Clinical Presentation

Acute pulmonary embolism (PE) is the third most common acute cardiovascular disease, after myocardial infarction and stroke.¹ It occurs when a blood clot dislodges from a vein, travels through the venous system, and lodges in the blood vessels of the lung.² Blockages of the pulmonary artery and its branches can lead to obstruction of blood flow to the heart. Resultant pressure in the lungs may increase right heart pressure, causing right ventricular strain, which can lead to cardiovascular compromise and hypoxemia.³ Other complications include pulmonary hemorrhage and infarct. PE is a major cause of emergency hospitalization, with clinical expression ranging from asymptomatic disease to sudden death. Acute PE can lead to chronic pulmonary hypertension, post-thrombotic syndrome, and right ventricular failure if it is not promptly diagnosed and treated.⁴ Further, untreated PE can be fatal in up to 30% of patients. If administered quickly, anticoagulation therapy is highly effective at preventing extension of thrombus and can prevent mortality and morbidity associated with PE.⁵

PE is part of the continuum of venous thromboembolism (VTE), which also includes deep vein thrombosis (DVT).⁶ Most PEs originate from thrombi in the leg or pelvic veins that have dislodged. Evidence of lower-limb DVT is found in about 70% of patients who have sustained a PE.¹ However, asymptomatic PE can also present in patients without DVT.^7,8

Assuming the incidence rate in Canada is similar to that in the United States (US), PE likely afflicts between 0.1% and 1% of the population.⁹ An accurate estimate of PE incidence is difficult to obtain, because a large proportion of pulmonary emboli are detected on autopsy,¹⁰ and not all of these cases are clinically relevant. About 80% of patients identified with PE at autopsy are unsuspected or undiagnosed before death.⁹

There are many challenges associated with diagnosing PE, one being the non-specific nature of common PE symptoms. The most common symptoms include dyspnea and chest pain.^11,12 The underlying cause of PE-related symptoms could be a plethora of alternative conditions, including rib or vertebral body fracture, acute myocardial infarction, pulmonary edema, pneumonia, neoplasm, or interstitial lung disease. Approximately 30% of patients with PE may be asymptomatic.¹³ Unspecific patient symptoms can potentially lead to over-testing, as PE may be considered in the differential diagnoses of a range of symptoms.

PE rarely occurs in the absence of risk factors and the likelihood of occurrence increases progressively where multiple risk factors are present. Factors associated with the development of PE can be inherited or acquired. They include, but are not limited to, malignancies, immobilization, surgery, extremity paresis, hormone replacement therapy and oral contraception, and factor V Leiden mutation or other acquired thrombophilia conditions.¹⁴ Patients who have DVT or who are taking medications that alter coagulation of the blood are also at risk of developing PE.^8,15 In addition, pregnant women are four to five times more likely to develop VTE, which is one of the leading causes of maternal death during childbirth.¹⁶ However, a proportion of patients who develop PE may have no risk factors.

Risk Stratification

The likelihood of PE can be estimated using various risk stratification approaches (see Table 1; Appendix 5). A patient may initially undergo assessment with a clinical prediction rule or clinical gestalt. Patients with high probability of PE may proceed directly to imaging, while patients with low probability may undergo further testing such as Pulmonary Embolism Rule-Out Criteria (PERC) or D-dimer testing to further assess the need for diagnostic imaging. This may be supplemented by additional biochemical or imaging studies to rule out differential diagnoses or strengthen estimates of PE risk.

Evidence supports the practice of determining the clinical pretest probability of PE before proceeding with diagnostic testing.¹⁷ The American College of Physicians (ACP) has provided best practice advice on the evaluation of patients with suspected acute PE, noting that the first step when evaluating a patient is to establish his or her pretest probability of PE.¹⁸ Clinical prediction rules (also called clinical decision rules) aim to determine risk profile and the necessity of undergoing diagnostic testing. The ACP recommends using either the Wells or Geneva clinical prediction rules.¹⁸

The Wells rule combines seven items based on both objective criteria from patient history or physical examination, and physician judgment, into a total score.¹⁹ Typically patients with scores lower than 4 are deemed low risk for PE, although there is variation in the cut-offs applied. The Geneva score differs from Wells in that additional diagnostic testing (electrocardiography, and/or chest radiography, and arterial blood gas) contributes to the score as well as consideration of risk factors and clinical presentation.¹⁹ A revised Geneva score has been developed that can be determined independently of the additional diagnostic tests.

There is controversy regarding which rules are the most accurate for predicting acute PE,²⁰ but the Wells and Geneva rules have received the most extensive validation in the widest range of settings.^20,21 There is some evidence from systematic reviews (SRs) to suggest that the Wells rule is more accurate than the Geneva rule, but that the most appropriate tool may depend on the setting (i.e., low prevalence versus high prevalence [referred population]).^20,22 Adherence to protocols incorporating the Wells score and D-dimer testing has been demonstrated to result in a 20% to 30% reduction in the number of computed tomography (CT) examinations performed.²³ D-dimer is one of several lab-based or imaging studies that are not used to diagnose PE, but may be used to increase confidence in the decision to forego testing or rule out differential diagnoses. A negative D-dimer test in a low-probability patient can support the decision to forego additional diagnostic investigation. In addition to D-dimer, other tests include lower-limb compression ultrasound, echocardiography (transthoracic or transesophageal), chest X-ray, capnography, and electrocardiography. These modalities are used for rule-out of PE or prognostic assessment of confirmed PE. PERC is an additional tool that can be applied in patients with low pretest probability following initial clinical assessment to help assess whether D-dimer testing can be deferred.²⁴ It is based on parameters that are available at initial emergency department assessment and uses an eight-factor decision rule. The clinician must answer “no” to all questions for a negative result, which can rule out PE and defers the need for further testing.

Given that symptoms of PE are not specific, clinical features alone cannot confidently rule PE in or out and they are rarely used in isolation.¹¹ There is also a risk of false-negatives, which can result in patients not receiving further diagnostic testing or necessary treatment. The positive predictive value of risk stratification strategies, particularly clinical prediction rules, may be influenced to some extent by the prevalence of disease in the population, as well as cut-off values used.²² Nevertheless, these scores may improve the efficiency of PE assessment and diagnostic yield of imaging studies,²¹ and decrease the volume of unnecessary imaging studies. Their use is in line with initiatives by Choosing Wisely and society partners, which recommend that clinicians avoid CT angiography in patients who are stratified at low risk of PE and receive either a negative PERC score or D-dimer measurement.²⁵

Diagnostic Imaging

Patients who are deemed at high risk of PE following pre-imaging risk stratification, or based on unstable presentation, usually undergo diagnostic testing for confirmation of disease positivity (see Table 1; Appendix 5). Conventional pulmonary angiography (PA) has been previously regarded as the gold standard, but due to the requirement for right heart catheterization and insufficient sensitivity, it has been overtaken by alternative modalities.^26,27 Other less-invasive methods of diagnosing PE include computed tomography pulmonary angiography (CTPA), magnetic resonance pulmonary angiography (MRPA), ventilation-perfusion (V/Q) scanning planar scintigraphy, V/Q single-photon emission computed tomography (SPECT), or V/Q SPECT-CT, positron emission tomography–CT (PET-CT) and thoracic ultrasound. Each of these imaging modalities has strengths and limitations, and the appropriate modality may depend on the available expertise of health care providers and technology, whether adherence to acquisition protocols are followed, and whether specific patient risk factors (e.g., allergy to contrast dye) and clinical conditions (e.g., pregnancy) are present.¹ Not all modalities are widely available or in routine clinical use in Canada and other developed countries. This may be due to lack of availability or expertise, or practical considerations such as increased time required and complexity of performing the exam.²⁸

CT overtook V/Q scintigraphy as the most frequently used imaging modality to diagnose PE in 2001.²⁹ Although CT is widely considered to be a more definitive test, a large multi-centre study reported that both CT and V/Q imaging used in conjunction with clinical probability assessment, D-dimer, and lower-limb ultrasound testing resulted in similar low rates of VTE events during three-month follow-up.³⁰ Because CT is associated with exposure to ionizing radiation and iodinated contrast agents (with the associated risk of malignancy and contrast allergy), there is concern about its overuse.³¹ A surge in CT use and improvements in technology have led to an observed escalation in the diagnosis of PE (including sub-segmental PEs of unclear clinical importance),^29,30 but there is no evidence linking its increased use with improved patient outcomes.^32–34 Major technical advances in CT technology have led to the use of CTPA combined with indirect CT venography, electrocardiogram (ECG)-gated CTPA, and dual source/dual energy CTPA.³⁵ However, in patients with known allergy to contrast media, those with severe renal failure, and pregnant women, alternative imaging modalities are often considered, especially in the emergency setting.⁴

Policy Issues

Of the total population of patients who are evaluated for suspected PE, few are confirmed to have the condition, indicating a low diagnostic yield of current evaluation methods.^22,36 Studies report a range of values for the diagnostic yield of CTPA, ranging from less than 5% to 30%, depending on the clinical characteristics of the patient pool, and use of risk stratification strategies.^37–40 False-positive test results, which, depending on pretest probability,⁴¹ can occur in approximately 10% to 42% of patients⁴² who undergo CT scanning, can lead to unnecessary anticoagulation therapy, which carries substantial risk of adverse effects including hemorrhage (occasionally devastating or fatal), interactions with other medications, inconvenience in terms of attendance for repeated blood tests (possibly requiring time off work), implications for future dental and medical procedures, and costs (both to the patient and society).⁴³ False-negative CT results, which also occur at high frequency (e.g., 1% to 11%),⁴⁴ can lead to bypass of necessary treatment, complications, and death. The uncertain benefit of increased testing and the significant expense of PE could suggest that current CT utilization patterns for the diagnosis of PE are not cost-effective.¹⁸ This is reflected in the increased diagnosis of mild PEs, which, if treated, may increase costs and possible harms, and may not reduce mortality. In light of these concerns, it is important to assess whether there are other cost-effective and safe alternatives.

The optimal diagnostic strategy for suspected PE among experts remains controversial,^46,47 and it can differ based on factors related to the health care setting (i.e., urban, rural, or remote) that may impact access to imaging. The optimal diagnostic strategy would, in theory, be one that has high diagnostic accuracy and clinical utility, at an acceptable cost. However, issues of access may also influence what is considered optimal for different populations. For instance, provision of timely diagnosis may be less feasible in rural and remote facilities due to lack of access to certain testing and imaging modalities and specialist expertise, as well as geographical barriers to care. Inability to access optimal diagnostic testing in a timely manner could increase the risk for missed diagnoses, as well as unnecessary anticoagulation due to either false-positives or long wait times to receive assessment.⁹ Patient safety concerns associated with exposure to radiation and contrast media that accompanies several imaging studies also disproportionately affect specific patient groups, including pregnant women, and young women for whom the risk of breast cancer associated with radiation is higher.⁹

Summary and Project Goals

Patients with suspected PE should be assessed using appropriate diagnostic tests in a timely manner.^2,45 Timing of access to diagnostic test results may have a significant impact on the management of the condition and the effective use of health care resources.⁹ The heterogeneous clinical presentation of PE and lack of specific symptoms can lead to myriad problems. These include the wide application of testing, which can be very costly and may result in over-diagnosis, false-positives, and unnecessary treatment. Although guidelines for PE diagnosis recommend the use of imaging tests,¹² the optimal diagnostic strategy for suspected PE remains uncertain,^46,47 and it may vary depending on the health care setting due to access to the technology. Thus, the goal of this health technology assessment (HTA) is to conduct an assessment of the evidence to inform formulation of recommendations regarding the optimal diagnostic strategy, including risk stratification, for acute PE in the current context of care, considering benefits, harms, and costs, as well as patient experiences, implementation issues, and environmental impacts.