|
Meeting the Challenge of Pulmonary Embolism
The so-called classic presentation of pulmonary embolism is rarely seen in clinical practice. Here’s how to make sense of the clues.
By Colin Kaide, MD, FACEP, FAAEM, Scott Williams, MD, David P. Bahner, MD, RDMS, and Douglas A. Rund, MD, FACEP
Pulmonary embolism (PE) may appear as dramatically as an acute massive embolism or as subtly as shortness of breath. Each patient encounter is unique, and, as is often the case in the clinical setting, important clues may be discovered only during a probing assessment—and sometimes reassessment.
WHEN IN DOUBT, ASK MORE QUESTIONS
Consider Ms. W, 81, who presents with pain behind her left knee. She also reports some mild swelling in the leg, which began shortly after the pain started. She can’t recall any injury to the leg. She says she has had no fever, chills, or other signs or symptoms of infection. She denies having any chest pain but does say that she is “always short of breath.”
Ms. W’s medical history includes systemic lupus erythematosus and a myocardial infarction many years ago. She also had an episode of deep vein thrombosis (DVT) and subsequent PE after a hip fracture. She was treated with a Greenfield filter inserted in the inferior vena cava. She had no previous blood clots; a hypercoagulation work-up was not performed at the time.
She is taking simvastatin, ezetimibe, esomeprazole, aspirin, isosorbide mononitrate, furosemide, potassium chloride, and losartan. She is not on anticoagulation or beta-blocker therapy.
Her physical exam is essentially normal, except for the knee pain and associated 2+ pitting edema in her lower leg and foot. Examination of her heart and lungs finds no abnormalities.
A review of her medical record reveals that she had been seen four days earlier for left flank pain. The evaluation at that time included a urinalysis, left rib x-ray series, and a noncontrast computed tomography (CT) scan of the abdomen and pelvis to assess for nephrolithiasis. All studies were unremarkable, and Ms. W was discharged home with a diagnosis of “muscle strain.”
After the history, physical exam, and data review, the patient’s signs and symptoms are reassessed because the clinical picture seems unclear. She still has posterior knee pain but no acute symptoms of chest pain or shortness of breath different from her baseline. Vital signs remain normal. More detailed questions reveal that although she denies chest pain, she is experiencing some “chest tightness.” Further exploration of her initial report that she is “always short of breath” elicits a complaint of increasing shortness of breath with exertion over the previous three days.
Ms. W is placed on a cardiac monitor, intravenous access is established, and she is given oxygen by nasal cannula at 2 L/min. An ECG and chest x-ray are ordered, along with a complete blood count (CBC), chemistry panel, cardiac biomarkers, D-dimer, and coagulation studies. Doppler ultrasound of the left lower extremity is not immediately available.
The ECG shows a normal sinus rhythm, without ST-segment changes or T-wave inversions. No evidence of acute disease is found on the chest films. The CBC, chemistry panel, and coagulation studies are all unremarkable. Her total creatine kinase is 33 U/L, with a troponin of 0.02 ng/ml (both negative). The D-dimer is elevated at 10.7 µg/ml (normal is less than 0.50).
At this point, the emergency physician has to rethink his approach. Despite normal vital signs and a previously placed Greenfield filter, the patient is reporting chest and respiratory symptoms, and the D-dimer result suggests an acute clotting process. A spiral CT pulmonary angiogram is ordered; it finds a large filling defect within the pulmonary vasculature compatible with a saddle embolus (see image below). The clot extends into the lobar and segmental arteries bilaterally.
 |
| Possible saddle embolus. This spiral CT pulmonary angiogram of the patient presented in this article shows a large filling defect within the pulmonary vasculature that is compatible with a saddle embolus. The clot extends into the lobar and segmental arteries. |
The patient remains hemodynamically stable throughout her emergency department stay. She is given a heparin bolus of 80 U/kg, followed by a heparin infusion at 18 U/kg/hr. She is then admitted to the intensive care unit.
FACTORS THAT PROMOTE COAGULATION
Deep vein thrombosis and PE are manifestations of the same disease, occurring at different places along the clinical spectrum. Together they are most accurately called venous thromboembolism (VTE). An estimated 2 million Americans suffer from DVT annually; about 300,000 die, almost always from a distal clot traveling to the lungs. These numbers do not take into account those PEs that go unrecognized. In a recent paper from the Carolinas Medical Center, we noted that patients who experience sudden death with pulseless electrical activity carry a high probability of having had a massive PE.
More than 150 years ago, Rudolf Virchow, a German physician widely regarded as the father of pathology, elucidated three factors that promote intravascular coagulation:
Venous stasis. Conditions that predispose the patient to venous stasis generally involve immobilization of the whole body or one or more extremities, which promotes in situ clot formation.
Vessel wall damage. Any condition, such as trauma or surgery, that damages the endothelium of the vessel walls and exposes the basement membrane may initiate the extrinsic path of the clotting cascade. The risk tapers off four weeks or so after an operation and during a similar time frame after trauma.
Hypercoagulability. This factor includes both inherited defects in the coagulation system and acquired factors that make patients more prone to pathological clotting.
The box above lists the various factors in Virchow’s famous triad that can cause VTE. Keep in mind, though, that while the general pathophysiology of thromboembolic disease is understood, as many as 20% of patients with VTE have no identified risk factors and are classified as idiopathic.
DIAGNOSING DEEP VEIN THROMBOSIS
Signs and symptoms of DVT include extremity swelling, warmth, redness, pain, and tenderness. Rarely, a “cord” or firm, clotted, deep lower extremity vein may be palpable on exam. Occasionally, DVT arises in the pelvic or uterine vessels, in which case the clinical presentation may include swelling of the whole leg and pelvic pain. Sometimes a Doppler ultrasound exam doesn’t show a clot in the leg, although a large clot burden may be present in the pelvis, making the diagnosis of DVT harder.
|
The differential diagnosis of extremity edema and pain is broad (see box above). To narrow the differential, physicians should process the patient’s history and physical in the context of clinical experience to determine the pretest probability for each condition. That is, the clinician assesses the patient’s risk of a specific disease before beginning diagnostic testing. This pretest probability may be arrived via a clinical gestalt (educated guess) or through the use of a formal, standardized assessment tool. After establishing a pretest probability, the results of diagnostic tests can be used to generate a post-test probability. This strategy is useful when the tests are unable to provide a clear-cut “yes” or “no” answer.
The American College of Emergency Physicians policy guidelines for diagnosing and managing DVT and PE strongly advocate calculating a pretest probability using a formalized assessment prior to testing. The Wells and Modified Charlotte Criteria are cited as options, with the Wells criteria (see tables below) being the most widely recognized.
The gold standard for diagnosing DVT is contrast venography. However, it is an invasive test, has multiple potential complications, and has fallen out of favor in the routine work-up of DVT. It has been almost entirely replaced by compression ultrasound, which has a sensitivity of approximately 90%, making false negative studies uncommon. Consider repeating the exam in three to five days in a patient who has a high pretest probability for DVT and a negative Doppler ultrasound exam.
Another diagnostic option is to use D-dimer in conjunction with a structured clinical pretest probability assessment. D-dimer is a breakdown product of cross-linked fibrin found in blood clots. Both ELISA D-dimer testing and immunoturbidimetric (also called LIA, or Latex Immuno-Assay) testing have a sensitivity of 95% to 97%, making them useful for ruling out DVT in low- and moderate-risk patients. When evaluating a patient for DVT using D-dimer, the first step is to assign a pretest probability. The patient is classified as low risk, moderate risk, or high risk based on a prospectively validated set of criteria such as the Wells criteria. Ms. W, for example, was deemed a moderate-risk patient using the Wells criteria (2 points, one each for pitting edema and tenderness).
In the high-risk population, the D-dimer does not offer diagnostic utility because the starting pretest probability is too high for the negative D-dimer to confer a low-enough post-test probability to exclude DVT. Because D-dimer testing lacks specificity, it is quite poor at ruling in disease, making further evaluation necessary in the low-risk patient with a positive D-dimer.
The successful use of a gestalt or clinical intuition to gauge the pretest probability of PE depends on the skill of the treating physician and can vary widely among practitioners. That being said, the ability of an experienced clinician to judge low to moderate pretest probability for PE seems to correlate well with the more structured approach of Wells. This has not been formally evaluated in diagnosing DVT.
OUTPATIENT TREATMENT
Recent studies and a great deal of clinical experience nationally suggest many patients with newly diagnosed DVT are candidates for outpatient treatment with low-molecular-weight heparin (LMWH) and warfarin. The LMWH would be continued until the patient has a therapeutic INR, which usually takes five days. Owing to the rapid onset of action of LMWH, warfarin can safely be started concomitantly with the first injection of the LMWH without conferring a higher risk of the transient hypercoagulability typically seen if warfarin is started without another anticoagulant.
The outpatient strategy is appropriate if certain criteria are met: the patient has no concomitant pulmonary emboli or comorbid diseases, is able to return if symptoms worsen, can inject heparin or LMWH, and can afford the necessary medications. In addition, the patient must have a primary care physician who can adjust warfarin dosing based on serial measurements of prothrombin/INR levels. At our institution, outpatient management is the rule; only 10% of patients with lower-extremity DVT are managed as inpatients.
The role of thrombolysis, either systemic or catheter-directed (delivered directly into the clot), is still undefined. Thrombolysis does increase bleeding complications. What is unclear is whether it reduces the incidence of postphlebitic syndrome (a post-DVT inflammatory condition that causes valvular incompetence), seen in 20% to 30% of cases. There is also a role for vena cava filters in patients who have contraindications to systemic anticoagulation or who experience a new thromboembolic event while on adequate anticoagulation. Filters can also be placed in the superior vena cava in patients with upper-extremity DVT. These filters, designed to catch clots and prevent PEs, have a limited effectiveness and only result in a decrease in PE risk of approximately 30%.
In the past, DVTs that were isolated to the calf were believed to pose no significant risk of embolism and were not routinely treated with anticoagulation. More recent evidence suggests that the risk of clot propagation and subsequent embolism from isolated calf DVTs is high enough to require anticoagulation to prevent proximal clotting.
Thrombectomy may be indicated if infection is present or if the clot burden is extensive.
PULMONARY EMBOLISM: MYTH AND MISCONCEPTION
Pulmonary embolism is the most concerning complication of DVT. More than 200,000 cases are diagnosed in the United States every year. Mortality rates as high as 25% at seven days have been reported. Most PEs are believed to arise from the deep veins of the lower extremities and pelvis, though they can also migrate from the upper extremities and the renal veins. Risk factors for PE are the same as those for DVT.
The most common signs and symptoms of PE are dyspnea, tachypnea, and pleuritic chest pain. Other common findings include tachycardia, fever (rarely above 102°F), and hemoptysis (rarely massive). Occasionally, in the case of massive clot burden, the patient may suffer syncope or sudden death. Cyanosis, rales, and heart murmurs are less common findings.
A common misconception is that reproducible chest pain with chest wall palpation makes PE very unlikely. Clinicians may also dismiss PE as a possible diagnosis when the patient says the symptoms didn’t start suddenly. However, in a review of data collected on a large population of patients undergoing evaluation for PE, only 20% of patients with reproducible chest pain had PE. Moreover, only 39% of patients with proven PE described a sudden onset of symptoms, compared to 51% of those without PE.
A chest X-ray and ECG are most useful to evaluate for the possibility of other diseases that can cause chest pain and dyspnea, such as acute coronary syndrome or pneumonia. The most common chest x-ray findings in PE are a normal film, nonspecific parenchymal abnormality (such as microatelectasis), and atelectasis. Other so-called classic findings such as the Hampton hump (pleural-based, wedge-shaped defect) and Westermark sign (truncation of the pulmonary vessels) are actually quite rare.
The most common ECG findings in PE are a normal tracing, nonspecific ST-T wave changes, and sinus tachycardia. Other frequent findings include signs of right heart strain. The often-mentioned S1Q3T3 is neither specific nor sensitive enough to rule PE in or out and is present in only 13% of PEs.
Arterial blood gases and pulse oximetry are rarely useful in the diagnostic algorithm for PE. However, hypoxemia does appear to correlate with the degree of pulmonary vascular occlusion and can predict outcome. It may therefore be helpful in risk stratification and determining appropriate triage to ICU, stepdown, or a monitored floor bed.
TAKING THE TESTING FURTHER
Just as in the evaluation of DVT, D-dimer testing can be used to evaluate patients for PE. In particular, the quantitative ELISA D-dimer and the LIA test have a sensitivity of approximately 95%, making a negative result in the appropriate low-risk patient quite useful for ruling out disease. A positive result, on the other hand, demands further testing in the absence of an alternative diagnosis.
As with DVT, pretest probability must be established to interpret the results of a D-dimer. When Wells did his original studies, he used a form of D-dimer testing that had a sensitivity of around 80%. This required a pretest probability of around 13% to allow a negative D-dimer to confer a post-test probability that would effectively rule out the disease. By using new high-sensitivity D-dimers (95% to 99%), a pretest probability can be as high as 40% and PE can still be ruled out with a negative D-dimer. Using the Wells criteria to assess low pretest probability confers a 12% to 13.5% risk of PE. The clinical gestalt assessment of a low probability for PE leads to a 19.9% chance that a PE will be present. Whether using a gestalt low probability or the more structured assessments, a modern, high-sensitivity D-dimer can be effectively employed in low and moderate pretest probability patients because it has a very good negative predictive value. However, in patients with a high pretest probability for PE, a negative D-dimer cannot confer a low enough post-test probability to rule out PE.
A helical CT angiogram of the chest has become the diagnostic modality of choice for PE in most medical centers, replacing the old gold standard, the pulmonary angiogram. Sensitivity for CT angiography varies with the interpreting radiologist and the type of CT scanner but approaches 95% in many studies. For situations in which a CT angiogram is contraindicated or unavailable, ventilation-perfusion (VQ) scanning is an option. However, it is important to note that only VQ scan results that are normal or near-normal, or that are low probability in a low-risk patient, are helpful in ruling out PE. As an example, according to PIOPED I (Prospective Investigation of Pulmonary Embolism Diagnosis), a clinically high-risk patient with a low-probability VQ scan result still has a 16% to 40% risk of having PE.
What about pregnant patients who need to have PE ruled out? According to the International Commission on Radiological Protection, the spiral CT study delivers less absorbed radiation to the fetus than VQ scanning. Both tests are considered acceptable, even in pregnancy, if the diagnosis of PE is considered. The radiation risk is not nearly as great a concern as the risk of undiagnosed PE.
|
A set of criteria has been derived to define a population of ultra-low-risk patients. These criteria have undergone prospective validation in three separate studies, the most recent of which involved more than 7500 patients from one German and two American hospitals. The pretest probability for PE in this group is low enough that any testing will have a false positive risk equal to the false negative risk. Patients who pass the PERC rule (Pulmonary Embolism Rule-Out Criteria) have a pretest probability for PE of less than 2%, and further testing cannot confer a post-test probability any lower. This rule (see box above) can be applied to patients when there is a clinical intuition of low pretest probability. Often the clinician does not believe there is a real need to pursue the diagnosis but feels obliged to do so for fear of possible medicolegal consequences. The PERC rule should not be used to obviate the need for a work-up in high-risk patients.
Treatment of PE is the same as for DVT—namely, anticoagulation with heparin or LMWH followed by warfarin. Indications for thrombolysis, either systemic or catheter-directed, are controversial and may include witnessed arrest, hemodynamic instability, and right heart strain on echocardiography. Surgical embolectomy is also an option for patients with contraindications to thrombolysis and profound hypotension or severe hypoxemia.
With the advent of outpatient anticoagulation, it has been suggested in the literature that stable patients with PE may be treated on an outpatient basis. The scientific literature in this area is not mature, however, and the standard of care remains inpatient admission at this time.
WHEN THE CASE ISN’T ‘CLASSIC’
The textbooks tell us that PE is clinically manifested by tachycardia, tachypnea, pleuritic chest pain, and in some cases hemoptysis. In severe cases, dramatic symptoms such as hypotension, profound hypoxemia, or even sudden death may occur. Unfortunately, thinking about the diagnosis in terms of these classic symptoms can lead to dangerous conclusions. The most problematic error is the assumption that the absence of a classic presentation means the absence of PE.
In Ms. W’s case, there were good reasons to suspect that she might have DVT in her lower extremity. What was her risk for DVT and PE: low, moderate, or high? These assessments and documentation of medical decision-making are not commonly performed yet are essential to ordering and interpreting test results. Since the test of choice, duplex Doppler ultrasound, was not immediately available, the diagnosis could not be confirmed. Although Ms. W initially denied chest symptoms, the treating physician’s experience and persistence allowed for clarification of the clinical presentation, which made PE seem more likely.
The diagnosis was confirmed with a spiral CT, and the appropriate treatment was initiated. When either DVT or PE is diagnosed, the ultimate diagnosis of VTE is made, mandating treatment. One may argue that simply finding DVT is adequate in this case, since treatment for DVT and PE is essentially the same: anticoagulation.
Because PE can be such a difficult diagnosis, many physicians will miss it during their career. The goal is to make this as infrequent as possible through clinical vigilance—asking the right questions and eliciting the details that lead to the best possible diagnostic plan.
Suggested Reading
American College of Emergency Physicians Clinical Policies Committee: Clinical policy: critical issues in the evaluation and management of adult patients presenting with suspected lower-extremity deep venous thrombosis. Ann Emerg Med 42(2):124, 2003.
American College of Emergency Physicians Clinical Policies Committee: Clinical policy: critical issues in the evaluation and management of adult patients presenting with suspected pulmonary embolism. Ann Emerg Med 41(2):257, 2003.
Brown MD, et al.: Turbidimetric D-dimer test in the diagnosis of pulmonary embolism: a metaanalysis. Clin Chem 49(11):1846, 2003.
Courtney DM, et al.: Pulseless electrical activity with witnessed arrest as a predictor of sudden death from massive pulmonary embolism in outpatients. Resuscitation 49(3):265, 2001.
Hogg K, et al.: Application of pulmonary embolism rule-out criteria to the UK Manchester Investigation of Pulmonary Embolism Diagnosis (MIOPED) study cohort. J Thromb Haemost 3(3):592, 2005.
Huda W: When a pregnant patient has a suspected pulmonary embolism, what are the typical embryo doses from a chest CT and a ventilation/perfusion study? Pediatr Radiol 35(4):452, 2005.
Kline J and Runyon M: Pulmonary embolism and deep venous thrombosis. In: Marx J, et al. (eds): Rosen’s Emergency Medicine: Concepts and Clinical Practice, 6th ed, Mosby/Elsevier, 2006, pp. 1368-1381.
Kline JA, et al.: Clinical criteria to prevent unnecessary diagnostic testing in emergency department patients with suspected pulmonary embolism. J Thromb Haemost 2(8):1247, 2004.
Kline JA, et al.: Criteria for the safe use of D-dimer testing in emergency department patients with suspected pulmonary embolism: a multicenter US study. Ann Emerg Med 39(2):144, 2002.
Kline JA, et al.: More on: clinical criteria to prevent unnecessary diagnostic testing in emergency department patients with suspected pulmonary embolism. J Thromb Haemost 3(1):190, 2005.
Le Gal G, et al.: Reproduction of chest pain by palpation: diagnostic accuracy in suspected pulmonary embolism. BMJ 330(7489):452, 2005.
Richman PB, et al.: Electrocardiographic findings in emergency department patients with pulmonary embolism. J Emerg Med 27(2):121, 2004.
Stein PD, et al.: Multidetector computed tomography for acute pulmonary embolism. N Engl J Med 354(22):2317, 2006.
Wells P, et al. Derivation of a simple clinical model to categorize patient’s probability of pulmonary embolism: Increasing the model’s utility with the SimpliRED D-dimer. Thromb Haemost 83:416, 2000.
Wells P, et al. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. Lancet 350(9094):1795, 1997. |
|
