Approach

Management of pulmonary embolism (PE) is based on an algorithmic diagnostic approach. History and physical exam are relatively insensitive and nonspecific, so must be combined with other diagnostic tests in the clinical decision-making process.

History

History can vary greatly between individuals. Many will report an acute onset of either chest discomfort or dyspnea, but PE may present with more unusual symptoms, or even be asymptomatic.[117]

Pleuritic chest pain and dyspnea are the common presenting features.[117][118]​​ A sense of apprehension is often reported.[4][119]​​ Hemoptysis and syncope are less common; the latter suggests a larger clot burden, more significant right ventricular dysfunction, and poorer prognosis.[3]​​​[117][118]

Risk factors for venous thromboembolism (VTE) should be ascertained.

Signs and physical exam

Signs of PE include tachycardia, tachypnea, increased respiratory effort, fever (usually low-grade), and, in more severe cases, hypotension and signs of hypoperfusion (shock).

Physical exam is often nonspecific.[117] In severe cases, findings of right ventricular overload may be present, such as elevated jugular venous pulsation, loss of palpability of the left ventricular apex (due to posterior displacement of the left ventricle by enlargement of the right ventricle), and a right-sided third heart sound.

PE which has progressed over a period of time may present with physical findings of pulmonary hypertension, such as a right ventricular heave, holosystolic murmur of tricuspid regurgitation, loud pulmonic component of the second heart sound, elevated jugular venous pressure, and pitting edema of the extremities.[120][121][122]​ As PE most often originates from lower extremity deep vein thrombosis (DVT), physical findings of this condition may be present.

Patients with suspected PE with shock or hypotension

Shock (end-organ hypoperfusion and a systolic BP <90 mmHg or vasopressor requirement to maintain systolic BP >90 mmHg) or hypotension (systolic BP <90 mmHg or >40mmHg decrease from known baseline for at least 15 minutes) occurs in a minority of cases but portends a high risk of mortality.[3]​​

More than 95% of patients who present with acute PE are hemodynamically stable.[123]

Ideally, PE should be confirmed by computed tomographic pulmonary angiogram (CTPA) before thrombolytic therapy is administered.[3]​​​[124] However, a negative V/Q lung scan effectively excludes PE, and is a radiation- and medium-sparing procedure; but imaging acquisition takes much longer than CTPA.[3]​​ If the patient is at risk of imminent cardiac arrest, treatment may be commenced on clinical grounds alone.[125][126]

Patients with suspected PE without shock or hypotension

When history and physical exam fail to rule out PE, the pretest probability of PE should be determined using clinical judgement. Where initial clinical assessment suggests a low clinical suspicion of PE, the application of the Pulmonary Embolism Rule-Out Criteria (PERC) is recommended to exclude PE in these patients.[3][127]​​​​ Patients with a high clinical suspicion of PE require further assessment using the original Well criteria (modified), simplified Wells criteria (modified), original Geneva score (revised), the simplified Geneva score (revised) or the YEARS criteria. For more information on these see below.[3][128][129][130]​​ If there is a high probability for PE but diagnostic testing cannot be performed promptly, then empiric anticoagulation at initiation dose should be provided, if not contraindicated, while awaiting results of diagnostic testing.[3]

Diagnostic algorithm for PE

[Figure caption and citation for the preceding image starts]: Summary: pulmonary embolism diagnostic pathwayCreated by BMJ Knowledge Centre [Citation ends].com.bmj.content.model.Caption@49aeb40a

Assessing clinical probability of PE

Clinical probability, assessed by clinical judgement and a validated prediction rule, is the basis for all diagnostic strategies for PE.[3]​ Suspected PE can often be excluded without an imaging test, avoiding the expense and the risks of radiation and contrast from imaging. Therefore, the first step in making the diagnosis of PE is to establish the probability that a PE is present using a risk assessment model. Using an algorithmic approach, more than 40% of patients suspected of PE can have the diagnosis safely excluded without imaging.[127]

Guidelines from the American College of Physicians recommend the application of the PERC to exclude PE in patients where initial clinical assessment suggests a low clinical suspicion of PE.[3][127]​​ One meta-analysis of studies that assessed the accuracy of PERC to rule out PE reported a sensitivity of 97%.[131]

In patients who meet all of the PERC criteria (age <50 years; initial heart rate <100 bpm; initial oxygen saturation >94% on room air; no unilateral leg swelling; no hemoptysis; no surgery or trauma within last 4 weeks; no history of VTE; no estrogen use), the risk for PE is considered to be lower than the risk of testing, and no further testing is indicated.[3][127]

Hemodynamically stable patients who do not meet all the PERC criteria, or patients with high clinical suspicion of PE on initial evaluation, require assessment using the original Wells criteria (modified), simplified Wells criteria (modified), original Geneva score (revised), the simplified Geneva score (revised), or the YEARS criteria.[3][128][129][130]

Each of these clinical decision tools assigns a value (a single point, or points) to a series of historic and physical exam features, the sum of which determines whether PE is likely or unlikely. The YEARS criteria use a subset of the Wells criteria to create two levels of pretest probability, and an adjusted D-dimer threshold to identify patients for whom PE can be safely ruled out without imaging.[130]

​If these scoring systems suggest PE is unlikely, D-dimer measurement is recommended to assess the need for imaging.[3][127]​​ See Criteria.​​ Patients where Wells/Geneva or YEARS criteria suggest that PE is likely, or with an abnormal D-dimer, should proceed immediately to imaging (CTPA; or ventilation-perfusion [V/Q] lung scan if CTPA is contraindicated), as should any patient with suspected PE with shock or hypotension.[127]​ In these patients, anticoagulation should be initiated while awaiting imaging results.[19][20]​​​​ [ Pulmonary Embolism Wells Score Opens in new window ] [ Revised Geneva Score for Estimation of the Clinical Probability of Pulmonary Embolism in Adults Opens in new window ] [ YEARS Algorithm for Pulmonary Embolism (PE) Opens in new window ] ​​​​​

When each clinical decision tool (original Wells, modified Wells, simplified Wells, revised Geneva, and simplified revised Geneva; all with D-dimer testing if PE unlikely) was validated in a primary care dataset, sensitivity ranged from 88% (simplified revised Geneva) to 96% (simplified Wells) and specificity from 48% (revised Geneva) to 53% (simplified revised Geneva).[132]​ The simplified versions of the modified Wells criteria or revised Geneva score may be preferred in clinical practice because of their ease of use.[133]​ Both simplified versions have been validated; neither has been shown to be superior to the other.[3][132]​ However, the Geneva score is based entirely on objective clinical items and may be more reproducible (the Wells criteria [original and simplified] include the subjective clinical item "alternative diagnosis less likely than PE").[134]

The Wells criteria and revised Geneva score categorize patients dichotomously (PE unlikely or PE likely). However, earlier iterations of each tool attributed low, intermediate, or high clinical probabilities of PE. If the two-level classification is used, PE is confirmed in 50% of patients in the PE-likely category compared with 12% in the PE-unlikely category. If the three-level classification is employed, the proportion of patients with confirmed PE will be around 10% in the low probability category, 30% in the intermediate probability category, and 65% in the high probability category.[135]

The YEARS criteria adjusts the D-dimer threshold to exclude PE based on number of criteria present. The criteria are: clinical signs of DVT; hemoptysis; and whether PE is the most likely diagnosis.[130] The YEARS criteria avoids imaging in the highest percentage of patients and so may result in reduced need for CTPA, but maintains the subjective elements from the Wells criteria.[130] The YEARS criteria have been evaluated in pregnant patients suspected of PE.[136]

Confirmation of PE

Confirmation of the diagnosis requires documentation of a blood clot in a pulmonary artery by an imaging study, such as multiple-detector CTPA or V/Q lung scan (radiation-sparing).[3]​ Confirmation of PE with a definitive test is essential because treatment is associated with significant bleeding risk.

Do not routinely order imaging for patients with suspected PE without an intermediate or high pretest probability of PE or with a negative D-dimer.[137][138]​​​​[139]

Quantitative D-dimer level

D-dimer is a breakdown product of cross-linked fibrin; hence, if there is an acute VTE, D-dimer level is likely to be elevated. A quantitative or highly sensitive D-dimer test is therefore a useful test to exclude the presence of an acute VTE. However an elevated D-dimer level is nonspecific, and is frequently abnormal in patients who are older, are acutely ill, have underlying hepatic disease, have an infection, or are pregnant. There are many tests available for D-dimer, but the best are highly sensitive enzyme-linked immunosorbent assay tests. Each of the multiple tests that are available on the market has its own normal cutoff value. D-dimer can be reported in different units, so the specific cutoff value for the test being used should be noted.[140][141]

D-dimer testing is indicated in hemodynamically stable patients who did dot meet all the criteria of Pulmonary Embolism Rule-Out Criteria (PERC) and where PE has been assessed as unlikely using a validated clinical decision tool (original Wells, modified Wells, simplified Wells, revised Geneva, simplified revised Geneva, or YEARS criteria).[3][127]

Clinicians should not obtain a D-dimer level in patients with a high clinical suspicion of PE; immediate imaging is indicated.[127]

D-dimer testing is highly sensitive (>95%) but nonspecific.

A plasma D-dimer level below threshold safely excludes PE in patients with an unlikely (intermediate or low) pretest probability of PE, and no further testing is required.[3][124]​ The risk of PE within 3 months is <1% in these patients.[142][143]

D-dimer may be adjusted to age (normal is <age × 10 micrograms/L in patients aged ≥50 years) or pretest probability of disease (1000 nanograms/mL cutoff in low-probability patients, or 500 nanograms/mL in intermediate-probability patients) to increase the specificity, and thus the percentage of patients who can avoid an imaging study.[3][144][145]​ In patients with cancer, using an age-adjusted D-dimer cutoff doubled the proportion of patients in whom PE could be excluded by clinical decision rule and D-dimer, without imaging.[146]​ The YEARS algorithm with risk-adapted D-dimer thresholds has been studied in pregnant patients suspected of PE.[136]

Patients with an abnormal D-dimer level should undergo multiple-detector CTPA (or V/Q lung scan if CTPA is contraindicated) to confirm or exclude a diagnosis of PE.[3][124][127]

Initial imaging studies - CTPA and V/Q Scan

CTPA has the best diagnostic accuracy of all advanced noninvasive imaging methods; CTPA confirms the diagnosis by direct visualization of thrombus in a pulmonary artery, where it appears as a partial or complete intraluminal filling defect.[139][147]​​​​​ The likelihood ratio to rule in a PE with a filling defect in the segmental or subsegmental branches is 24.1 (range of 12.4 to 46.7), whereas the likelihood to rule it out is 0.11 (range of 0.06 to 0.19).[147] Specificity is 96%.​[148]​ Three-month incidence of a subsequent venous thromboembolic event following a negative computed tomography (CT) scan is <2%.[147][149]

CT scans are contraindicated in approximately 25% of patients due to contrast allergy or renal failure.[148]​ In pregnant women, radiation exposure through CT scan is at a dose much lower than the exposure associated with fetal harm and, if necessary, CTPA should not be withheld.[150]

CTPA should be used with discretion in adolescents and young adults, and alternative imaging strategies employed where possible.[151][152]​​ If it is used, ionizing radiation exposure should be carefully monitored and minimized.[153][Figure caption and citation for the preceding image starts]: Contrasted CTPA scan showing subsegmental right pulmonary artery emboli (see arrows)From the collection of Seth W. Clemens; used with permission [Citation ends].com.bmj.content.model.Caption@5b593d75

V/Q lung scan, preferably using single-photon emission computed tomography (SPECT; which may reduce the number of inconclusive scans), is an alternative to CTPA.[139][154]​​​ One prospective study of patients with suspected acute PE reported a sensitivity of 97% and a specificity of 88% for V/Q SPECT scan.[155] In one retrospective study of 2328 patients with clinically suspected PE, 601 of 608 patients with a final diagnosis of PE had a positive V/Q SPECT scan (99% sensitivity), and 1153 patients without final PE diagnosis had a negative V/Q SPECT scan (98% specificity).[156]​ A negative V/Q scan effectively excludes PE. V/Q scan is a radiation- and medium-sparing procedure and may be appropriate for patients with contraindications or relative contraindications to CT (e.g., contrast allergy, moderate to severe renal failure, pregnancy, young patients).[3]​​​ Consider a V/Q lung scan rather than CTPA to diagnose PE in young women with a normal chest radiograph to reduce the overall radiation dose to the breast.[3][157]​​​​​ Methods include planar V/Q, single photon emission CT (SPECT) and SPECT with concomitant low-dose computed tomography (Q-SPECT-CT). SPECT based methods have fewer indeterminate results.[158]

Other imaging studies

Point-of-Care Ultrasound (POCUS)

POCUS is a rapid imaging study performed at the bedside by the physician or other trained clinician. Guidelines support its use to evaluate potential etiologies for acute dyspnea, but the precise fit into the diagnostic algorithm for patients clinically suspected of PE requires further exploration.[159]

Chest x-ray

A normal chest x-ray does not confirm or eliminate PE as a diagnosis.[139] However, a chest-ray may rule out other causes of a patient’s symptoms such as pneumothorax or pneumonia.[139]​ Chest x-ray findings associated with PE include the Fleischner sign/prominent central pulmonary artery (20%); Westermark sign/oligemia in PE area of distribution (11%); Hampton hump/pleural-based areas of increased opacity corresponding to the distribution of the PE (27%).[129][160]​​ None of these findings are sufficiently sensitive or specific to allow a definitive diagnosis.

Magnetic resonance angiography (MRA)

MRA can be used to evaluate the central and segmental arteries, magnetic resonance (MR) based imaging is rarely used due to the ease and performance characteristics of CTPA.[139] Three different techniques are available: gadolinium-contrast enhanced angiography (Gd-MRA), real-time angiography (RT-MRA), and MR-perfusion images.[161]​ MRA may be used to evaluate the central and segmental arteries.[139]​ The test has high specificity (91% to 98%) but low sensitivity (75% to 93%), and so cannot reliably exclude PE with a negative test.[161]​ Gadolinium (Gd) contrast, used in Gd-MRA and MR perfusion studies, is relatively contraindicated in pregnancy. Due to cost, longer acquisition time and operating characteristics, MR modalities are used infrequently.

Pulmonary angiography

Despite its diagnostic accuracy, invasive pulmonary angiography is rarely used for the diagnosis or exclusion of PE.[3][162]​​ It is associated with risk of morbidity/mortality, and (less invasive) CTPA affords comparable diagnostic precision.[3][163]​​ The negative predictive value may be as high as 99%, however, this is an invasive test with morbidity of 3% to 6% and a mortality of 0.2% to 0.5%.[163][164][165]​​​ Involves the use of contrast; a relative contraindication in pregnancy and renal failure.

Transthoracic echocardiography (TTE)

TTE is generally not indicated in the diagnosis of acute PE, but it is useful in identifying right ventricular strain and assisting with severity classification and determining prognosis.[3]​​[139] Echocardiographic evidence of right heart thrombi is significantly associated with increased 30-day mortality in patients diagnosed with acute PE.[166][Figure caption and citation for the preceding image starts]: Gd-MRA showing a right main pulmonary artery pulmonary embolism (see arrow)From the collection of Seth W. Clemens; used with permission [Citation ends].com.bmj.content.model.Caption@5d43cfbc​ ​ If a definitive imaging modality is unavailable, echocardiography may be considered for patients with suspected PE presenting with shock or hypotension.[3][124]

Ultrasound

Bilateral venous compression ultrasound to establish the presence of thrombosis suggestive of PE is recommended in pregnant patients with suspected PE (see special patient populations section below).​​[3][167]

Laboratory investigations

Baseline laboratory tests including prothrombin time (PT), activated partial thromboplastin time (aPTT), and international normalized ratio (INR) are important to aid decisions about the safety and type of initial anticoagulation selected. Prolonged values, especially aPTT, in the absence of exposure to an anticoagulant, may suggest antiphospholipid antibody syndrome. Renal and hepatic function panels also help determine the appropriate choice of anticoagulant therapy, as different agents carry precautions or are contraindicated in renal or hepatic dysfunction.[168]​ Complete blood count may detect hematologic abnormalities. Heparin therapy can be associated with heparin-induced thrombocytopenia; platelet count should be measured at baseline and regularly throughout heparin treatment. Elevated platelet count may suggest essential thrombocytosis or a myeloproliferative disorder, which may represent a secondary hypercoagulable state.

Troponin can be used to assist in determining the severity category of acute PE, which impacts management decisions. It is suggested in patients who have either an elevated Pulmonary Embolism Severity Index (PESI) category, or abnormalities of the right ventricle on imaging.[3]

Thrombophilia screen

Thrombophilia commonly refers to five hereditary conditions (Factor V Leiden, prothrombin gene 20210A, deficiencies in antithrombin, protein C deficiency, and protein S deficiency) and antiphospholipid syndrome (an acquired condition). However, many gene variants and acquired conditions modify thrombosis risk.[169]

Indications for screening are controversial.[169][170]​​ Hereditary thrombophilia does not sufficiently modify the predicted risk of recurrent thrombosis to affect treatment decisions, and a conservative approach to testing is reasonable.[19] If a thrombophilia screen is indicated, it should be deferred until a minimum of 3 months of anticoagulant therapy has been completed because some thrombophilia tests are influenced by the presence of acute thrombosis or anticoagulant therapy.[169]  Some guidelines suggest testing only in situations where the result is likely to change a clinical decision (such as in patients with unprovoked DVT or PE who are considering stopping anticoagulants).[20][61][171]​​​​ However, the presence of a hereditary thrombophilia does not significantly increase the predicted risk of recurrent VTE after a provoked VTE, and guidelines discourage testing in this setting.[3][60]

Antiphospholipid syndrome

Antiphospholipid antibodies may predict a higher risk of future thrombosis following an initial VTE event and may impact selection of therapy.[65] Controversy exists regarding whether broad screening for antiphospholipid antibodies or screening only on the basis of clinical suspicion should be preferred.[172][173]​​ Some guidelines suggest testing only in situations where the result is likely to change a clinical decision (such as in patients with unprovoked DVT or PE who are considering stopping anticoagulants, however these guidelines recommend seeking specialist advice as these tests may be affected by anticoagulants and acute thrombosis).[20][169]

For antiphospholipid antibody screening, cardiolipin and beta-2 glycoprotein-I antibodies can be performed without regard to the presence of anticoagulants; however, most anticoagulants interfere with assays for lupus anticoagulant.[65]

Other investigations

Electrocardiography (ECG)

ECG cannot definitively establish or eliminate PE as a diagnosis, and specific findings may only be suggestive of PE.[118][174][175]​​ ECG can, however, be used to assess right ventricular function in patients with confirmed PE without shock or hypotension.[3]​​[19]​​​ Right ventricular dysfunction is predictive of adverse outcome and enables risk stratification in these patients.[176][177][178]

Arterial blood gasses (ABG)

ABG analysis is of very limited diagnostic utility, alone or in combination with other clinical variables, in suspected PE.[179]​ Hypoxemia is considered to be a typical finding in acute PE, but ABG analysis is of very limited diagnostic utility, alone or in combination with other clinical variables, in suspected PE.[3][179]

A PaO₂ <80 mmHg, a PaCO₂ <36 mmHg, or an abnormal alveolar-arterial gradient (A-aO₂) are not predictive of PE in patients suspected of having PE.[179]​ In patients with suspected acute PE with normal ABG results, PE could not be excluded in 38% of those without cardiopulmonary disease and 14% with preexisting cardiopulmonary disease, respectively.[180]

Tests for underlying conditions

Occult cancer is present in approximately 3% to 5% of patients with an unprovoked DVT.[181]​ However, extensive investigations (beyond routine laboratory tests and age-appropriate routine screening) for cancer in patients with a first unprovoked DVT are not routinely indicated, because they have not been convincingly shown to improve prognosis or mortality.[182][183][184][185]​​ Signs or symptoms that suggest a possible malignancy should be pursued, if present.

Special patient populations

Pregnancy

Symptoms and signs of VTE may be less specific in pregnant women than in nonpregnant patients.[124]

D-dimer levels increase through normal pregnancy, complicating its use as a test to exclude suspected PE.[3]​ However, increasing evidence supports the use of pretest probability algorithms adapted to pregnant patients with D-dimer.[136][186]​ Guidelines support use of multi-step algorithms over universal imaging in pregnant patients with suspected PE.[3]​​

When an algorithm indicates that imaging is indicated in a pregnant patient suspected of PE, exposure to radiation-associated imaging should be minimized. Bilateral venous compression ultrasound to establish the presence of thrombosis suggestive of PE is recommended in pregnant patients with suspected PE, as identifying DVT can prompt institution of therapy without the need to confirm PE. However, normal ultrasound does not exclude PE.[3]​​[167] Chest x-ray is the first radiation-associated procedure if PE is suspected.[167]  In the setting of a normal chest x-ray, American Thoracic Society consensus guidelines recommend lung scintigraphy (with V/Q scan).[167] Pregnant women with a nondiagnostic V/Q scan, in whom further investigation is deemed appropriate, may undergo CTPA.[167] European Society of Cardiology guidelines suggest that CTPA in patients with abnormal chest x-ray and either V/Q scan or CTPA if chest x-ray is normal.[3]​ The administered dose of radiopharmaceutical should be reduced by a factor of 2 when lung scans are indicated in pregnant women; longer acquisition times should be used to achieve adequate imaging.[154] 

Adolescents and young adults

CTPA should be used with discretion, especially if PE can be ruled out by other noninvasive methods with less radiation exposure.[151][152]​ If it is used, ionizing radiation exposure should be carefully monitored and minimized.​[153]​​

Use of this content is subject to our disclaimer