Chapter 27. Pulmonary Embolic Disorders: Thrombus, Air, and Fat

• Pulmonary embolism (PE) is common, underdiagnosed, and lethal, yet readily treatable.
• Prophylaxis and accurate diagnosis are essential to improving outcome.
• The cause of death in PE is most often circulatory failure (acute cor pulmonale) due to right heart ischemia.
• There is no perfect diagnostic test for PE; accurate diagnosis requires both an informed clinical pretest probability and a stepwise application of tests including D-dimer, helical CT angiography, and lower extremity duplex.
• A careful risk assessment may identify patients ideal for outpatient therapy. Conversely, patients with hypotension, cancer, heart failure, hypoxemia, and present or prior deep vein thrombosis are at significantly higher risk for death, recurrence, or major bleeding from PE, and are best managed in an appropriately monitored setting.
• Low molecular weight heparin (LMWH) is approved and recommended as the initial therapy for PE, and should be used in most patients unless there exists a compelling reason to do otherwise. When LMWH is not used, unfractionated heparin is typically used to maintain the partial thromboplastin time at 1.5 to 2.5 times control. Numerous new anticoagulants are being tested and may soon be approved for the treatment of PE.
• Critically ill patients may especially benefit from aggressive use of vena caval interruption.
• Thrombolytic therapy is life-saving in patients with massive embolism and circulatory instability.
• Air and fat embolism usually present as acute respiratory distress syndrome, and are managed with mechanical ventilation, oxygen, and positive end-expiratory pressure.

This chapter will cover diseases involving embolism to the pulmonary circulation, including pulmonary thromboembolism, as well as the less common conditions of venous air embolism and fat embolism. Thromboembolism is predominantly an acute circulatory insult, with important but less dramatic consequences for gas exchange. In contrast, both air and fat embolism usually present as acute hypoxemic respiratory failure (AHRF). All three of these forms of embolism may cause acute right heart failure, which is more fully discussed in Chap. 26.

Pulmonary embolism (PE) is a dramatic and life-threatening complication of underlying deep venous thrombosis (DVT). Therefore, much of the management of PE is grounded in the prophylaxis, diagnosis, and treatment of DVT. Much of our knowledge about DVT and PE is derived from patients who are not critically ill. When generalizations regarding clinical manifestations, utility of diagnostic tests, and efficacy of therapeutic approaches are extrapolated to the critically ill population, it is with some risk.

Pulmonary thromboembolism is a common yet underdiagnosed illness which accounts for substantial morbidity and mortality. It was estimated nearly 25 years ago that 630,000 persons each year suffer PE in the United States alone, with nearly 200,000 deaths (Fig. 27-1).1 The incidence of PE over the past two decades may be decreasing2 but this is controversial.3,4 Recent hospital-based studies estimate an incidence of 1 case per 1000 persons per year, with 200,000 to 300,000 annual hospitalizations.5,6 Data on mortality from PE are less problematic; a recent analysis of all U.S. death certificates between 1979 and 1998 found that PE mortality has decreased substantially over the past 20 years, from over 35,000 deaths in 1979 to fewer than 25,000 deaths in 1998.7 Accurate estimates for its incidence and lethality are notoriously difficult to calculate, for reasons largely related to difficulties in diagnosis and data collection.3,8 However, there is general consensus that venous thromboembolism is underdiagnosed. Failure to diagnose PE is a serious management error since 30% of untreated patients die, while only 8% succumb with effective therapy.

ICU patients are not a homogenous group of patients and the incidence of thromboembolism ranges widely across subsets of patients (Table 27-1). For example, in one study of medical ICU patients, 33 percent were found to have deep venous thrombosis, despite prophylaxis in more than half of them.9 Among patients with DVT, half of them had thrombi in the proximal lower extremity, a site associated with a high risk of embolization. In addition, 15% of thrombi were catheter-associated thrombi in the upper extremities. One prospective study found that 33% of ICU patients with a central venous catheter had a sonographically detectable thrombus in the internal jugular or subclavian vein; of note, none of these patients were symptomatic, and none developed a symptomatic pulmonary embolus.10 In trying to devise prophylactic and diagnostic plans, intensivists should take into account the patient mix in their own ICUs.

In spite of great strides in the understanding of venous thromboembolism, PE continues to cause substantial morbidity and mortality. Critically ill patients form a unique and challenging subset of those at risk. The presence of indwelling lines and forced immobility make these patients particularly susceptible to venous thromboemboli. Diagnosis, which is difficult even in ambulatory patients, is further impeded by barriers to communication and physical examination. Moreover, alternate explanations for hypoxemia, lung infiltrates, respiratory failure, and hemodynamic instability are readily available, such that a diagnosis of pulmonary thromboembolism may not be seriously considered. Finally, critically ill patients are likely to have limited cardiopulmonary reserve, so that pulmonary emboli may be particularly lethal.

Pathophysiology

Venous thrombosis begins with the formation of microthrombi at a site of venous stasis or injury. Thrombosis impedes flow and generates further vascular injury, favoring progressive clot formation. In some patients, clot becomes substantial and propagates to a proximal vein where it has the potential to embolize to the pulmonary circulation. Most clinically relevant pulmonary emboli originate as proximal venous thrombi in the leg or pelvic veins. However, in the ICU, the routine placement of upper body catheters for vascular access, monitoring, drug administration, and nutrition raises the likelihood of important upper body sources of thrombi. How significantly these line-related upper body thrombi contribute to PE in critically ill patients is unknown, but they clearly have embolic potential.15 Such upper extremity thrombi have important implications for diagnostic strategies based on detection of lower extremity thrombi and therapies such as inferior vena caval interruption which presume a lower extremity source.

PE occurs when thrombi detach and are carried through the great veins to the pulmonary circulation. Pulmonary vascular occlusion has important physiologic consequences which lead to the manifestations of illness as well as to clues to diagnosis. PE has an impact most notably on gas exchange and the circulation.

Gas Exchange

Physical obstruction to pulmonary artery (PA) flow creates dead space in the segments served by the affected arteries. This creation of dead space has several effects on the partial pressure of carbon dioxide (PCO2) and end-tidal CO2 (ETCO2), which can provide clues to diagnosis. If minute ventilation (V̇e) does not change, as occurs in a mechanically ventilated, muscle-relaxed patient, PCO2 will rise. However, most patients augment V̇e more than necessary to maintain elimination of CO2, so that PCO2 typically falls with PE. In health, the ETCO2 is nearly the same as arterial CO2. After pulmonary embolization, since end-tidal gas is a mixture of alveolar gas (in which the partial pressure of carbon dioxide [PaCO2] approximates PaCO2) as well as the newly created physiologic dead space gas (in which PCO2 approximates inspired PCO2, or is nearly zero), ETCO2 falls in proportion to the degree of dead space and no longer approximates PaCO2 (Fig. 27-2). This principle of a fixed alveolar-to-arterial gradient for PCO2 has been used to distinguish acute exacerbations of chronic obstructive pulmonary disease (COPD) from pulmonary embolism in patients with acute ventilatory failure.16 While not yet studied in the critically ill population, the steady-state end-tidal alveolar dead space fraction—which can be easily derived once one has both an accurate PaCO2 and end-tidal pressure of carbon dioxide (ETCO2)—has a sensitivity of 79.5% and a negative predictive value of 90.7% in hospitalized patients with PE.17

A widened alveolar-to-arterial gradient for oxygen [(a-a)PO2] is present in the majority of patients with PE. However, since in PE hyperventilation is the rule, PaO2 may not be low. In fact, only 63% of patients with proven PE demonstrate a PaO2 <70.18 Therefore, a normal PaO2 does not conclusively exclude a diagnosis of PE. There have been several efforts to use various combinations of the PaO2, the PCO2, and the (a-a)PO2 to predict the likelihood of (or adequately exclude) pulmonary embolism, but these have all been shown to have insufficient discriminant value, even in patients without antecedent cardiopulmonary disease.19

The mechanisms of hypoxemia have been elucidated by applying the multiple inert gas elimination technique (MIGET) to patients with PE.20,21 Shunt is found in only a few patients. In some, this may be due to opening of a probe patent foramen ovale when right atrial pressure rises following PE (see below), with a consequent intracardiac right-to-left shunt. Since shunt is more likely to be found when patients are studied after some delay, atelectasis due to impaired surfactant production may contribute. In most patients, however, the most important pulmonary derangement is mismatching of ventilation and perfusion. In addition, the fall in cardiac output (Q̇t) that accompanies most pulmonary emboli leads to a fall in mixed venous saturation. This lowered venous saturation magnifies any hypoxemia due to shunt or ventilation-perfusion (V̇/Q̇) mismatch (Fig. 27-3).

While impaired oxygenation is important and often provides a clue to the diagnosis of PE, it is typically responsive to modest oxygen enrichment of inspired gas. Severe hypoxemia is usually seen only in patients with profound shock. If oxygen-refractory hypoxemia is present without obvious hypoperfusion, a patent foramen ovale should be suspected. Ventilatory failure is usually not a significant problem either, since most patients are able to double or triple V̇e if necessary, in order to maintain a normal (or reduced) PCO2. In fact, the greatest impact of PE is on the circulation, not on gas exchange. Although there are exceptions (e.g., some patients with severe COPD), morbidity and mortality from PE relate to cardiovascular compromise, not respiratory failure.

Circulation

PE obstructs the pulmonary vascular bed mechanically as well as via humoral and other mechanisms. This increases right ventricular afterload, which when compounded by tachycardia, increases right ventricular oxygen consumption. The right ventricle dilates and thins, its wall tension rises, and coronary perfusion is impeded. At the same time, pulmonary vascular obstruction hinders Q̇t and contributes to hypoxemia. Therefore, just at the time when the right ventricle demands an increase in oxygen delivery, the left ventricle may not be able to supply it. The superimposition of increased right heart oxygen demand on decreased oxygen supply puts the right ventricle at risk for ischemia, precipitating failure of the right heart (cor pulmonale). Acute right heart dysfunction is the likely cause for sudden death in patients with massive PE. There are substantial animal and human data supporting this hypothesis, which is further discussed in Chapter 26.

Of particular interest to the intensivist is the patient with a sublethal, yet large, PE. With increased afterload, the right ventricle dilates to a larger end-diastolic volume. This is associated with elevated right atrial and ventricular pressures and abnormally low Q̇t. One consequence of raised right atrial pressure is the potential for right-to-left shunting across a probe patent foramen ovale, causing oxygen-refractory hypoxemia. This could allow paradoxical embolization of thrombus to the systemic arteries. The increase in right ventricular pressure and volume also affects the left heart. A change in shape of the right ventricle and corresponding shift of the interventricular septum from right to left alters the diastolic pressure-volume characteristics of the left ventricle (Fig. 27-4) and may be detected echocardiographically. The consequent reduction in left ventricular compliance impairs diastolic filling, thereby reducing left ventricular preload and creating yet another obstacle to cardiac output. An alternative explanation for the apparent decrease in left ventricular performance is reduced preload due to a fall in transmural filling pressure. Pericardial constraint might explain the fact that the left ventricular end-diastolic pressure (or wedge pressure) does not fall.22

History, Examination, and Laboratory Data

Most patients with PE will complain of dyspnea, chest pain, and apprehension.23 Less common symptoms are cough, diaphoresis, and hemoptysis. It is notable that most patients will not have symptoms of venous thrombosis. Syncope is uncommon but described in all large series of PE (Table 27-2).

The majority of patients will demonstrate tachypnea and tachycardia. Pleural rub and signs of DVT are seen only occasionally. Fever is more common than is generally appreciated and seen in half of patients,23 but only rarely is the temperature >38.5°C. Patients with large emboli may have the typical findings of any patient with low output shock such as hypotension, narrow pulse pressure, and poor peripheral perfusion. Occasionally, unanticipated failure to come off mechanical ventilation or unexplained episodes of respiratory distress may be hints to a diagnosis of PE. Rare patients with PE have disseminated intravascular coagulation, systemic embolization, or acute respiratory distress syndrome (ARDS) as their presenting manifestation.

Most patients will demonstrate hypoxemia or at least a widened (a-a)PO2. However, since a small but significant fraction will have normal oxygenation,18 the blood gas value should not dissuade a physician from considering the diagnosis when the rest of the clinical picture is suggestive. In fact, as described above, the blood gas does not provide data that are useful in discriminating patients with PE from those without. The chest x-ray may demonstrate areas of oligemia (Westermark's sign) and rare patients will develop a pleural based, truncated cone (Hampton's hump); however, most films have only nonspecific findings. In fact, the greatest utility of the chest film is in making (or excluding) alternative diagnoses such as pneumonia, pneumothorax, or aortic dissection. Nevertheless, the typical (albeit nonspecific) findings of basilar atelectasis, elevation of the diaphragm, and pleural effusion should always suggest PE when there is no apparent alternative explanation. Electrocardiography (ECG) may reveal signs of right heart strain such as rightward axis shift, right bundle-branch block, or right precordial strain, but often shows only sinus tachycardia.

Signs from More Invasive Monitoring

Valuable signs of PE may come from many of the devices used to monitor critically ill patients. The intensivist may derive clues from the ventilator, expired gas analysis, the PA catheter, or during echocardiography. The sensitivity and specificity of these monitors for the diagnosis of PE are not known. Nevertheless, wise physicians attempt to incorporate all available data into their synthesis of the patient. This is especially important in the ICU, where the history and physical examination are often difficult or impossible to obtain.

The Ventilator

To maintain PCO2, the patient with PE must augment the V̇e. Therefore, any unexplained increase in V̇e should prompt consideration of PE. Of course, any cause of rising dead space (airflow obstruction, hypovolemia, or positive end-expiratory pressure [PEEP]) or CO2 production (anxiety, pain, fever, or sepsis) will also increase V̇e. However, when none of these conditions is apparent, especially when supporting clues are evident, PE becomes more likely.

Expired Carbon Dioxide

As described above, the increment in dead space after PE causes a detectable fall in ETCO2. With technologic improvements in ventilators, noninvasive assessment of expired CO2 is becoming increasingly practical in the ICU. A corollary of the fall in ETCO2 with PE is that if V̇e does not rise (e.g., in a muscle-relaxed patient), the total excretion of CO2 (expired CO2 concentration × V̇e) must fall. Therefore PaCO2 will rise progressively until a new steady state is reached at a higher PaCO2 . This can be demonstrated numerically by the alveolar dead-space fraction (AVDSf) as follows:

When combined with a negative D-dimer value, as will be discussed shortly, an AVDSf of less than 0.15 has been shown to exclude PE in hospitalized patients with a sensitivity of 97.8% and a negative predictive value of 98%.17 In calculating the AVDSf, however, one must take care to ensure proper calibration of the blood gas analyzer, as even small changes in PaCO2 measurements will cause large differences in AVDSf. Again, in the muscle-relaxed patient there are many explanations for a rising arterial CO2, but if no explanation is forthcoming, especially if the CO2 production is known to be constant, consideration of PE is warranted.

Pulmonary Artery Catheter

The most obvious clues from the pulmonary artery catheter (PAC) are the elevations in right atrial, right ventricular, and PA pressures and concomitant fall in Q̇t that occur with PE. With the reduced Q̇t, one also sees a widening of the arterial-to-venous oxygen content difference [(a-v)PO2] (Fick principle) and a decrement in the mixed venous oxygen saturation (Sv̅O2) or the central venous oxygen saturation. An oximetric catheter may facilitate early recognition of the fall in venous oxygenation. A final clue from the PAC may lie in the difference between the PA diastolic pressure and the pulmonary capillary wedge pressure (Ppw). Normally, flow through the pulmonary circulation is pulsatile, so that by the end of diastole, there is no more flow from the PA to the left atrium. Without flow, there can be no pressure gradient from the PA to the left atrium. Thus the end-diastolic PA pressure and the Ppw are nearly equal. However, when there is obstruction of the pulmonary vascular bed, flow is not completed by the end of diastole and a pressure gradient remains. A discrepancy between the PA diastolic pressure and Ppw may provide a clue to PA obstruction.24

Unfortunately, each of these observations is certainly nonspecific (and likely insensitive) so that only rarely do such changes indicate PE. For example, cardiac dysfunction (systolic or diastolic) causes a rise in right heart pressures and a fall in Q̇t; any cause of low Q̇t causes a widened (a-v)PO2; and any cause of acute lung injury may raise the PA diastolic-to-Ppw gradient. A further layer of complexity is added by the recent observation in a trial of surgical patients that pulmonary embolism was disproportionately associated with the PAC compared with central venous catheters (0.8% vs. 0%).25 Given the limitation of the PAC as a diagnostic tool and the risk of actually causing PE, it cannot be advocated for the diagnosis of PE.

Echocardiography

The role of echocardiography in the evaluation of suspected PE is evolving, but seems most helpful in determining a patient's prognosis. While its attractions include portability—especially for the evaluation of critically ill patients—noninvasiveness, potential to elucidate competing diagnoses (such as myocardial infarction or pericardial disease), and rapid availability, echocardiography is insensitive, and should not be used to diagnose PE. In prospective trials of unselected patients, sensitivities of between 29% and 52% are reported for various echocardiographic criteria of right ventricular strain or dysfunction or tricuspid regurgitation.26,27 Many patients with PE have normal echocardiograms. Occasionally, a study requested for evaluation of a low flow state may unexpectedly reveal findings strongly suggestive of PE.28,29 These include a dilated, thin-walled, poorly contracting right ventricle, and bowing of the interventricular septum to the left. Very rarely, echocardiography may demonstrate a thrombus in the right atrium or right ventricle (Table 27-3), clinching the diagnosis of PE.

When echocardiography exhibits right ventricular dysfunction, it reliably predicts an increased risk of mortality from pulmonary embolism. One study examined 126 patients with PE with echocardiography on the day of diagnosis, and found moderate RV dysfunction to impart a sixfold increased risk of in-hospital death compared to normal RV function.30 Even in patients assessed to be hemodynamically stable at presentation, right ventricular dysfunction portends a worse prognosis; one study found that 10% of such patients develop shock and 5% died in the hospital, compared to a 0% mortality among patients with normal RV function.31 In another series of hemodynamically stable patients, recurrent embolism was strongly associated with baseline echocardiographic abnormalities in right ventricular wall motion.32 A word of caution is prudent, however, in that the classic echocardiographic findings of PE are nonspecific, and are common to a number of causes of acute right ventricular pressure overload such as ARDS, other forms of severe hypoxemia, or status asthmaticus (see Chap. 26).

Special Problems in the ICU

The typical critically ill patient is unable to complain of the usual symptoms of PE, has numerous explanations for tachycardia and tachypnea, is hemodynamically unstable, and is a poor candidate for transport for radiographic studies. For that reason, it is important to have a clear sense of the probability of PE in any given patient. Such a judgment is complex, and validated algorithms for determining prior probability in critically ill patients are not available. Synthesis of the patient's cardiopulmonary physiology, combined with an assessment of risk factors, is all the clinician has to go on. Obviously, there is no substitute for experience in managing these very difficult patients. In the following sections, the contribution of various tests in evaluating suspected PE is discussed. An approach to diagnosis is summarized in Fig. 27-5.

Risk Factors

Since the symptoms, signs, and laboratory findings of PE are usually nonspecific, delay before pursuing a diagnosis in a patient with classic, unmistakable clues risks missing this potentially lethal disease. However, since nonspecific indicators of potential PE are ubiquitous, indiscriminant pursuit of the diagnosis is prohibitively costly and dangerous. Most patients with PE have identifiable risk factors (Table 27-4). Absence of such risk factors should lead the physician to seek alternative explanations for the patient's findings other than thromboembolism. On the other hand, when numerous risk factors are present, the diagnosis should be more seriously considered.

Diagnostic Tests

The gold standard for the diagnosis of PE has long been the pulmonary angiogram. However, its use in critically ill patients is limited by invasiveness, expense, need for dye infusion, and the risks attendant to transport out of the ICU. In many patients, the pulmonary angiogram can be replaced by its noninvasive cousin, the helical computed tomography (CT) angiogram. Helical CT angiography is rapidly replacing both the ventilation-perfusion (V̇/Q̇) lung scan and pulmonary angiography in the diagnosis of PE. The limitations of pulmonary angiography and V̇/Q̇ scanning, combined with residual doubts regarding the reliability of helical CT angiography, have led to attempts to supplant or augment these tests with alternatives. These include noninvasive leg studies, MRI imaging of the thorax, and various blood tests such as the D-dimer assay. An integrated approach to the diagnosis of PE is described in Figure 27-5.

Computed Tomography and Magnetic Resonance Imaging

Advances in CT and MR imaging have rekindled great interest in the potential for these relatively noninvasive studies to replace the V̇/Q̇ scan and pulmonary angiogram as diagnostic procedures of choice.

Magnetic resonance imaging and angiography (MRA) technology has progressed greatly in the past 10 years, with rapidly improving resolution and speed of image attainment. MR does not require iodinated contrast, making it ideal for patients with renal insufficiency, or for those who should not be exposed to ionizing radiation. Among initial reports in 1997, one study listed a sensitivity of 100% and specificity of 95% when MRA was performed in 30 patients undergoing pulmonary angiography for suspected PE.33 Three of the 22 “negative” MRAs were also termed “poor quality,” yet were considered negative readings for calculation of sensitivity and specificity. A larger (118 patients) study with an arguably less selected group of patients found the sensitivity of MRA to be 77% but with a wide 95% confidence interval; specificity was better, at 98%.34 In both studies, MRA proved difficult either due to patient factors that made them ineligible for magnetic resonance (9% to 15% of eligible patients), or due to poor image quality (6% to 9%), which made the study uninterpretable. Thus while promising, MRA has not yet attained a sensitivity that allows one to rule out PE.

CT scanning has received even more attention and is widely used. Helical CT angiography produces a two-dimensional image of the lung and its vessels at very small collimator, or slice thickness. It has been shown to detect central emboli—out to fourth-division vessels—with a high degree of reliability. Like MRI, helical CT angiography is minimally invasive and may provide alternative diagnoses in the work-up of PE. It does, however, require the injection of intravenous contrast, it is expensive (though less so than either MR or pulmonary angiography) to perform and interpret, and it can be difficult to perform in patients who are unable to hold their breath or are hemodynamically unstable. The most significant criticisms of this technique have been its inability to detect emboli to the level of subsegmental arteries and its variation in interobserver interpretation.

In a prospective trial of 42 patients with suspected PE who underwent helical CT angiography followed by pulmonary angiogram, all 23 patients with a normal CT were found to have a negative pulmonary angiogram, for a sensitivity of 100%.35 The test likewise appeared to be extremely specific, with only one false-positive, for a reported specificity of 96%. However, with further studies, a range of sensitivity emerges of 53% to 100%, and of specificity between 81% and 100%. Two well-performed recent reviews36,37 of the CT data came to similar conclusions: that the available literature on CT in the diagnosis of PE remains limited by small studies with incomplete use of a reference standard; that the sensitivity and specificity of CT in diagnosing PE are as yet undetermined; and that the safety of withholding anticoagulation based on a negative CT study is untested and unknown. Both studies raised concerns that interobserver variation and the potential inability of CT to reliably determine subsegmental PEs might preclude its usefulness as a diagnostic test. It is interesting to recall that in PIOPED, subsegmental PEs accounted for 6% of the PEs detected, and that the interobserver agreement for pulmonary angiograms among these small PEs was wider than that of larger PEs (66% for subsegmental PEs compared to 90% for segmental and 98% for lobar PEs).38 It would appear that subsegmental PEs are challenging to diagnose even by pulmonary angiography, and that the clinical significance of such clots remains difficult to ascertain.

One recent study may shed some light on the safety of withholding anticoagulation following negative CT scan.39 One hundred seventeen patients in whom PE was suspected but who were found to have a negative helical CT angiogram were followed for a mean duration of 21 months. The study population was elderly (mean = 65 years) and with frequent comorbidity; 70% of patients had known cardiac or respiratory disease. The recurrence rate for PE was 4.5%, based on three patients who died of unknown cause within 1 month of their initial suspected PEs, and two cases of repeat presentation with suspected PE which was confirmed by imaging. One of the patients who died was receiving anticoagulation for severe cardiac insufficiency; excluding his death yields a recurrence rate of 4.9%. If one were to consider the three unexplained deaths—two were older than 90 years with significant comorbidities—as not being due to recurrent PE, the recurrence rate would fall to 1.8%. For pulmonary angiogram, the reported rate of recurrence of PE following a negative study ranges from 0.6% to 4.9%. Thus it may be that helical CT angiography, while by no means a perfect test, has a PE recurrence rate similar to that of pulmonary angiography following a negative study, and is therefore adequate grounds for withholding anticoagulation. Further investigations to confirm or contrast the above findings are necessary to validate this claim.

Both CT and MRA have the advantage of being able to diagnose alternative conditions that may explain the patient's symptoms. Both are less invasive and less expensive than pulmonary angiography, and are probably easier than V̇/Q̇ scan to perform in a critically ill patient. Yet lacking comprehensive prospective studies using angiography as a gold standard—lacking PIOPED—both MRA and CT may remain suspect to some practitioners.

Noninvasive Leg Studies

Noninvasive leg studies include impedance plethysmography (IPG), phleborheography, venous Doppler, and B-mode ultrasound scanning of leg veins. The technical details of these procedures and differences between them are beyond the scope of this chapter, but which particular test to choose is largely a function of local expertise. Both IPG and venous ultrasonography are extremely helpful in assessing a patient with symptomatic proximal deep vein thrombosis. In this population, across numerous studies venous ultrasonography has a sensitivity of 97%, a positive predictive value (PPV) of 100%, and a negative predictive value (NPV) of 100%;40 IPG was similar though slightly inferior, with a sensitivity of 92%, PPV of 78%, and NPV of 96%.40 Falsely abnormal results on IPG can be seen with nonthrombotic compression of the vein, increased central venous pressure, and chronic obstructive pulmonary disease.41 The troublesome aspect to both of these imaging modalities is their performance in asymptomatic patients, particularly asymptomatic patients with high clinical probability of thromboembolism. Patients with symptomatic DVT are far more likely to have a proximal than a distal, or isolated calf vein, DVT.42,43 In contrast, among asymptomatic patients screened for thromboses, the majority—almost two thirds—are distal.44,45 When the natural history of postoperative deep vein thrombosis was studied in 1969, approximately 20% of distal DVTs subsequently extended into the proximal veins, and while nonextending distal DVT caused no clinical cases of pulmonary embolism, 4 of 9 cases with proximal DVT did cause PE.45 Subsequently, studies have observed that distal clots which subsequently progress tend to do so within a week of presentation,46 and that proximal extension of distal DVT after more than a week is unusual,40 spurring the evaluation of serial noninvasive leg studies for the diagnosis of proximal DVT. In evaluating the safety of withholding anticoagulation following a negative test, both IPG and venous duplex appear to have similar rates of subsequent DVT (1.5% to 2%) during the 6 months following a negative test.40 However, pooling data from combined studies of IPG, 4 of 1625 patients (0.25%) with an initial negative IPG died from PE during serial testing or follow-up, whereas one of 1747 patients (0.06%) with initial negative venous ultrasound died of PE.40

Precisely because PE more commonly follows asymptomatic, rather than symptomatic, DVT,38 noninvasive leg studies cannot make or exclude the diagnosis of PE. Moreover, the ultrasonic characteristics of any thrombus appear to predict poorly the likelihood of associated PE.47 Nevertheless, the wide availability of accurate noninvasive leg studies provides a simple method for managing many patients who cannot be transported or who have nondiagnostic helical CT angiography. In patients with angiographically proven PE, 43% to 57% will have detectable DVT.48,49 In some patients, diagnostic testing can be avoided by performing serial noninvasive leg studies. Demonstration of venous thrombosis provides grounds for treatment (usually anticoagulation), so the question of PE may cease to be important. On the other hand, failure to detect DVT may sufficiently reassure the intensivist that neither further diagnostic measures nor empirical treatment is urgently needed. The rationale for this approach is that if deep venous thrombosis cannot be detected, the likelihood of re-embolism is very low. Furthermore, if deep venous thrombosis develops, or a subclinical thrombosis propagates to the proximal deep veins, serial testing will reveal it before it has much chance to embolize. In ambulatory patients, this strategy is effective (Table 27-5). It is important to note that this approach has not been validated in critically ill patients and does not take into account the role of upper extremity thrombosis.

Ventilation-Perfusion Lung Scan

V̇/Q̇ scanning traditionally was the initial test of choice in the evaluation of PE, but two important studies have led to a sobering reappraisal of this test.38,48 V̇/Q̇ scans can be extremely helpful to the clinician when they provide either a high probability result—with an attendant specificity of 85%, confirming the diagnosis—or a normal result, when the diagnosis of PE is virtually excluded.38 Although 15 of 100 patients with “high probability” V̇/Q̇ scans will not have PEs, the risk of treating them is felt to be less than the risk and cost of performing pulmonary angiography in all 100 patients. The frustration with V̇/Q̇ scanning stems from the large number of tests which yield either intermediate probability or indeterminate results. Scans of intermediate probability indicate a substantial likelihood of PE (about 40%) necessitating further evaluation to prove or exclude the diagnosis. Among patients with low probability scans, the prevalence of PE ranges from 16% to 40%, necessitating further diagnostic work-up in all but those deemed to have a low clinical likelihood of PE, among whom only 4% of patients had PE. Thus the V̇/Q̇ scan results, which are sufficient in themselves to terminate the evaluation (because PE is either highly likely or has been reliably excluded) include all normal scans, all high probability scans, and low probability scans in patients with a low clinical pretest probability of PE. Unfortunately, a minority of patients fall into these three categories, so that for the majority of patients undergoing V̇/Q̇ scan, it prompts further testing.

How useful the ventilation-perfusion scan is when applied to the subset of critically ill patients remains an open question. In a retrospective analysis of the PIOPED database, 223 patients were defined to be “critically ill” (room air PaO2 <50 [n = 89]; mechanically ventilated [n = 46]; hypotensive, but not ventilated [n = 3]; or physically in an ICU, but not ventilated [n = 85]).53 While it is difficult to know precisely how ill these patients were (no severity of illness scores were reported), only 29 of 223 had a high probability result, and no critically ill patients had normal or near normal scans. Sixteen critically ill patients had the combination of low clinical pretest probability and low probability scan. Thus even in this group of ICU patients presumably stable enough to be enrolled in a clinical trial and in whom the scan could be performed, the test was definitive in only 20%. Although the investigators concluded that the sensitivity and positive predictive value of the V̇/Q̇ scan were similar in “critically ill” and “non–critically ill” patients, it is clear that the V̇/Q̇ scan is not helpful in the majority of patients. Given the surging popularity of CT scanning, combined with ever better CT and MR detectors, there seems to be little role for perfusion or ventilation-perfusion scanning in critically ill patients.

Other Noninvasive Means of Diagnosis

The most recent potentially useful blood test is the D-dimer assay. Very low levels of D-dimer argue against a diagnosis of PE, with a sensitivity of 95% and a negative likelihood ratio of 0.9 to 0.13 using the widely available quantitative rapid enzyme-linked immunosorbent assay (ELISA).54 Essential to the use of a D-dimer test in diagnosing patients with thromboembolic disease is an assessment of clinical pretest probability; one clinical prediction rule for DVT found a prevalence of DVT of 6%, 9%, and 28% in patients whose clinical prediction rule was low, intermediate, or high, respectively.55 Converting the prevalence to odds, and applying the D-dimer test with a 0.9 negative likelihood ratio, posttest probability is as low as 0.006 for the initial low-probability patients, whereas it is an order of magnitude greater (0.03%) in the high-probability patients. If one were to imagine extremely high-risk patients, with a pretest probability of 80%, the posttest probability of PE could be as high as 0.28 despite a negative D-dimer result. D-dimer assays can be considered unidirectional, in that a negative result can be extremely useful, yet a positive result has little effect on the likelihood of either DVT or PE.54 This fact unfortunately limits the utility of the D-dimer test in the critically ill population, in whom so many patients will have a positive D-dimer level. For example, in a retrospective analysis of patients with severe sepsis randomized to undergo aspirin versus placebo, 100% of patients had an elevated D-dimer.56 On the other hand, when D-dimer testing was applied to patients with cancer suspected of having deep vein thrombosis, it proved to be an effective test, with a 98% sensitivity, a 97% negative predictive value, and a negative likelihood ratio of 0.03.57 Despite concern that the high prevalence of thromboembolic disease in a population with cancer would limit the utility of D-dimer, the test performed well and allowed outpatient evaluation in 29% of the patients studied.

The combination of pulse oximetry and static compliance of the respiratory system yielded a very high sensitivity and specificity for PE in critically ill trauma patients.58 In patients with COPD, capnography during a prolonged passive exhalation had a negative predictive value of 100% when values of arterial PCO2 and end-maximal expiratory PCO2 were nearly equal.16 Combining steady-state end-tidal alveolar dead space fraction and D-dimer was also quite sensitive in diagnosing PE in hospitalized, though not critically ill, patients, with a sensitivity and negative predictive value of 98%.17 The proper place of each of these tests depends on further study.

Pulmonary Angiography

Pulmonary angiography is considered the definitive test for the diagnosis of PE. Positive findings include an intraluminal filling defect or a cutoff of a 2-mm or larger vessel seen in more than one view. Experienced radiologists agree on 98% of studies showing lobar embolism.59 However, agreement falls to 90% with segmental embolism, and only 66% in those with subsegmental clots. Thus even this test has limitations. Its sensitivity is sufficiently high to occasionally detect embolism even weeks to months after the acute episode. The earliest documented resolution of an angiogram to normal following a pulmonary embolus is 1 week, although faster resolution has been seen in animal studies.60 Thus this test can reliably exclude embolism if a negative study is obtained even a week following the onset of symptoms.61 Since it is invasive, costly, riskier than CT angiography, and requires the presence of an interventional radiologist (rather than a CT technician), it is usually reserved for patients in whom the diagnosis cannot be made or excluded by less invasive means. However, pulmonary angiography is safer than generally appreciated. Mortality is around 0.2% as shown in several large series.62,63 Several patients who have died had severe pulmonary hypertension and cor pulmonale at the time of the procedure, leading some to conclude that pulmonary hypertension is a contraindication to pulmonary angiography. The degree of risk in such patients was directly addressed in one review of pulmonary angiography.63 Elevated pulmonary systolic pressure (greater than 70 mm Hg) and elevated right ventricular diastolic pressure (greater than 20 mm Hg) were identified as risk factors. However, even in these sickest patients the mortality was only 2%. In patients with severe chronic pulmonary hypertension who are being considered for surgical endarterectomy (admittedly a different population) angiography has been performed quite safely.64 Nonionic, low osmolality contrast agents improve image quality and may be able to reduce risk even further.

Empirical Diagnosis

Occasionally, an empirical diagnosis of PE seems clear cut to the managing physician. No alternative diagnoses may seem plausible, or further diagnostic steps seem risky or unnecessary. Although this approach may appear attractive, it rarely proves correct. First, the clinical diagnosis is simply too unreliable, even for the most experienced clinician. Second, one can always expand the differential diagnosis to include alternative causes. Finally, the doubt that lingers after an empirical diagnosis too frequently haunts subsequent management. Progression of symptoms or signs despite therapy raises questions about failure of treatment or the need for alternative treatments. Complications of treatment (usually hemorrhage due to heparin, occasionally thrombocytopenia) create uncertainty about the necessity of the toxic therapy, or precipitate more diagnostic interventions in a newly unstable state. Since critically ill patients are more likely to have complications of therapy, empiricism is rarely appropriate.

The majority of patients with PE will not die from the clot that leads to diagnosis. As long as re-embolization is prevented, the patient will survive, while intrinsic fibrinolysis restores pulmonary blood flow. Therefore the primary goal of all therapies for PE is to prevent re-embolization. Some patients, however, survive the initial embolus yet remain in shock. These patients, who are overrepresented in ICU populations, may succumb to the initial embolus. Additional therapy to hasten clot resolution, aimed at more promptly restoring the circulation, is useful in such patients. Supportive care, anticoagulation, vena caval interruption, thrombolysis, fluid and vasoactive drug administration, and rarely surgical embolectomy all must be considered in the treatment of this disease. An integrated approach to the treatment of PE is presented in Figure 27-6.

Risk Assessment

Having made a diagnosis of PE, the clinician and patient face numerous potential therapies and outcomes. Pulmonary embolism is spectacularly inconsistent in its clinical presentation, and can range from asymptomatic or mildly symptomatic dyspnea to profound shock due to right ventricular dysfunction. Several characteristics of each presentation can allow the clinician to identify patients with the poorest prognosis, who almost certainly benefit from close observation in a monitored setting, and who may benefit from a more aggressive therapeutic approach. Equally important, the clinician may also identify those patients at low risk for complications, in whom a strategy of anticoagulation alone, potentially as an outpatient, will suffice. The Geneva prognostic index, generated from a prospective study of 296 patients with PE admitted through the emergency room, identified six predictors of adverse outcome, defined as death, recurrent thrombotic event, or major bleeding.65 Hypotension imparted an odds ratio of 15; cancer of 9.5; and prior DVT, DVT by ultrasound, heart failure, and hypoxemia odds ratios of between 2.6 and 3.8. Cardiac biomarkers have also become potential prognostic indicators; when elevated, both troponin T and troponin I portend a high risk for complications, and troponin I was significantly associated with increased overall mortality in one study.66 Low brain natriuretic peptide (BNP) was positively associated with survival, and high BNP with higher mortality and PE-specific mortality in another study of 110 normotensive patients with PE.67 As previously discussed, echocardiographic evidence of right ventricular dysfunction clearly identifies a subgroup of patients with PE at high risk of shock or death.31 To date no data have shown a decreased mortality or decreased recurrence of thromboembolic events by adopting a more aggressive treatment strategy for patients deemed to be high risk, and we do not advocate basing therapeutic decisions, such as whether or not to use thrombolytic medications, on such criteria. However, it seems prudent to use a careful risk assessment of death, recurrence, or major bleeding when contemplating which patients may benefit from outpatient therapy, treatment on the general ward, or surveillance in an ICU.

Supportive Care

Oxygen and Bed Rest

Patients typically present with hypoxemia, which responds well to oxygen therapy since the underlying pathophysiology is usually V̇/Q̇ mismatch. Bed rest, once advocated as the standard of care in treating venous thromboembolic disease, has recently been called into question by two randomized prospective studies.68,69 Both studies found that allowing patients with DVT to ambulate on day 2—compared to imposed strict bed rest for between 4 and 10 days—failed to increase the incidence of PE as detected by V̇/Q̇ scanning. Ambulating patients wore thigh-length compression stockings and walked up to 4 hours per day. Admittedly, the applicability of these studies to an ICU population may be problematic, as patients were excluded if they had clinically overt PE, free floating thrombus, pregnancy, renal insufficiency, or were unable to ambulate. Many of our ICU patients, and all patients with symptomatic PE, would have failed to qualify for these studies. Bed rest may have advantages to the critically ill patient beyond the theoretical advantage of reducing clot dislodgment. By reducing oxygen consumption (V̇O2) and thus Q̇t to their absolute minimum, bed rest—combined with sedation and mechanical ventilation in selected patients—is indicated in the treatment of patients with PE and shock.

Specific Therapies

Anticoagulation

Unfractionated Heparin (UFH)

Heparin has long been the mainstay of therapy for PE, although its pre-eminent role in treating thromboembolic disease has lately been challenged by newer medications. Unfractionated heparin is a mixture of acidic glycosaminoglycans typically extracted from porcine intestinal mucosa, with a variable molecular weight of between 5000 and 30,000 da, depending on its clinical preparation.70,71 Along with a coumarin derivative, it was the first anticoagulant to be prospectively shown to decrease mortality and recurrent PE, decreasing mortality by 25%.72 While it has proven effective, it requires monitoring of coagulation parameters as well as dose adjustment due to unpredictable plasma levels even within individuals. Although a recent study has shown that UFH given subcutaneously is as effective and safe as fixed dose nadroparin,73 UFH is generally given by continuous intravenous infusion and we expect this practice to continue. These limitations functionally restrict the use of heparin to the hospital setting. Heparin essentially catalyzes the effect of antithrombin to rapidly inhibit several of the factors from the intrinsic and common coagulation pathways, factors Xa, IXa, XIa, and XIIa.70 It also inhibits the activation of factors V and VIII by thrombin.71 Heparin is cleared rapidly from the plasma by binding to cell surface receptors on endothelial cells and reticuloendothelial elements,70 and its clearance is not affected by renal or hepatic insufficiency.

In dosing heparin, a singular concept emerges: one must give enough heparin to surpass a minimal level of anticoagulation in order to prevent further thromboembolism. This threshold level appears to be doses at which the activated partial thromboplastin time (aPTT) is at least 1.5 times control74 (or alternatively at which a heparin level of 0.2 to 0.4 U/mL by the protamine sulfate assay or 0.3 to 0.6 IU/mL by amidolytic anti-Xa study is reached), and it should be reached as quickly as possible. When continuous intravenous heparin was compared with intermittent subcutaneous heparin in the treatment of proximal deep vein thrombosis, only 1 of 61 patients who achieved an aPTT of greater than 1.5 times control suffered recurrent venous thromboembolism (VTE). In contrast, 13 of 53 patients whose aPTT was below 1.5 times control for 24 hours had recurrence, independent of the method of administration of heparin (relative risk = 15:1).74 Rapid, adequate anticoagulation is facilitated by a weight-based nomogram.73,75 Heparin is typically initiated with a bolus of 80 U/kg followed by a continuous infusion of 18 U/kg per hour. For most patients, oral anticoagulation (coumadin) is begun within 3 days of initiating heparin.

While the evidence supporting a lower therapeutic limit for heparin of 1.5 times control as judged by the aPTT is quite strong, the conventional upper limit (2.5 times control) is arbitrary. For years it had been assumed that the risk of hemorrhage was significantly related to the level of the aPTT, but one trial has cast doubt on this assumption.76 In this study, hemorrhagic complications were related to underlying predisposing factors (eg, recent surgery, ulcer disease, or cancer) rather than to supratherapeutic levels of the aPTT. Patients were initially treated with heparin plus warfarin or with heparin alone. Of 99 patients in the combined treatment group, 69 (69%) had aPTT levels greater than 2.5 times control for more than 24 hours. In contrast, in the heparin only group, 24 of 100 patients (24%) had supratherapeutic aPTT levels. Bleeding complications were similar in the two groups (combined therapy, 9.1%; heparin alone, 12.0%). In addition, bleeding complications were seen in only 8 of 93 patients with supratherapeutic aPTT regardless of therapeutic arm, compared with 13 of 106 patients without such elevation. These findings refute an association between modestly elevated aPTT and hemorrhage, and combined with the importance of prompt, adequate anticoagulation, suggest the value of an approach that aims to assure enough heparin in the first hours of treatment, rather than to avoid too much.

Complications of Heparin

Complications of heparin, in addition to hemorrhage, include heparin-induced thrombocytopenia (HIT), osteoporosis, hypersensitivity, and (rarely) hyperkalemia. The most important complication of heparin is bleeding. In a pooled analysis of its use in deep venous thrombosis, major bleeding (bleeding greater than 1 L, bleeding requiring blood transfusion, or intracerebral bleeding) was reported in 2 of 59 patients (3%).77 In a group of 121 patients given heparin for all indications, 8% developed major hemorrhage (fatal, life-threatening, potentially life-threatening, or bleeding leading to reoperation or requiring at least three units of blood).78 When heparin dosing is guided by nomogram as described above, the risk of hemorrhage was 9% to 12%.76 Hemorrhage typically occurs from the gastrointestinal or urinary tract, or from surgical incisions. Less common sites of serious bleeding include the retroperitoneum, adrenal glands, soft tissues, nose, and pleural space. Intracranial hemorrhage is uncommon in patients anticoagulated with heparin.

The approach to treatment of the patient who bleeds on heparin depends on the severity of bleeding. When bleeding is minor, simply stopping the heparin may be sufficient. Bleeding related to needlesticks may respond to sustained direct pressure. When hemorrhage endangers life or organ function, a more aggressive approach is mandatory. Transfusion of fresh frozen plasma is usually ineffective since circulating heparin inhibits the function of transfused factors. Protamine sulfate is an antidote to heparin. The dose of protamine depends on heparin levels, and is therefore related to dose, route of administration, and time since the last dose. When hemorrhage immediately follows a bolus of heparin, sufficient protamine to completely neutralize the heparin (1 mg protamine per 100 U heparin) should be administered. In the more usual situation in which heparin therapy is ongoing, the dose of protamine should be based on the approximate half-life of heparin (90 minutes). Since protamine is also an anticoagulant, the dose should be calculated to only half-correct the estimated circulating heparin. Protamine has been known to cause hypotension, shock, dyspnea, and pulmonary hypertension upon intravenous injection. The incidence is reduced by giving the drug very slowly (no more than 50 mg in 10 minutes). When protamine is administered, a physician should be present in case of an anaphylactoid reaction. Alternatives for the treatment of underlying conditions should be considered once heparin is stopped. For example, the patient with pulmonary embolization should have vena caval interruption with a percutaneous filter (see below).

Thrombocytopenia is a relatively common complication of heparin administration, typically occurring after several days of therapy. In large series, roughly 1% to 3% of patients given full-dose intravenous heparin have developed thrombocytopenia.79 Interestingly, in our experience at a tertiary care center where many of our patients are repeatedly exposed to heparin, the incidence of heparin-induced thrombocytopenia seems to be much higher. The mechanism in most patients is thought to be related to specific IgG directed against heparin. Thrombocytopenia is less commonly seen with highly purified low-molecular-weight heparins, suggesting that higher-molecular-weight components may be responsible. It is also rare in patients given prophylactic, minidose heparin, occurring in only one of 348 patients.80 The onset of thrombocytopenia is usually on the third to fifteenth day (mean = day 10), but it can occur after several hours in patients previously sensitized. The severity of thrombocytopenia is variable (commonly to 50,000/mm3), but can be severe (<5000/mm3). Most patients remain asymptomatic, but some suffer major arterial or venous thrombosis, or life-threatening hemorrhage. Rarely, this syndrome is associated with skin bullae that progress to necrosis (“heparin necrosis”). These may occur at sites of injection or at distant sites. Any time heparin is given, it is prudent to measure platelet counts on a daily basis. An otherwise unexplained drop in platelet count of 30% has been suggested as the threshold that should prompt discontinuation of heparin,80 although there are few data upon which to base this. Heparin-dependent IgG should be sought in the patient's serum, since a positive result may simplify management decisions. None of the available assays are fully sensitive, so a negative test is no guarantee against recurrent thrombocytopenia or thrombosis if the patient is rechallenged with heparin.

Low Molecular Weight Heparin

The promise of low molecular weight heparins (LMWHs) is a simple, once or twice daily therapy for venous thromboembolism that does not require monitoring of anticoagulant effect. As a class, the LMWHs encompass a number of fragments of unfractionated heparin which are on average 5000 da in molecular weight. The different LMWHs have been depolymerized by various means, and thus differ in their pharmacokinetic and pharmacodynamic properties.70 One LMWH may not be interchangeable with another. LMWHs achieve excellent bioavailability when given subcutaneously (about 90% of that achieved with an equal intravenous dose), have a long half-life (2 to 4.4 hours), correlate well between anticoagulant response and body weight, and have equal or better antithrombotic effects than unfractionated heparin. There have now been numerous trials comparing LMWHs (subcutaneously or intravenously) with unfractionated heparin in both DVT and PE.81–85 In an early meta-analysis pooling the results of over 2000 patients, LMWHs appeared superior with regard to venographic score improvement, recurrence rate of VTE, and hemorrhage. There is also a trend towards an all-cause mortality benefit for LMWHs.81 Similarly, a recent meta-analysis of LMWH compared to unfractionated heparin for the treatment of PE found non–statistically-significant decreases in recurrent symptomatic VTE, PE, and bleeding complications.86 With such data, groups such as the American College of Chest Physicians now recommend that clinicians use LMW heparin over unfractionated heparin.71

Aspects of caring for critically ill patients may limit the full implementation of such recommendations in the ICU. Many patients with critical illness have coincident renal failure, bleeding diatheses, or need for ongoing invasive procedures, all of which make unfractionated heparin, which can be quickly discontinued and rapidly cleared from the body, preferable. Conversely, certain situations appear to be ideally suited to the use of LMWH, such as in the long-term treatment of cancer patients with venous thromboembolism.87,88 As stated, the incidence of HIT is lower with LMWH than with unfractionated heparin, although in many trials of LMWH for DVT or PE, the incidence of HIT was sufficiently low in both arms to preclude a statistical comparison. In orthopedic and surgical studies as well as the initial industry-sponsored trials of various LMWHs, the incidence of HIT varied between 0% and 0.8% of patients, compared with approximately 3% of patients treated with unfractionated heparin.89 However, antibodies from patients with previous HIT cross-react with every commercially available LMWH, therefore LMWHs cannot be safely used in a patient known to have HIT.89

Alternatives to Heparin

Neither heparin nor low molecular weight heparin can be used in patients with HIT; furthermore, no safe dosing regimen has been approved for LMWH in the presence of renal failure. Current alternatives to heparin include selective factor Xa inhibitors, direct thrombin inhibitors, and oral anticoagulants.

Direct thrombin inhibitors include the hirudin family, first isolated from leeches, and the argatroban family of active-site inhibitors. The advantage of such agents over unfractionated heparin is their ability to inhibit fibrin-bound thrombin with a predictable dose response, and their inability to produce HIT.70 Hirudin and its derivatives, especially lepirudin and bivalirudin, are potent anticoagulants which have been proven more effective than enoxaparin, a LMWH, in DVT prophylaxis of patients undergoing hip replacement.90 The risk of life-threatening bleeds appears to be equal between hirudin and heparin,90,91 although an increased risk of moderate bleeding was observed with hirudin.91 aPTT must be monitored with hirudin agents (goal aPTT = 1.5 to 2.5 times control). Dosages must be adjusted for renal insufficiency. The agents are approved for use in patients with HIT, but they have not been extensively studied in prospective comparison to heparin or LMWH in treatment of pulmonary embolism.

Additional direct thrombin inhibitors are the class represented by argatroban, melagatran, and the new oral agent ximelagatran. Argatroban is an effective anticoagulant in patients with HIT, and has been licensed for use in this capacity.70 Like the hirudins, argatroban has not been studied in any randomized trials in PE, but it appears to be as safe or safer than heparin when given to patients with myocardial infarcation.92 It is cleared principally by the biliary system and the dose needs no adjustment for renal failure. Monitoring of both the aPTT and the prothrombin time (PT) is recommended. Ximelagatran, an oral prodrug converted to melagatran, does not require monitoring. Having been proved favorable in preventing DVT following total knee replacement compared to both coumadin93 and enoxaparin,94 ximelagatran was tested for secondary prevention of DVT. In this study, the oral drug was compared to warfarin therapy for patients who had already received 6 months of oral anticoagulation with coumadin.95 Recurrent thromboembolism was significantly reduced in the ximelagatran group (2.8% vs. 12.6% in the warfarin group, p <0.001), with similar rates of major bleeding. Three cases of fatal pulmonary embolism were diagnosed in the warfarin group, compared to none in the ximelagatran group, though overall mortality was equivalent between groups. Ximelagatran has been shown to cause elevated liver aminotransferases, especially in the first 4 months of therapy; in this study, all cases of elevated alanine aminotransferase were transient, and no case (out of 600) progressed to hepatic failure.95

Among the newest of antithrombotic agents approved for the prophylaxis of DVT and pulmonary embolus is fondaparinux, a novel synthetic drug inhibiting factor Xa. By binding to antithrombin with tight affinity and inducing a conformational change, fondaparinux facilitates the binding of antithrombin with factor Xa, thus blocking the common pathway of coagulation.70 It has been approved by the FDA as a subcutaneous injection of 2.5 mg once daily for use in DVT and PE prophylaxis in patients undergoing orthopedic surgery for hip fracture, hip replacement, or knee replacement, after several randomized controlled trials found superior reduction of DVT for fondaparinux compared to enoxaparin.96 A study comparing fondaparinux and heparin for pulmonary embolism found the new agent to be as effective and safe as heparin, with a nonsignificant trend toward fewer recurrences of VTE in the fondaparinux group.97 Bleeding events were similar, and no difference in mortality was detected. HIT has not been observed with fondaparinux use. A number of additional anti–factor Xa agents are in development, but they have not yet been widely tested in humans. While fondaparinux shows promise in the acute treatment of PE, more studies are necessary before recommending that factor Xa inhibition replace our traditional anticoagulant strategies.

Warfarin anticoagulation should be instituted in the first few days (5 mg PO qhs for the first 2 days, then adjust the dose to achieve an International Normalized Ratio [INR] between 2.0 and 3.0) and continued to overlap with heparin therapy for at least 5 days. If the patient is clinically unstable or likely to require ongoing invasive procedures, heparin or a rapidly adjustable agent such as lepirudin or argatroban should be used preferentially, to facilitate rapid adjustment of anticoagulation if necessary.

Vena Caval Interruption

Occasionally patients have compelling reasons that preclude anticoagulation, or continue to embolize despite adequate anticoagulation. As most thromboemboli originate in the legs, pelvis, or inferior vena cava (IVC),98 inferior vena caval interruption (VCI) has the potential to prevent subsequent embolization. Conventional indications for the use of VCI in patients with venous thromboembolism have included contraindications to anticoagulation, hemorrhage following anticoagulation, failure of anticoagulation to prevent recurrent embolization, and prophylaxis of extremely high-risk patients. Only one randomized prospective trial has evaluated using VCI to prevent PE.99 This trial randomized 400 patients with proximal DVT who were at risk for PE to receive a vena caval filter or standard anticoagulation. Over the first 12 days, the vena caval filter group had significantly fewer cases of PE (1.1% vs. 4.8%, p <0.03). However, over the subsequent 2 years, during which all patients were anticoagulated for at least 3 months, the rates of PE, major bleeding, and death were similar in both groups. Moreover, the vena caval filter group actually had a significantly higher rate of recurrent DVT, with a nearly twofold increase in odds ratio for DVT compared to the no-filter group. Thus VCI appeared to be a successful short-term strategy to prevent PE, but came at the expense of increased long-term DVT and a trend toward increased thromboembolic disease. This landmark study prompted widespread discussion on two fronts: whether anticoagulation should be considered for all patients receiving a vena caval filter, and whether a temporary or retrievable filter might be more successful.100 Unfortunately, no prospective trial evaluates either question. Progress has been made in the technology of retrievable filters, and small case series of several different types of temporary filters have been published.101,102 Whereas initially, temporary filters had to be removed within a 2-week period in order to minimize the risk of endothelialization, newer filters have been retrieved after up to 134 days without complication.101

As VCI devices have evolved, they have become smaller and more easily placed. As physicians have become more experienced with them, they have become safer as well, although published complication rates have varied widely, from 2% to 19%.101 Proposed indications for VCI continue to expand as the technology improves,100 yet caval filters are probably underutilized in the sickest patients. Those with critical cardiopulmonary compromise may be least able to tolerate the admittedly low risk of heparin failure. Patients who survive sublethal embolism but remain hemodynamically compromised may benefit from VCI since recurrent embolism, while unlikely, will be fatal. In addition, some patients have such a predictably high risk of venous thromboembolism that VCI may be indicated even before any thrombosis is detected. For example, we have placed caval filters prophylactically in patients with severe COPD and proximal venous thrombosis and in a woman with Crohn's disease complicated by toxic megacolon and ARDS on her way to the operating room. Additional potential indications include extensive free-floating vena caval thrombus (since the efficacy of heparin in this setting is controversial), pre–pulmonary embolectomy, and as primary treatment in patients with malignancy (since anticoagulants seem less effective and less safe in this setting). The advent of safe, effective retrievable filters may greatly expand our use of VCI.

Complications include filter fracture (occasionally with embolization of fragments), improper placement of the filter, venous thrombosis at the insertion site (seen in 8% to 25%103), caval occlusion (which is now far less common than when the Mobin-Uddin “umbrella” filters were in use), inadvertent dislodgment by guidewires during central venous catheterization, and erosion or perforation of the caval wall and other viscera. A final point about VCI is that while it prevents most recurrent emboli, it does not treat the (presumed) leg source. Therefore, when these devices are used, concomitant anticoagulation is necessary (unless the indication for VCI is contraindication to anticoagulation). The significance of upper extremity sources of thromboemboli in ICU patients, who often have central catheters, is unknown but probably small. Nevertheless, inferior VCI would clearly be ineffective so that when the clinical presentation suggests an upper source, due consideration should precede placement of this device. Filters have been placed in the superior vena cava and in the suprarenal IVC with apparent safety, but the experience is limited.

Thrombolytic Therapy

It has been clear for more than three decades that thrombolytic therapy more rapidly lyses pulmonary emboli and improves hemodynamics.104 However, 7 days after therapy, patients given thrombolytic therapy cannot be distinguished from those treated only with heparin on the basis of clinical findings or lung perfusion.105 Combined with the clearly increased risk and cost of thrombolytic agents, this has led to a general skepticism about the value of thrombolytic therapy for the treatment of PE. In a survey of pulmonary physicians, only 11% would give thrombolytic therapy for a large PE without hypotension, severe hypoxemia, or right ventricular strain.106 However, nearly all would give thrombolytics to a patient with shock due to PE, and there is now convincing data to support this approach. For example, when patients with massive PE were given alteplase (1 mg/kg over 1 hour) in an uncontrolled trial, there was significant clinical improvement in 11 of 15 with shock.107

In another study comparing alteplase (100 mg over 2 hours) and heparin, right ventricular function assessed echocardiographically improved more rapidly in patients given alteplase.108 This more rapid hemodynamic improvement could translate into improved survival in critically compromised patients. Furthermore, of the 55 patients treated with heparin alone, 5 suffered recurrence (2 fatal) within 14 days, whereas none of 46 patients given alteplase recurred. Finally, one very small (eight patients) randomized clinical trial comparing thrombolysis (streptokinase, 1,500,000 U over 1 hour) with heparin in patients with massive embolism and hypoperfusion showed a mortality benefit for thrombolytic therapy; all four patients receiving heparin alone died, while there were no deaths in the streptokinase group.109

Recently, the debate over thrombolytic therapy has moved from discussion of massive PE to that of submassive PE: PE which has not yet provoked hypotension, but for which the risk for death is high. Data from echocardiographic studies show an indisputably higher mortality in patients with moderate RV dysfunction.30 The utility of echocardiography in prognostication has led to great interest as to whether echocardiographic criteria should be used routinely to guide the use of thrombolytic therapy.110,111 Two studies address this question. In a retrospective cohort study of patients with radiographically large emboli and right ventricular dilation (but without hypotension or shock), those who received thrombolytic therapy had better pulmonary perfusion on day one, but this advantage disappeared by day seven.112 Mortality was 6% in the thrombolytic group compared to no deaths in the heparin group, and severe bleeding in the thrombolytic group approached 10%, with intracranial bleeds in 4% of patients. In a second, prospective study, patients with acute PE and right ventricular dysfunction (but without shock) were randomized to heparin plus alteplase or heparin plus placebo.113 While there was no difference in mortality between groups, the alteplase group had a significantly lower rate of treatment escalation, mostly consisting of secondary thrombolysis for worsened symptoms. No significant difference in bleeding was found between groups, which was surprising, as previous studies have never failed to show an increased risk of major hemorrhage with thrombolytic therapy compared to heparin. In a large prospective registry of over 2400 patients with PE, the rate of major hemorrhage in patients receiving thrombolytics was almost 22%, and intracranial bleeds were noted in 3%.114 The retrospective, nonrandomized nature of the first study raises concerns for its validity, but the latter has also been cited as potentially flawed in its design, both in its definition of “right heart dysfunction” and its use of secondary thrombolysis based on vague clinical criteria.115 As no mortality benefit has yet been established for patients with submassive PE who receive thrombolysis, we advocate restricting thrombolytic therapy for those patients with clinically apparent shock.

The optimal regimen for thrombolytic therapy has not been established. A number of approaches are listed in Table 27-6.104,107,109,116–118 There has been a general trend towards bolus, as opposed to long-duration, infusion, based on both animal studies and human trials in patients with myocardial infarction showing increased efficacy and reduced hemorrhage. Yet the increasing benefits of front-loaded regimens has a limit as demonstrated in a comparison of alteplase, 0.6 mg/kg (maximum dose = 50 mg) over 15 minutes versus 100 mg over 2 hours, in which there was no significant difference in efficacy or rate of hemorrhage.119–121 The particular thrombolytic agent is probably not of much importance. In a head-to-head comparison of alteplase (100 mg over 2 hours) and urokinase (1,000,000 U over 10 minutes followed by 2,000,000 over 110 minutes), the two regimens yielded similar efficacy and safety.122 A powerful case can be made that streptokinase is the agent of choice since it is the one agent for which a mortality benefit has been shown,109 and it is fourfold cheaper than any of the current alternatives. Careful attention should be given to selecting patients appropriately to reduce the rate of hemorrhagic complications (Table 27-7). Especially important is a concerted effort to avoid invasive procedures, including arterial blood gases, arterial catheters, central venous punctures, and pulmonary angiograms, where possible.123

Various measures of the lytic state correlate poorly with both efficacy and incidence of bleeding, so that outside of clinical research protocols, routine monitoring is not indicated. When streptokinase is given, the manufacturer recommends that the thrombin time be assayed at 4 hours to ensure that a lytic state is achieved. An adequate lytic state can be assumed if the thrombin time is prolonged above the normal limits of the laboratory, or if the fibrinogen level is reduced. Clinical monitoring should include serial neurologic examinations to detect central nervous system hemorrhage and frequent vital signs to detect gastrointestinal or retroperitoneal hemorrhage. Patients who have undergone catheterization should have the groin puncture examined, and preferably have repeated measurements of thigh girth. Huge volumes of blood can be lost into the thigh and groin, especially in obese patients, with little external evidence of bleeding.

After the thrombolytic agent is discontinued, heparin is typically begun (without a bolus) when the thrombin time or the aPTT falls to less than 2 times control. Heparin is begun as an intravenous infusion at 1300 U/h and titrated to a PTT of 1.5 to 2.5 times control.

No advantage to infusing the thrombolytic agent directly into the pulmonary artery has been shown in controlled trials despite claims of benefit in small case series.124 Investigational approaches include pharmacomechanical thrombolysis125 and augmentation of cardiac output to enhance thrombolysis.126

Thrombolytic Therapy Following Surgery

Although recent surgery is generally included among the absolute contraindications to thrombolytic therapy, there is an evolving literature supporting its use. For example, 13 patients with angiographically confirmed embolism within 2 weeks of major surgery (mean = 9.6 days) were given a modified regimen of urokinase (2200 U/kg directly into the clot, followed by 2200 U/kg per hour for up to 24 hours).127 Complete lysis was achieved in all and there were no deaths or bleeding complications. In another report, two patients in shock due to PE were given bolus regimens of urokinase (1,200,000 U) or alteplase (40 mg, followed by another 40 mg over 1 hour) only 2 days after lung resection.128 There was prompt clinical improvement, although one patient had delayed hemorrhage. Finally, nine patients were treated with urokinase (1,000,000 U over 10 minutes, followed by 2,000,000 U over 110 minutes) following neurosurgery (mean = 19 days following surgery).129 All of the patients survived their acute episode of PE and no intracranial hemorrhage occurred, although one patient developed a subgaleal hematoma. These reports suggest that recent surgery is a relative, not an absolute, contraindication, and that the risks and benefits should be considered on an individual basis.

Complications

The greatest limitation of the thrombolytic drugs, and the factor which has limited their acceptance for the treatment of venous thromboembolism, is the consequential incidence of bleeding. In patients treated for pulmonary embolism the risk of major hemorrhage is reported to be around 15%,107 but these data were gathered in an era of frequent pulmonary angiography. As mentioned, intracranial bleeds have been observed in as many as 4.7% of patients,112 although larger series report a 3% incidence.114 When serious bleeding occurs, the lytic agent should be immediately discontinued, and reliable, multiple, large-bore catheters secured. Direct compression of bleeding vessels may stop or slow ongoing blood loss. If heparin has been given, it too should be stopped and consideration given to reversing heparinization with protamine. Most patients will be adequately managed without the transfusion of clotting factors. If it becomes necessary to reverse the lytic state, cryoprecipitate, which contains fibrinogen and factor VIII (both of which are consumed by plasmin) is the preferred blood product.130 The initial dose is 10 units, after which the fibrinogen level should be assayed. Fresh frozen plasma (as a source of factors V and VIII), platelets, and fibrinolytic drugs (e.g., epsilon aminocaproic acid 5 g over 30 minutes) all may play a role in the critically bleeding patient.

Allergic Effects

Allergic reactions, including skin rashes, fever, and hypotension are rare except with streptokinase. Mild reactions can be treated with antihistamines and acetaminophen. More severe reactions should prompt the addition of hydrocortisone. Hypotension usually responds to volume administration.

Fluid, Vasoactive Drugs, and Nitric Oxide

Volume administration with saline or colloid has generally been advocated in patients with PE and shock on the grounds that it will increase filling pressures and thereby augment Q̇t. However, in a patient with elevated right heart pressures and a grossly distended right ventricle, it is possible that further distention of the right ventricle during volume administration will increase myocardial oxygen consumption, yet fail to increase Q̇t and oxygen supply. In addition, to the extent that fluids increase right ventricular end-diastolic volume, the interventricular septum will bulge further to the left, impede left heart filling, and further compromise Q̇t. Experimental studies to determine the effect of fluids have shown a detrimental effect on hemodynamics.131 Therefore, volume administration should not be routine therapy unless the patient is clearly hypovolemic. When fluids are given, central venous catheterization (to measure venous oxyhemoglobin saturation and changes in right atrial pressure) and echocardiography may provide useful guidance. These issues are further discussed in Chapter 26.

There is also controversy regarding the use of vasoactive drugs to treat the hypoperfusion caused by PE. Successful use of vasoconstrictors, inotropes, and vasodilators has been reported. Since no controlled studies in patients have been performed, it is hard to give firm recommendations. However, the pathophysiology of this form of shock, the results of some animal experiments, and limited human data (all discussed in Chapter 26) provide some guidance. When any of these drugs are used, serial assessment of the effect of the intervention is mandatory. Any drug which does not result in the intended salutary effect should be discontinued promptly.

The vasoactive drug of choice, based on the largest published experience in patients, is dobutamine.132 Dobutamine is infused beginning at 5 μg/kg per minute and increased to effect. If dobutamine is ineffective or incompletely effective, norepinephrine should be tried. The rationale for the use of this vasoconstrictor is based on the assumption that right ventricular ischemia is the fundamental problem leading to shock. A vasoconstrictor that increases systemic arteriolar tone could raise aortic pressure and augment coronary blood flow, without increasing right ventricular load. In animal models of sublethal PE, norepinephrine was shown to be superior to no therapy, to volume administration, and to isoproterenol in the maintenance of cardiac output as well as in survival time.133 Infusion is initiated at 2 μg/min and adjusted (up to 30 μg/min) based on the hemodynamic response. In the clinical setting, hypoperfusion may have additional contributors such as left ventricular dysfunction or ischemic heart disease, so that a vasoconstrictor might be less beneficial than in controlled animal experiments. If dobutamine and norepinephrine fail to improve cardiac output, epinephrine may succeed.134 Finally, nitric oxide, which can lower the pulmonary artery pressure, boost cardiac output, and improve oxygenation, can be tried if available.135,136

Embolectomy and Mechanical Therapies

Surgical embolectomy is a major procedure rarely resorted to in most institutions. In part, this is related to the availability of other, more benign, therapies such as heparin and thrombolysis. Additionally, it takes time to organize a surgical team, operating room, cardiopulmonary bypass and so on, by which time the patient is often hemodynamically improved or moribund. Yet embolectomy has its advocates, who maintain that thrombolytic therapy is often contraindicated in patients who could benefit from it, the operative mortality for embolectomy is now acceptable, and chronic cor pulmonale can be averted.

In one institution's review of 87 patients with PE, 34 were treated with heparin, 28 with streptokinase, and 25 with embolectomy.137 Pretreatment embolic scores were most severe in the embolectomy group. Hospital mortality in the heparin, streptokinase, and surgery groups was 6%, 21%, and 20%, respectively. However, cumulative survival at 5 years was 68%, 64%, and 80%, a trend favoring embolectomy. However, most late deaths were due to malignancy, not recurrent PE or chronic pulmonary hypertension. Although the authors recommended surgical embolectomy for all patients with emboli in the main pulmonary arteries based on their results, regardless of the hemodynamic impact, this study was not randomized and the possibility seems large that the long-term benefit for embolectomy was related to selection of patients. A more recent trial showed that surgical embolectomy was comparable to thrombolytic therapy in patients with massive PE.138

Mortality due to embolectomy appears to be in the range of 30% to 40%, but may be as low as 8% in those who have not sustained cardiac arrest preoperatively.139,140 Even if this lower number reflects improvements in anesthetic or operative technique, this mortality is still comparable to that of patients with massive embolism treated less invasively.141 The argument that embolectomy might reduce the long-term consequences of chronic pulmonary hypertension lacks force, even though this complication is more common than previously thought.142 Thus it has never been demonstrated in a well-performed trial that embolectomy confers any advantage over thrombolytic therapy, or for that matter heparin. The population of patients who might benefit from embolectomy appears to be those who meet the following criteria: those with a hemodynamically significant embolism, in whom thrombolytic therapy is contraindicated; and in a center with a rapidly responding cardiopulmonary bypass team and a surgeon experienced in the technique of embolectomy.

Several new devices have been tested which aim to remove pulmonary emboli less invasively than the direct surgical approach. For example, a 10F suction catheter, inserted through a jugular or femoral venotomy and advanced into the pulmonary artery has been used to extract clot.143 Eleven of eighteen patients improved immediately. Suction embolectomy was more likely to be successful in patients treated promptly after hemodynamic deterioration. More recently, patients with shock underwent mechanical fragmentation of their massive PE with a rotational pigtail catheter, followed by thrombolytic therapy.144 Nine of ten patients survived, and 6 of the survivors achieved hemodynamic stability within 48 hours of the procedure. Alternative methods to re-establish pulmonary artery patency include endovascular stents,145 but experience with all of these techniques is limited.

Pregnancy

The pregnant woman who may have PE presents unique challenges.146 Pregnancy is thought to be a risk factor for venous thrombosis, and PE is the second leading cause of death among gravidas (next to trauma).147 The addition of the fetus, as well as anatomic considerations, leads to several key differences in management.

Diagnosis

Diagnosis can be more difficult because of reluctance to perform potentially risky procedures, particularly involving diagnostic radiology. However, it is important not to lose sight of the risk of failing to make the diagnosis. Thus when the diagnosis is seriously considered, it should be pursued. V̇/Q̇ scans probably pose little risk to the fetus. The estimated radiation dose is small, and the risk is clearly less than that of missing a diagnosis.148 Impedance plethysmography is also less useful in pregnancy due to mechanical compression of the IVC by the enlarged uterus, especially in the third trimester. A positive result in the first two trimesters is sufficient basis on which to anticoagulate, but may not be in the last trimester unless the vascular laboratory has particular expertise in pregnancy.149 The risk of helical CT angiography is not known, but when indicated, it should not be withheld out of concern for fetal radiation exposure. The maternal risks of undiagnosed and untreated PE are clear. A diagnosis of thromboembolism in pregnancy has serious implications for the mother, not just in the current pregnancy, but during any subsequent pregnancies. Because of the risk of heparin-induced osteoporosis and the remaining uncertainties about effective prophylaxis against recurrence during future pregnancies, the diagnosis of thromboembolism should never be made lightly in a gravida, but instead should be based on solid evidence.

Treatment

The use of heparin in pregnancy is no different than in the nonpregnant patient. Heparins do not cross the placenta, and pose a low risk to the fetus. However, warfarin, which is teratogenic, should not be instituted. Rather, long-term treatment should consist of subcutaneous heparin150 or LMWH.151 LMWHs are preferred in that they have a lower incidence of both bleeding and of heparin-induced thrombocytopenia.151 Osteoporosis is a serious complication of full-dose heparin during pregnancy, with fractures occurring in 2% of women.152 There are indications that bone mineral loss is reversible,153 but several cases of debilitating back pain have been reported. Low molecular weight heparins, when prospectively studied, do not appear to accelerate bone mineral density loss.154 Thrombolytic therapy risks spontaneous abortion and uterine bleeding. Nevertheless, several case reports of the successful use of thrombolytic treatment of PE have appeared. Vena caval interruption requires some modification as well. The left ovarian vein (a potential source of clot) drains into the left renal vein. Therefore, when a caval filter is placed, it should be inserted to a suprarenal position, rather than below the renal veins. Again, several cases of successful use of these devices have been described. Pregnancy may be an ideal situation for the temporary vena caval interruption device.

Chronic Obstructive Pulmonary Disease

Patients with COPD are at increased risk for PE. In addition, their preexisting respiratory compromise and abnormal pulmonary vasculature leave them particularly vulnerable to the cardiopulmonary consequences of PE. Ironically, diagnosis of PE in the setting of COPD is unusually difficult. Patients commonly complain of dyspnea, chest pain, cough, and anxiety, and occasionally note hemoptysis and leg swelling. Their examinations, chest radiograms, ECGs, and arterial blood gas values are usually abnormal at baseline. V̇/Q̇ scans are most often unhelpful. For example, in PIOPED 108 patients were identified as having COPD (although objective data were available in only 43).155 Scans were intermediate probability in a full 60%. Only 20 patients (19%) had results which made pulmonary angiography unnecessary by being normal, high probability, or low probability paired with a low pretest clinical estimate. Nevertheless, for the occasional patient, V̇/Q̇ scanning obviated pulmonary angiography.

When patients with COPD present with symptoms that are atypical for their usual exacerbation, particularly when the PaCO2 is reduced from previously elevated values, it is worth considering the diagnosis.156 Positive leg studies may provide a rationale for anticoagulation and obviate the need for further investigation, although this approach has been called into question.41 CT angiography or pulmonary angiography may be necessary to establish a diagnosis. The physician and patient may be in the unfortunate position of having to repeatedly consider an invasive procedure or risk a missed diagnosis. In fact, it is in just this type of situation that PEs go undiagnosed. There is no simple answer to this problem and PEs will continue to be underdiagnosed until better, less invasive tests are available.

Patients with Coagulation or Platelet Disorders

The risk of venous thromboembolism in patients with chronic liver disease or marked thrombocytopenia is not known. While it seems sensible to conclude that the risk must be lower than if clotting and platelet function were normal, PEs occur even when the bleeding tendency is severe.157,158 Therefore, when the clinical presentation strongly suggests PE, thrombocytopenia and coagulopathy should provide little reassurance, and diagnostic testing is indicated. Patients with chronic renal failure—excepting those with nephrotic syndrome—do seem to be at a remarkably low risk of venous thromboembolism, so alternative diagnoses should always be sought.

A discussion of prophylaxis has been left to the end of the section on thromboembolism because here it is particularly easy to emphasize several points. PE is common, lethal, usually missed, difficult to evaluate, and costly to treat. Critical illness makes these statements especially true, therefore the goal must be to prevent this disease. A full treatment of prophylaxis is beyond the scope of this chapter, but a few points are worth making. Prophylaxis in the ICU is discussed further in Chapter 10.

Over the last 20 years, a large volume of information regarding the risks and benefits of numerous prophylactic strategies has been collected. The greatest accumulated experience is with low-dose heparin, typically 5000 U administered subcutaneously either bid or tid both in surgical and medical patients. Other methods of thromboprophylaxis are also available, and may have a role in special subsets of patients, patients in whom heparin is ineffective or contraindicated, or those who are at unusually high (or low) risk of VTE. These alternatives include low molecular weight heparins (LMWHs), adjusted-dose heparin, aspirin, warfarin, dextrans, antiembolism stockings, intermittent pneumatic (or sequential) compression cuffs, and vena caval interruption. In some patients combined prophylaxis is superior. For example, the combination of minidose heparin and intermittent pneumatic compression was more effective than minidose heparin alone in patients following cardiac surgery.159 In trauma patients160 and in patients suffering hip fracture, LMWH is superior to low-dose heparin. In many settings the optimal approach has not yet been clarified and there are far fewer data regarding medical patients than surgical patients. Nevertheless, for most patients at risk of VTE, some method which is both safe and effective is available. In one survey of 152 medical ICU patients, only 33% were given prophylaxis and there was a mean delay of 2 days before prophylaxis was prescribed, although risk factors were identified in 87% of patients.161 In the absence of a contraindication, nearly all critically ill patients should receive minidose heparin. When heparin is unsafe (e.g., neurosurgical patients), alternative methods of prophylaxis, such as intermittent pneumatic compression cuffs, should be instituted. In some patients at very high risk of thromboembolism, such as those with acute spinal cord injury, minidose heparin is ineffective and LMWH is preferred.83 In most patients, however, the increased cost of LMWHs is not matched by clearly incremental efficacy or improved safety.83 Finally, although controversial, some patients at extremely high risk of PE should have consideration given to prophylactic insertion of vena caval filters (see above).

The syndrome of air (or gas) embolism results when air enters the vasculature, travels to the pulmonary circulation, and causes circulatory or respiratory embarrassment. It is uncommonly recognized in critically ill patients, but is quite likely underdiagnosed.

Pathophysiology

The syndrome is triggered when a gas, usually air, enters a vessel, typically a vein. It travels with the venous return to the right heart and lungs, where it may have circulatory or respiratory consequences. Occasionally, air reaches the arterial circulation leading to systemic manifestations. Although air embolism is often abrupt and short-lived, intriguing cases of continuous streaming of bubbles in mechanically ventilated patients have been reported.162 At times, such embolism may persist over many days.

Entry of Air into the Vasculature

Development of air embolism requires an abnormal communication between air and the blood vessel. In addition, there must be a pressure gradient to favor entry of air into the vessel, rather than bleeding from the vessel. Trauma, surgical incisions, and intravascular catheters create the most common sources of air entry. In addition, there are more subtle paths through which air can reach the vasculature, such as in damaged, mechanically ventilated lungs of patients with ARDS. The driving gradient for air entry may be provided by air under pressure, as during positive pressure ventilation or high pressure wound irrigation. Alternatively, the air may be at atmospheric pressure, but the intravascular pressure is subatmospheric. For example, any vein which is above the heart by an amount exceeding the right atrial pressure is likely to be at less than atmospheric pressure (and therefore appears collapsed). Table 27-8 lists some of the causes of the air embolism syndrome.

Circulatory Consequences

Massive air embolization can fill the right heart, impede venous return, and thereby stop circulation; thus sudden death is one of the possible outcomes. It is estimated that >100 mL of air must be acutely infused to arrest circulation. Most often, however, air passes through the right heart into the lungs. There it raises PA pressure, but has predominantly respiratory consequences. Since unilateral experimental air embolism causes pulmonary hemodynamic changes similar to bilateral embolism, humoral or reflex vascular changes probably account for some of the increase in pulmonary vascular resistance. In an experimental canine model of massive air embolism, systemic hypotension and pulmonary hypertension were attenuated by pretreatment with an intravenous endothelin-receptor antagonist.163 Such work suggests that the circulatory effects of air embolism may be due to cytokine release following release of endothelin—a potent pulmonary vasoconstrictor—and that the activation of the cyclooxygenase pathway may contribute. One case study described a patient with suspected air embolism who developed the systemic inflammatory response syndrome,164 perhaps lending further credence to the idea that air in the circulation causes downstream humoral, not simply mechanical, consequences.

Respiratory Consequences

Air is carried into the pulmonary vasculature where it embolizes in pulmonary arterioles and capillaries. The abnormal air-blood interface is thought to denature plasma proteins, creating amorphous proteinaceous and cellular debris at the surface of air bubbles.165 This debris attracts and activates white blood cells, facilitating injury to the pulmonary capillaries. Endothelial injury increases capillary permeability, which leads to alveolar flooding. The resulting noncardiogenic pulmonary edema accounts for the majority of symptoms and signs due to air embolism (see Chap. 38). In addition, air embolization leads to bronchoconstriction, a point which may be useful in diagnosis.166

Although the dominant gas exchange abnormality is hypoxemia, carbon dioxide elimination is impaired as well. As pulmonary vessels become occluded, alveoli subtended by them are ventilated, but unperfused. This increment in dead space may be signaled by a drop in end-tidal CO2, if this is being monitored. In the patient with fixed minute ventilation (e.g., if the patient is on a muscle relaxant), PCO2 will rise. Either of these may lead to suspicion of the diagnosis.

Extrathoracic Manifestations

Air embolism is occasionally accompanied by systemic findings. If air directly enters the pulmonary veins, as may occur in patients being mechanically ventilated with acute lung injury, bubbles pass directly to the arterial circulation. However, since air typically enters a systemic vein, the arterial circulation is protected from embolization by the filtering effect of the pulmonary circulation. Nevertheless, bubbles can pass to the left side of the heart via the foramen ovale, which is probe patent in up to 30% of people. This type of foramen ovale does not ordinarily allow right-to-left shunting, due to the higher pressures in the left atrium. However, after significant embolization to the pulmonary circulation, right heart pressures rise, reversing the interatrial gradient. This allows bubbles to pass directly from the right to left atrium, then to the systemic circulation. Even in the absence of a foramen ovale, air can reach the arterial circulation since the lungs do not fully filter air, especially when a large amount is embolized. Air may pass through large extra-alveolar vessels or through the pulmonary capillaries themselves. In animal experiments, the threshold rate of venous air infusion that overwhelms pulmonary filtering is 0.30 mL/kg per minute.167 For a 70-kg man, this value translates to only 21 mL/min.

Once air reaches the arterial circulation, peripheral embolization leads to ischemic manifestations in the brain, heart, skin (livedo reticularis),168 and other organs. Some of the ischemic manifestations in the periphery are probably mediated by polymorphonuclear leukocytes and oxygen radicals, as is the injury in the lung.169

Presentation

Air embolism is usually recognized when it presents as acute hypoxemic respiratory failure. As noted above, it may also manifest as an acute hypoperfusion state or as peripheral embolization. The chest x-ray shows diffuse alveolar filling. We have seen one case in which intracardiac and intra-arterial air was grossly evident on the chest radiograph. Increased dead space may be indicated by increased V̇e, increased PCO2, or decreased ETCO2. Rarely, echocardiography will demonstrate residual air (or ongoing embolization) in the heart. Precordial Doppler monitoring during high-risk surgery is well suited for detecting air (Table 27-9).

A diagnosis of air embolism is usually considered when air is witnessed to enter an intravascular catheter. It is also likely to be considered in extremely high-risk settings such as upright neurosurgery. However, if air embolism is only considered when it is grossly apparent, many episodes will go unappreciated. It should also be included in the differential diagnosis of patients with hypoperfusion, systemic embolization, obtundation, and respiratory failure, especially when more likely causes are lacking. It is also worth emphasizing that many cases are related to central lines, not only during placement, but while the catheters are in place (catheter disconnection, hub fracture, or gas in the line), while being changed over a wire, and after they are removed (through a persistent cutaneous tract).

The differential diagnosis of air embolism includes other forms of noncardiogenic pulmonary edema, as well as cardiogenic edema. Thus volume overload, sepsis, and gastric acid aspiration must be excluded.

Management

The goals of treatment are to prevent re-embolization while supporting respiration and circulation. In most cases, resolution is prompt. The source of air entry should be identified and closed if possible. Alternatively, the gradient favoring air entry can be lessened, as for example by saline administration to raise intravascular pressures. When air embolism complicates positive pressure ventilation, it is advisable to lower airway pressures by lowering tidal volumes, reducing positive end-expiratory pressure (PEEP), or intentionally hypoventilating. Oxygen hastens the reabsorption of air from bubbles, so all patients with significant air embolism should receive 100% oxygen during the initial resuscitation.170

In certain situations, it may be possible to retrieve air from the venous circulation or right heart, especially intraoperatively when a catheter is in place for that purpose. However, this should not be routinely attempted in other settings because significant amounts of air cannot usually be removed. Positional maneuvers to prevent air from embolizing to the lungs (such as head down left decubitus) are largely unproved. Similarly, the distribution of arterial emboli seems little affected by the Trendelenburg position, since the force of arterial flow greatly outweighs the buoyancy of the bubbles.171

Standard treatment is similar to that of any patient with ARDS. Mechanical ventilation to reduce the work of breathing, with oxygen and PEEP to maintain arterial saturation are usually necessary. Although the pulmonary edema is not related to hypervolemia, the degree of lung leak is probably sensitive to filling pressures. Therefore we reduce filling pressures to the lowest value that allows an adequate Q̇t (see Chap. 38).

In animal experiments, corticosteroids given before embolization, or shortly following embolization, reduce the degree of lung injury. In high-risk neurosurgical patients on steroids, the incidence of the syndrome seems less than in similar patients who are not so treated, suggesting a prophylactic benefit in human beings as well. Nevertheless, no clear treatment role for these agents has been demonstrated, and they should not be routinely given. Potential future therapies include other anti-inflammatory drugs and agents directed against oxygen free radicals.

Hyperbaric treatment is of theoretical benefit since compression reduces the size of bubbles. This reduces the surface area for activation of white blood cells and can thereby limit pulmonary and systemic injury. Such therapy is standard when the mechanism of gas embolism is decompression, such as in professional and recreational divers. It is not routinely used in other critically ill patients, however. Since patients usually respond readily to standard supportive measures, and since the syndrome typically resolves in only 24 to 48 hours, it is unlikely that hyperbaric treatment will find a prominent role in treatment. It is probably better to keep patients in the ICU where they remain under the watchful eye of the health care team, rather than to risk transport for an unproven treatment. When such patients are transported by air, a pressurized craft flying at low altitude should be requested.

The fat embolism syndrome (FES) is associated with fat particles in the microcirculation of the lung. It consists typically of lung dysfunction, neurologic manifestations, and petechiae, usually following a latent interval. It is most common following long bone fractures, typically presenting as dyspnea and confusion. However, FES is seen after other forms of trauma and in several nontraumatic conditions as well. For example, FES has been proposed as a major cause of the acute chest syndrome in patients with sickle cell disease172(see Chap. 108). After long bone or pelvic fracture, the incidence of the syndrome is at least 10% when evidence for it is prospectively sought,173 although serious clinical manifestations are seen in only 1% to 3%. Since the clinical presentation is usually mild, FES is often unrecognized. Even when lung injury is obvious, its cause may be attributed to infection, aspiration, or traumatic ARDS, rather than to fat embolization. Some of the causes of FES are presented in Table 27-10.

Pathophysiology

Nontraumatic Embolism

Fat globules are seen in pulmonary (and other) vessels at autopsy and can be found in venous blood. In contrast to traumatic embolism, the fat is probably not derived from bone marrow, but rather arises from lipids in the blood. Serum from acutely ill patients has the capacity to agglutinate chylomicrons and very low density lipoproteins, as well as liposomes of nutritional fat emulsions.174 It has been proposed that C-reactive protein (CRP), which provokes the calcium-dependent agglutination of each of these lipid-containing substances, may underlie nontraumatic FES. Since CRP is dramatically elevated in trauma, sepsis, and inflammatory disorders, this provides a mechanism for fat embolization.

An alternative, but less attractive, hypothesis implicates the liberation of free fatty acids (FFAs) from fat stores. Although FFAs are known to injure the pulmonary vascular endothelium, their concentration in the systemic circulation during critical illness does not rise sufficiently to account for lung injury. An understanding of the pathophysiology is potentially valuable in the search for prophylactic and therapeutic approaches to these patients, since corticosteroids, for example, may ameliorate the toxic effect of embolized fat but act to mobilize FFA.

Traumatic Embolism

Fracture of bone releases neutral fat which embolizes into the pulmonary vasculature. The derivation of this fat from bone is supported by the finding of coincident particles of bone marrow at autopsy in patients with long bone fractures and by echocardiographic studies showing frequent, often dramatic embolism at the time of medullary reaming.175 Local hydrolysis of fat by lung lipase releases toxic FFAs, which generate endothelial injury. Systemic findings in FES probably relate to passage of venous fat emboli across the pulmonary circulation, although serum-derived fat may play a role. Elevated right heart pressures following embolism may open a probe patent foramen ovale, causing severe, even fatal systemic embolism.176 Further, fat can cross the pulmonary circuit even in the absence of a right-to-left shunt, as has been shown in experimental animals.177 Fat was able to traverse the pulmonary microcirculation even though 15-micron radiolabeled microspheres could not, perhaps due to enhanced deformability of fat emboli.177

Clinical Manifestations

Following injury, there is usually a latent interval of 12 to 72 hours before the syndrome becomes evident. The dominant findings are related to lung injury. Patients with FES present as ARDS, with dyspnea, hypoxemia, and a diffuse lung lesion. In addition, there is often confusion, obtundation, or coma, signs due to cerebral fat embolism rather than coincident hypoxemia. The typical neuropathologic findings include fat microemboli and diffuse petechial hemorrhagic infarcts. Petechiae are also seen on the skin, particularly over the upper chest, neck, and face. On funduscopic examination, embolized fat may be detected in retinal vessels (Purtscher's retinopathy). Often thrombocytopenia and anemia are present. Rare patients will develop a full blown acute right heart syndrome176 (see Chap. 26).

The diagnosis is usually based on the clinical findings in a patient at risk for FES. Fat globules in the urine are neither sensitive nor specific for the diagnosis of FES. Attempts have been made to find alternative, objective means of diagnosis, since this might be useful in devising prophylactic or therapeutic strategies. Fat can be detected in bronchoalveolar lavage specimens in many patients following trauma, but this finding appears not to be a reliable means of diagnosis.178 Fat can also be seen in spun samples of blood withdrawn from a wedged pulmonary artery catheter, but this finding is also not clinically useful.179

Prophylaxis and Treatment

The substantial incidence of the syndrome and the large number of patients with a well-defined risk factor (long bone fracture) have provided the opportunity to evaluate prophylactic approaches. The least controversial strategy has involved a shift toward early fixation of long bone fractures, even in patients with multiple trauma. Early fixation decreases the incidence of FES, as well as of ARDS and pneumonia, and reduces length of stay.180–182 Orthopedic surgery, and particularly total hip arthroplasty, carries a high risk of fat embolism, which has been demonstrated by transesophageal echocardiography to occur during the preparation of the femur and insertion of the femoral component. Specific surgical techniques, such as using bone vacuum with the preparation of the femur,183 have shown decreased fat emboli detected by transesophageal echocardiography, although the clinical significance of such of echo-detected emboli is less clear. More controversial is the use of prophylactic corticosteroids. Nearly all trials of methylprednisolone, in both high (7.5 mg/kg every 6 hours) or low dose (1.5 mg/kg every 8 hours), have shown a reduction in the incidence of the FES, as well as less severe hypoxemia.184–186 Nevertheless, concerns regarding the risk of infection and impairment of wound healing have limited the routine use of these drugs. Since most cases of FES are mild and the great majority of patients recover, an acceptable prophylactic regimen would have to be quite safe and inexpensive.

Once the syndrome becomes evident, treatment is that of ARDS (see Chap. 38). Prevention of re-embolization by fracture fixation should be attempted, and supportive management with oxygen and PEEP initiated. It has been suggested that corticosteroids may be of benefit even once the syndrome is established, but it is not clear that the benefits outweigh the risks, and we do not recommend them. No clear role has been established for glucose and insulin, heparin, ethanol, and albumin, despite studies seeking a useful therapy for FES.