What every physician needs to know:

Venous thromboembolism (VTE) includes the spectrum of deep venous thrombosis (DVT) and pulmonary embolism (PE). Acute pulmonary embolism (PE) is responsible for 100,000 to 300,000 deaths per year in the US. It is commonly not diagnosed or even suspected until after the patient dies. When it is suspected, patients should undergo diagnostic testing, and if there is a high clinical suspicion and low perceived risk of bleeding, initiation of therapy should be considered even before securing a diagnosis.

Acute PE commonly occurs in hospitalized surgical and medically ill patients. It is usually, but not always, associated with specific risk factors. It is less common in the pediatric setting, but it can occur in patients in any medical or surgical specialty setting. The spectrum of presentations ranges from minimal symptoms to massive emboli that cause sudden death or that progress rapidly to death from right-heart failure. The epidemiology and pathophysiology of acute PE, as well as the clinical manifestations and diagnostic approach to this disease, are discussed below.


Acute PE is a worldwide disease, although the risk is somewhat lower in Asian populations. The vast majority (95%) of cases of acute PE originate from thrombi in the leg or pelvic veins. However, emboli may arise from other sources, such as the axillary subclavian system or renal veins. There is clear comorbidity associated with the chronic aspects of this disease including post-thrombotic syndrome and chronic thromboembolic pulmonary hypertension, as well as a less clearly defined entity referred to as chronic thromboembolic disease (CTED), which includes patients with persistently abnormal lung imaging, but without pulmonary hypertension, but we focus on patients who present with suspected acute PE.

Continue Reading

Patients with acute DVT and/or PE often have one or more underlying risk factors that arise from Virchow’s triad of stasis, venous injury, or hypercoagulability (thrombophilia). Thus, reduced mobility, trauma, and malignancy are among the common predisposing factors (Table 1). Certain clinical settings, such as hip fracture or total hip or knee replacement, place patients at substantial risk for VTE. Such settings should serve as “red flags” when a patient presents with compatible symptoms.

Table I.
Hereditary factors**Antithrombin deficiencyProtein C deficiency Protein S deficiencyFactor V Leiden Activated protein C resistance without factor V LeidenProthrombin gene mutationPlasminogen deficiencyDysfibrinogenemiaAcquired factors*Reduced mobilityAdvanced ageCancerAcute medical illnessMajor surgeryTraumaSpinal cord injuryPregnancy and the postpartum periodOral contraceptivesHormone replacement therapyPolycythemia veraAntiphospholipid antibody syndromeHeparinsChemotherapyObesityCentral venous catheterizationImmobilizer or castProbable factorsElevated homocysteineElevated factors VIII, IX, XIElevated fibrinogenElevated thrombin-activated fibrinolysis inhibitorLow levels of tissue factor pathway inhibitor

*In a compatible clinical setting, acute DVT and/or PE should be considered even in the absence of known risk factors.

**It remains unclear whether some of the disorders listed above are hereditary, acquired, or both.

Death from acute PE is caused by right ventricular failure. The right ventricle is not accustomed to pumping against a significantly increased afterload, so it becomes dilated and hypokinetic with large embolic burdens. When the clot burden reaches a critical threshold, the right ventricle is unable to generate enough force to achieve an adequate cardiac output and it fails, resulting in hypotension and cardiac arrest that often manifests as pulseless electrical activity.

Therefore, a key part of risk-stratifying patients with acute PE (i.e., determining whether thrombolytic therapy or embolectomy should be considered) is to ascertain the status of this chamber, generally by echocardiography. Because of obstructed pulmonary arteries, dead space ventilation is an important contribution to the pathophysiology. In most patients, this ventilation perfusion mismatch results in hypoxemia and an increased alveolar-arterial gradient.

Are you sure your patient has acute PE? What should you expect to find?

Approximately 90% of patients with acute PE present with acute dyspnea, which, if unexplained, should prompt a diagnostic evaluation. Chest pain, a common symptom in acute PE, may manifest as vague chest discomfort or as pleuritic chest pain caused by smaller peripheral emboli that cause pulmonary infarction. Patients may also develop low grade fever with infarction, and/or hemoptysis. High fever almost always indicates a process other than PE or an associated process. Palpitations may occur related to sinus or atrial tachycardia.

Patients with large clot burdens may develop angina-like pain that may arise from right ventricular ischemia. Lightheadedness, presyncope, and/or syncope may occur, suggesting more extensive embolism. Cough is common in the setting of acute PE, but as a symptom, it is nonspecific and may or may not be due to PE. Pulmonary infarction may be associated with cough. While the vast majority of patients with acute PE have a lower-extremity source, there may or may not be symptoms or signs of acute DVT. When present, these symptoms may manifest as calf pain and swelling or simply as the “charlie horse” sensation of a pulled muscle.

Physical examination may reveal a calm, mildly symptomatic patient, and an anxious individual in extremis, or anything in between. Unexplained tachycardia suggests the possibility of acute PE. Younger patients may have relative increases in heart rate from baseline, but may maintain rates below 100/min. Larger emboli that cause right ventricular dysfunction may cause hypotension. This symptom merits prompt evaluation, as massive PE (PE associated with hemodynamic compromise) may require aggressive measures, such as thrombolytic therapy or pulmonary embolectomy. Tachypnea is common in acute PE, but it may not be present. Chest wall tenderness can occur with acute PE because of pulmonary infarction.

As suggested above, clinical evidence of DVT is often absent. When cardiac arrest occurs with acute PE, it may manifest as pulseless electrical activity or simply as asystole. A smaller percentage of patients may experience ventricular tachycardia or fibrillation. Patients with acute PE may present immediately with the onset of symptoms. However, as many as 25% identify the onset of their symptoms as more than two weeks prior to the time of diagnosis.

When PE is suspected, clinical assessment alone cannot confirm or exclude it. The history, risk factors, physical exam, and ancillary studies should be integrated to form a differential diagnosis and determine the need for specific testing for acute PE. Because a number of factors must be considered, formulation of a pretest probability can facilitate the clinician’s approach. This can be done simply by gestalt, weighing in the clinician’s experience, comfort level with the disease, and knowledge of the PE literature.

However, increasing data support the use of clinical prediction models to guide the diagnostic approach. The most widely studied models include the Wells score, the PERC score, and the Geneva scores. While these models have clear utility, a high level of clinical suspicion for acute PE should not be ignored solely because a clinical predictive model suggests that it can be ignored.

Wells Score

Wells and colleagues, created a clinical prediction rule for suspected acute PE and tested it nearly two decades ago, but the scoring system has evolved over time. They subsequently simplified the model by using logistic regression analysis to select seven variables that were significantly related to PE (Table 2).

Table II.
PE is the most likely diagnosis = 3 pointsSymptoms and signs of DVT are present = 3 pointsHeart rate higher than 100 / minute = 1.5 pointsImmobilization for at least three days = 1.5 pointssurgery in the previous four weeksPrevious objectively diagnosed DVT or PE = 1 pointHemoptysis = 1 pointMalignancy with treatment within six months = 1 point

In the validation cohort for the newest model, a score of less than 4 points (PE unlikely) combined with a negative Simpli-Red D-dimer assay (not an ELISA-based assay) accurately excluded a diagnosis of acute PE in 98% of patients. As per the first three-point item in the score, gestalt is part of the method, so it is not entirely objective. It has also been suggested that, commonly, the subjective 3-point “PE is the most likely diagnosis” is what tips the score in favor of PE.

Consider a 40-year-old obese woman on oral contraceptives who presents with sudden-onset dyspnea and a heart rate of 96/minute, much faster than her usual baseline. With PE being considered the most likely diagnosis, her Wells score would be only three (PE unlikely). In this situation, a D-dimer test would be done. If PE were, in fact, present, the D-dimer would very likely be positive, and imaging would then be performed. In such a classic setting, many clinicians would simply do the imaging study without the score and without the D-dimer.


The PERC rule was designed to rule out acute PE without further testing in patients who present to the emergency room. The eight variables are listed in Table 3.

Table III.

Age less than fifty yearsPulse less than 100/minuteOxygen saturation greater than 94 percentAbsence of:

Unilateral leg swelling


Recent surgery

Prior DVT / PE

Oral contraceptive use

As a diagnostic test, low suspicion and a negative PERC status have been shown to have a sensitivity of 97.4% (CI 95.8% to 98.5%) and specificity of 21.9% (CI 21.0% to 22.9%). That is, a gestalt estimate of low clinical suspicion and PERC negative status reduced the probability of acute PE to 2% in about 20% of outpatients with suspected PE. However, recent data applied retrospectively to consecutive patients who presented with suspected PE suggest that the PERC rule alone or combined with another pretest probability approach (revised Geneva score) cannot safely identify very low-risk patients in whom PE can be ruled out without additional testing, at least in populations with a relatively high prevalence of acute PE.

Should PERC be used? As with other PE predictive models, it is useful as a guide. However, some patients who have entirely negative PERC scores also prove to have PE. A 45-year-old man with active colon cancer has had nausea related to chemotherapy and has spent the past week on the couch at home. He presents with sudden-onset dyspnea and pleuritic chest pain. In the emergency department, his heart rate is 90/minute, and his O2 saturation is 97 percent. His PERC score is zero, yet he has several clear VTE risk factors, including active cancer and reduced mobility. He is young, and it is not surprising for a young individual with baseline normal cardiopulmonary reserve to have a heart rate of <100 / minute and a normal O2 saturation in the setting of acute PE. In fact, his arterial blood gas reveals a normal alveolar-arterial gradient.

Chest-computed tomography angiogram (CTA) revealed acute PE in this patient. To use the PERC rule, however, the clinician must first rate the patient as low clinical risk for PE. In this case, most clinicians would have rated the patient as high risk, eliminating the use of the score. If we consider the same 45-year-old man with sudden onset dyspnea, pleuritic chest pain and cough, but without cancer and reduced mobility, whom a clinician believes may have infection, a PERC score of zero might be followed by antibiotic therapy and no PE evaluation. This scenario could be acute idiopathic PE. Therefore, while these scoring systems are useful guides, they are not foolproof.

Geneva Scores

The Geneva score was originally designed as a somewhat complex clinical prediction rule that required arterial blood gas analysis. It did not include a subjective analysis of whether PE was most likely, as the Wells score did. It was ultimately revised to include only clinical data, as with the Wells score, and it was more recently simplified (Table 4). The Geneva score has similarities to the Wells score, and a recent study suggests that the Wells rule may be more accurate among inpatients and patients presenting to the emergency department, while the revised Geneva score can be used in the emergency department with high reliability.

Table IV.
Revised Geneva score*Variable ScoreAge 65 years or older = 1 pointPrevious DVT or PE = 3 pointsSurgery or fracture within the last month = 2 pointsActive malignancy = 2 pointsHemoptysis = 2 pointsHeart rate 75-94/minute = 3 pointsHeart rate higher than 95 / minute = 5 pointsUnilateral lower limb pain = 3 pointsPain on deep palpation of lower limb and unilateral edema = 4 point0-3 points low probability for acute PE (8%)4-10 points = intermediate probability (28%)more than 10 points = high probability (74%)

* The simplified Geneva score includes the same parameters as the revised score, but the score for each parameter is uniformly one point, and if the heart rate is higher than 95/minute, an additional point is added. It is suggested that the likelihood or PE in a patient with a simplified Geneva score of less than 2 and a normal D-dimer is 3 percent.

In summary, clinical prediction rules can be very useful, although there is not an established standard of care that requires them. The prediction rules described above have been the most widely studied. Clinical gestalt, even while utilizing a score, is very important.

Beware: there are other diseases that can mimic acute PE.

Pulmonary embolism may be mistaken for pneumonia, asthma, bronchitis, a COPD flare, congestive heart failure, acute myocardial infarction, and other cardiopulmonary disorders associated with dyspnea or chest pain, as well as nephrolithiasis. If fever and cough dominate the clinical presentation, then infection is most likely.

However, patients may present with underlying infection or congestive heart failure and, based upon that illness, their reduced mobility may result in their developing acute PE which may be mistaken for infection or heart failure alone. Postoperative patients may be at risk for pneumonia as well as PE, and depending on the clinical scenario, both may need to be considered. While PE may be associated with wheezing, wheezing is much more commonly present in the setting of other diagnoses.

How and/or why did the patient develop acute PE?

Venous thrombosis occurs as a result of one or more of the following abnormalities, which are termed Virchow’s triad:

1. Hypercoagulability

2. Stasis of venous blood flow

3. Venous endothelial injury or dysfunction

Which individuals are at greatest risk of developing acute PE?

The risk of acute VTE increases with age, probably because of reduced mobility and increased comorbidity. The risk may become more important at approximately age 40, increasing exponentially in subsequent decades. However, a patient’s young age should not deter a clinician from considering acute VTE in a compatible setting. Pregnancy and the postpartum state also increase the risk, as do several inherited and acquired thrombophilias. Long-distance travel appears to be a risk, although patients who get VTE with travel as the sole risk factor appear to have unusual susceptibility.

What laboratory studies should you order to help make the diagnosis, and how should you interpret the results?

Laboratory testing may be helpful in establishing or refuting the diagnosis of acute PE, but they cannot rule PE in with certainty. Severe anemia may be found in ill patients at high risk for acute VTE and may serve as an explanation for dyspnea, but its presence does not rule out concomitant PE. Leukocytosis is much more common with infection than with acute VTE; a white blood cell count greater than 20,000 is present in less than 20% of cases of acute PE.

D-dimer testing has been used widely for ruling out acute VTE. The most sensitive assays have been quantitative D-dimers including enzyme-linked immunosorbent assays (ELISA) and turbidometric assays with sensitivities of greater than 95%. The D-dimer assay is best utilized in patients with low or moderate clinical probability, although some data suggest excellent sensitivity and specificity regardless of pretest probability.

Despite the excellent sensitivity of these assays, when clinical suspicion is high for acute VTE, imaging studies should be performed to rule out VTE definitively. Recent clinical data suggest that higher levels of D-dimer may be more specific for acute PE; nonetheless D-dimer cannot be used to rule in PE. While D-dimer testing in the setting of pregnancy and suspected acute PE appears to be sensitive, there are case reports of false negative tests, and D-dimer is not currently recommended in this setting. The D-dimer test positivity increases progressively in pregnancy such that, by the third trimester, the majority of individuals have a positive test even without thromboembolism.

Arterial blood gas analysis may demonstrate hypoxemia and hypocapnia (i.e., an abnormal alveolar-arterial gradient), but these may be normal as well, particularly in younger patients without cardiopulmonary disease. Because many cardiopulmonary diseases are associated with abnormal gas exchange and ventilation-perfusion mismatching, the specificity for these abnormalities is poor. In the setting of a normal or near-normal CXR and significant unexplained hypoxemia however, PE should be considered.

Patients who present with chest pain may have a positive troponin, which should not automatically be assumed to represent left ventricular ischemia. The serum troponin may be positive in acute PE; this positive reading generally occurs in the setting of a substantial clot burden, resulting in right ventricular ischemia. The brain natriuretic peptide level may also be elevated in acute PE because of right ventricular dilation, which may serve as a clue to the diagnosis but which is, again, nonspecific.

What imaging studies will be helpful in making or excluding the diagnosis of acute PE?

An established diagnosis of acute DVT or PE requires either radiographic imaging or autopsy proof. While a high level of clinical suspicion is crucial, and while treatment should be considered in such settings even before the diagnosis is established, imaging is required for proof and for continued therapy.

The CXR is often abnormal in acute PE, but it may also be normal or minimally abnormal, demonstrating, for example, atelectasis alone. Pulmonary infarction may be associated with pleural-based wedge-shaped infiltrates (Hampton’s hump), which may be mistaken for pneumonia. In some cases, pulmonary infarction may show cavitation by imaging. A paucity of lung markings (Westermark’s sign) may suggest PE, but again, these are all nonspecific findings.

In patients who present with suspected acute PE, chest CTA has become the standard diagnostic test. However, the ventilation perfusion (VQ) scan still has substantial utility in certain settings.

Echocardiography may also establish the diagnosis in certain settings, such as when emboli in-transit are visualized in the right atrium. In addition, “McConnell’s sign,” defined as right ventricular free-wall hypokinesis in the presence of normal apical contractility, appears to be specific for acute PE (as opposed to settings in which chronic pulmonary hypertension occurs), though it has been shown that the same echocardiographic appearance can occur in acute right ventricular infarction.

When patients present with suspected acute DVT, an aggressive approach to leg imaging, most commonly compression ultrasound, should be taken. When ultrasound is used in this setting (as opposed to screening in asymptomatic legs), it is highly sensitive and specific.

Ventilation Perfusion Scanning

The utility of VQ scanning was clearly established in the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) study published in 1990. The study provided evidence that the majority of patients with acute PE had non-high-probability VQ scans based on subsequent pulmonary angiography. A high level of clinical suspicion in the setting of a nondiagnostic VQ scan should lead to another imaging study (CTA, pulmonary angiography, or leg imaging).

The presence of underlying lung disease is the most common reason for a nondiagnostic scan. Therefore, when conditions like emphysema or pulmonary fibrosis are present, CTA should be considered. Ideal candidates with suspected PE to consider for VQ scanning are younger patients–generally, those under age 40 who have no underlying cardiopulmonary disease. While most centers do both a ventilation and perfusion scan in suspected acute PE, a normal perfusion scan alone rules out acute pulmonary embolism. Portable perfusion scanning can be done at some centers and may be useful in very ill patients who cannot be moved. While VQ scanning may be nondiagnostic in settings of severe deterioration, with severe hypoxemia and/or hypotension, a scan without large defects may be enough to rule out PE as a cause of the deterioration.

Computed Tomographic Angiography

CT technology and its use in suspected acute PE has advanced significantly over the past decade. With the evolution of multi-array scanners, a good-quality CTA that is negative for acute PE essentially rules out the diagnosis. Specificity is excellent as well. CTA is also highly useful in demonstrating other potential causes of dyspnea and chest pain. Small, subsegmental emboli are difficult to visualize, so when the study is suboptimal or there is doubt, additional lung or leg imaging should be considered.

Another utility of CTA may be to evaluate right atrial and ventricular size, although standard transthoracic echocardiography offers more useful information, including right ventricular contractility. Intracardiac emboli are easily visualized as well. With the introduction of dual-energy CT technology, electrocardiographic-gated CTA may become a practical way to evaluate cardiac function in acute PE. Furthermore, dual energy CTA offers information on both clot burden and perfusion. This scanning technology is more expensive but could become more mainstream in the near future.

Imaging of the leg veins by CT venography can also be performed to establish the diagnosis of concomitant DVT or to look for DVT when the chest CTA is negative, but radiation exposure is increased, so such imaging has generally not been deemed necessary. Recent data suggest that mortality that is due to acute PE is higher in the setting of residual DVT, so the concept of evaluating the legs in acute PE may require additional consideration.

Incidental PE is sometimes discovered in patients in whom CTA is done for another reason. In some patients, therapy has been recommended (American College of Chest Physicians guidelines, 2016), but the risks and benefits have not been clearly established. There is risk of long-term anticoagulation, so such cases are best individualized. For example, in patients in whom incidental PE is discovered during cancer staging, the risk of VTE is increased and therapy may be appropriate. Randomized clinical trials should be considered for incidentally discovered acute PE.

Patients should be questioned regarding any history of iodine/contrast allergy. Steroids can be administered to reduce the risk of adverse effects. Risk factors for contrast-induced renal insufficiency include a creatinine of more than 1.5 mg/dL, dehydration, diabetes mellitus, and age of more than 70 years.

Traditional Pulmonary Angiography

While traditional pulmonary angiography has been the gold standard for establishing the diagnosis of acute PE for decades, it is rarely done now in this setting. It requires more expertise and support staff than CTA and is not available at smaller institutions, especially at night. It is most commonly done now in the setting of catheter-directed techniques for more extensive emboli. While there are no clear guidelines for such approaches, they offer the potential for mechanical disruption (fragmentation and/or extraction), thrombolysis with lower than traditional doses of thrombolytic agents (potentially reducing bleeding risk), or both.

Magnetic Resonance Imaging

Magnetic resonance angiography (MRA) has been studied in the setting of suspected acute PE. This technique takes more time to complete than CTA, and the diagnostic yield has been shown to be institution-dependent. Furthermore, with increasing information about nephrogenic fibrosing dermopathy (caused by gadolinium exposure in the setting of baseline renal insufficiency), enthusiasm has waned. MRA, which is highly sensitive for acute DVT, visualizes the deep veins up into the inferior vena cava, with better views proximally than can be obtained with ultrasound. However, ultrasound is simpler, faster, and adequate in the majority of cases of suspected acute DVT.

Compression Ultrasound

Patients who present with suspected acute PE should undergo chest imaging unless the clinical probability is deemed low or moderate and D-dimer testing is negative. If chest imaging cannot be performed for some reason, ultrasound of the legs can be performed to rule in DVT and establish the need for therapy. While Doppler tracings are done, and direct visualization of thrombi can also sometimes be accomplished, compression is the most sensitive and specific technique.

If tests for thrombi are positive, treatment can be initiated (or continued). However, if they are negative, DVT (and PE) cannot be excluded with certainty. Ultrasound imaging of asymptomatic legs (i.e., screening) may be less sensitive than it is when patients present with pain, tenderness, and/or swelling, particularly in the setting of morbid obesity. However, when patients present with leg symptoms and suspected DVT, the sensitivity and specificity of ultrasound are excellent.

What non-invasive pulmonary diagnostic studies will be helpful in making or excluding the diagnosis of acute PE??

Patients with acute PE may demonstrate oxygen desaturation on pulse oximetry but normal oxygen desaturation does not reliably exclude the diagnosis. The O2 saturation does not take into account the pCO2. That is, a patient may have acute PE and have an O2 saturation of 99% but be working hard to achieve that; that is, by arterial blood gas, the pCO2 may be 20 mm Hg. Practitioners should not rely solely on the O2 saturation. In fact, in proven PE, the alveolar-arterial (A-a) gradient may actually be normal; that is, pO2 and pCO2 may both be normal. Thus, while a normal oxygen saturation or A-a gradient is reassuring, it does not rule out acute PE.

Electrocardiography may be normal in patients with acute PE, although it may reveal sinus tachycardia or an atrial arrhythmia. In addition, the S1Q3T3 pattern may be present, but these are all nonspecific findings. With extensive emboli, a right ventricular strain pattern may be present.

What other diagnostic procedures will be helpful in making or excluding the diagnosis of acute PE?

Occasionally, very proximal PE may be seen on echocardiography, particularly on transesophageal echocardiography. Right heart clot-in-transit may lead to the diagnosis of concomitant PE. No other diagnostic procedures assist the diagnosis of acute PE.

What pathology/cytology/genetic studies will be helpful in making or excluding the diagnosis of acute PE?

This category of studies does not assist the diagnosis of acute PE.

If you decide the patient has acute PE, how should the patient be managed?

Please see the chapter, “Acute Pulmonary Embolism: Prevention and Treatment.”

What is the prognosis for patients managed in the recommended ways?

Please see the chapter, “Acute Pulmonary Embolism: Prevention and Treatment.”

What other considerations exist for patients with acute PE?

While much of the patient education in the setting of acute PE rests on the importance of understanding and adhering to the therapeutic regimen, knowledge of the disease process and characteristic symptoms is also important. Patients should understand that there is a risk of recurrence both on and off therapy. While the initial presentation might have been sudden onset dyspnea without chest pain, recurrence could manifest as pleuritic pain with hemoptysis of leg pain and/or swelling. Patients should realize that symptoms suggestive of recurrence should be investigated immediately.