presence of documented pulmonary embolism, a negative lower extremity study does not preclude the need for further evaluation of the pulmonary arteries if there is a reasonably high suspicion of pulmonary thromboembolism.

The technique of CTA (discussed in Chapter 3) has become the most commonly used modality in the diagnosis of pulmonary embolism (Fig. 13-2). CTA offers the significant advantage of high-quality visualization of the lung parenchyma, which is helpful in considering the likelihood of competing diagnoses. In addition, in many centers, CTA can be performed quickly and is more readily available than ventilation-perfusion scanning. However, the radiation dose associated with CTA, especially to the breast and chest, is significantly higher than with ventilation-perfusion scanning, so despite its advantages, CTA has not completely replaced perfusion scanning as a diagnostic modality. The clinician must weigh the risk and benefits for the individual patient as well as the practical issues of test availability and interpretation for each patient in whom the diagnosis of pulmonary embolism is being considered.

Long thought to be the gold standard for confirmation of pulmonary embolism, conventional pulmonary angiography is now rarely used for this indication (Fig. 13-3). In this technique, direct evaluation of the pulmonary arterial system to identify intraluminal thrombus is accomplished invasively by advancing a catheter through a jugular or femoral vein through the right heart and injecting radiographic contrast material directly into the pulmonary arteries. Compared with CTA, pulmonary angiography is less sensitive for pulmonary emboli, is invasive and more expensive, and is therefore not recommended in a standard evaluation. Although its lack of radiation or iodinated contrast exposure makes magnetic resonance imaging (MRI) seem like a potentially attractive option for diagnosing pulmonary emboli, technical aspects of the study make it currently much less useful and thus rarely considered for evaluation of pulmonary emboli.

Treatment

Standard treatment of a pulmonary embolus involves anticoagulant therapy, initially intravenous unfractionated heparin or subcutaneous low-molecular-weight heparin (LMWH), and then an oral coumarin derivative (warfarin), the latter usually given for at least 3 to 6 months. LMWH has a number of potential advantages over unfractionated heparin, including less risk of heparin-induced thrombocytopenia. In addition, in most cases it does not require frequent laboratory monitoring of coagulation tests to guide dosage adjustment, and it can be given subcutaneously in one or two daily doses, avoiding the need for continuous intravenous infusion. A number of newer anticoagulants are now available, including direct thrombin inhibitors (e.g., dabigatran and argatroban) and factor Xa inhibitors (e.g., fondaparinux and rivaroxaban). The best use of these costly newer agents in managing patients with venous thromboembolic disease is an area of active research.

By far the most important aspect of treating pulmonary embolism is achieving prompt and adequate anticoagulation. For the most part, after a patient with an embolus that has already entered the pulmonary circulation has sought medical attention, the biggest danger derives from a recurrent embolus that causes hemodynamic instability. For this reason, treatment in high-risk patients can be started before the diagnosis is confirmed.

The rationale for use of anticoagulants is to prevent formation of new thrombi or propagation of old thrombi (in the legs), not to dissolve clots that have already embolized to the lungs. As a result, there has been a great deal of interest in the use of thrombolytic agents such as tissue plasminogen activator (tPA), streptokinase, and urokinase for treatment of pulmonary emboli. These agents, which lyse recent blood clots, ideally are given within the first several days of the embolic event, but they may be effective even up to 2 weeks after the embolism. A difference in overall mortality rate has not been demonstrated when a thrombolytic agent versus heparin is used for initial therapy. However, there may be specific subgroups of patients more likely to benefit from thrombolytic therapy, namely patients with massive pulmonary embolism and hemodynamic compromise as a result of vascular occlusion. When one of these agents is used, the initial intravenous infusion of the thrombolytic agent is followed by standard anticoagulant therapy.

In some circumstances, treatment of pulmonary embolism involves placement of a filtering device into the inferior vena cava (IVC), with the goal of trapping thrombi from the lower extremities en route to the pulmonary circulation. This type of device, often called an IVC filter, is used most frequently if there are contraindications to anticoagulant therapy (e.g., bleeding problems), if thromboemboli have recurred despite adequate anticoagulation, or if the patient already has such limited pulmonary vascular reserve that an additional clot to the lungs likely would be fatal. If there is no contraindication, anticoagulation is continued after the device is in place to decrease the risk of clotting in the veins distal to the filter. Once in place, some newer filters can be removed once the need for the filter has resolved.

No discussion of the treatment of pulmonary embolism is complete without considering prophylactic methods used to prevent deep venous thrombosis in the high-risk patient. The most common forms of prophylaxis traditionally have been (1) external compression of the lower extremities with an intermittently inflating pneumatic device and (2) heparin administered subcutaneously in low dosage. LMWH is at least as effective and safe as low-dose unfractionated heparin. One or another method of prophylaxis is now generally used in patients about to undergo thoracic or abdominal surgery and in a variety of other high-risk patients who are on bed rest in the hospital. In the highest-risk patients (i.e., those with hip fracture or hip or knee replacement surgery), either LMWH or warfarin is more effective for prophylaxis than low-dose subcutaneous standard heparin.

References