Diagnosis, Diagnostic Tools, and Risk Stratification for Contemporary Treatment of Pulmonary Embolism

Abstract

Pulmonary embolism (PE) is quite common and is associated with significant morbidity and mortality. It is estimated that it is the cause of approximately 100,000 annual deaths in the United States. With great variability in presenting symptoms of PE, poor recognition of PE can be fatal. As such, many risk scores have been created to identify the sickest patients. Choosing the appropriate imaging modality is also critical. Invasive pulmonary angiography was once the gold standard to establish the diagnosis. With the advent of nuclear imaging, V/Q scans, invasive angiography has been phased out for diagnosing acute PE. At present, the standard for diagnosis of acute PE is computed tomography pulmonary angiography. In select patient cohorts, nuclear studies remain the modality of choice. Once the diagnosis of acute PE is established, there is a broad spectrum of severity in outcome which has led to substantial focus and development of risk stratification prediction models. We will discuss making the proper diagnosis with contemporary diagnostic tools and risk stratifying patients with PE to receive the correct treatment.

The true incidence of pulmonary embolism (PE) in the United States is unknown. However, it is estimated that close to 100,000 deaths, annually, are due to PE. ^{1 2 3 4 5} Its clinical impact depends on the hemodynamic complications that depend on chronicity, size, and location of the obstruction. ^{6 7} Risk factors for PE include increased age, female gender, tobacco use, immobilization, recent surgery, autoimmune disorders, and malignancy. ^{1 2 3 4} The risk of developing a PE increases as the number of risk factors increases. While the impact of PE is still undetermined, case fatality appears to be improving. ^{5 8 9} This is due to earlier detection of PE by improved diagnostic tools and advances in management such as the emergence of mechanical thrombectomy as an alternative to thrombolytic treatment in the higher risk patients.

Diagnosis

Variability of symptoms can make PE difficult to diagnose. Common symptoms such as dyspnea, chest pain, or palpitations are shared with several cardiopulmonary conditions. Tachycardia, tachypnea, and evidence of right ventricular (RV) strain on physical exam suggest that PE should be part of the differential diagnosis. Hypotension and cool extremities may be present if there is hemodynamic compromise resulting from obstructive shock. Asymmetrical lower extremity edema, calf tenderness, and a positive Homan's test may occur if there is simultaneous deep vein thrombosis.

T-wave inversions, ST elevation, and evidence of a deep S-wave in lead I, as well as deep q-wave and T-wave inversion in lead III can be seen on an electrocardiogram. ¹⁴ A wedge opacity on a chest X-ray can be suggestive of pulmonary infarct from PE. An echocardiogram may show right heart dilatation with a septal shift as well the McConnell's sign where the RV free wall is akinetic except for at the apex.

Pretest probability plays a role in the selection of diagnostic tools to rule out a PE. Younger patients (age < 50), without evidence of tachycardia or hypoxia, with lack of leg swelling or hemoptysis and without a history of previous PE, trauma, or use of hormonal therapy can likely be ruled out for PE without imaging. ¹⁰ These criteria make up the pulmonary embolism rule-out criteria (PERC) score. Validated risk scores such as the Wells score or PERC have been shown to be more specific and as sensitive as clinical impression for PE. ¹¹ Wells score can only be used if a patient has had symptoms for < 30 days. It is not validated for use in pregnant women, or in patients who have been on anticoagulation. A Wells score ≤ 4 makes PE unlikely but does not fully exclude it. In such cases the PERC score can be used to safely exclude a PE if it is zero. However, the PERC score has not been validated in patients with cancer, morbid obesity, or long-standing hypoxemia.

In cases of low or intermediate probability, as determined by risk scores such as the Wells or PERC scores, a D-dimer test is useful. ^{12 13 14 15 16} Normal D-dimer levels have a high negative predictive value (NPV), but a low specificity therefore a negative result would rule out PE. Patients with intermediate or high pretest probability or those with an elevated D-dimer should have further imaging for evidence of PE.

Imaging

There are multiple imaging modalities that assist in the diagnosis of PE. Chest radiographs in the setting of PE are typically unremarkable. However, there are associated signs and eponyms that are occasionally observed. These include the Westermark sign and oligemia distal to large pulmonary artery occlusion. ¹⁷ The Fleischner sign, or a prominent central artery resulting from pulmonary hypertension or PE may also be observed from enlargement of pulmonary artery. ¹⁷ The Palla sign refers to the enlargement of the right descending pulmonary artery, causing a “sausage” appearance toward the right middle lobe. ¹⁸ A Hampton hump refers to a dome-shaped, pleural-based opacification in the lung from a PE or pulmonary infarct. When it occurs secondary to a PE, the opacification typically is wedge shaped. ¹⁹ Historically, invasive pulmonary angiography was the main modality for diagnosing PE and was considered the reference standard. ^{20 21} However, its sensitivity is not 100%. ²² It also has a rather poor sensitivity (45–66%) for identifying subsegmental PE. ^{23 24} A major limitation to angiography is its inherent invasiveness requiring right heart catheterization and contrast administration. With the advent of nuclear ventilation and perfusion imaging, invasive angiography has been phased out as the diagnostic imaging modality of choice.

V/Q scans use select radiotracers, instead of a contrast medium, to identify areas of ventilation and perfusion mismatch. This leads to lower radiation exposure for patients without the use of nephrotoxic agents. As such, in patients with renal failure, young females, or those with contrast allergies, V/Q scans are preferred. ²⁵ However, they are limited by their availability and their results are frequently inconclusive in up to 50% of exams. ²⁶ The use of multiple-head cameras, V/Q single-photon emission computed tomography (SPECT) is a newer method that provides better sensitivity and specificity with lower inconclusive test rates (< 3%). ²⁷ SPECT images allows for three-dimensional imaging, instead of the two-dimensional (2D) imaging using planar V/Q. As such, there is improved images of the lungs that allows for less segmental overload and shine-through of adjacent lung allowing for better definition of size and shape of perfusion defect in individual defects. ²⁸ SPECT imaging has advantages over planar imaging as there is less interobserver variation and better delineation of mismatched defects. There is also a higher sensitivity and specificity (97–100% vs. 76–85% and 91–96% vs. 78–85%, respectively). ^{29 30} SPECT imaging also has significantly lower studies that are indeterminate, usually less than 5%. There are few studies that compared V/Q SPECT with CT pulmonary angiogram (CTPA). Although limited, there is some evidence that SPECT is more sensitive, but less specific in the detection of PE and uses significantly less radiation doses. ^{30 31 32 33 34} However, there are no prospective management outcomes studies using V/Q SPECT.

CTPA is currently the gold standard imaging modality. ²⁶ This imaging modality allows for visualization of subsegmental pulmonary arteries. ³⁵ According to the PIOPED II trial, the sensitivity and specificity of CTPA for PE is 83 and 96%, respectively. ³⁶ The NPV was found to be largely dependent on the pretest probability of PE. Low and intermediate pretest probability of PE had a NPV of 96 and 89%, respectively, whereas patients with high pretest probability had NPV of 60%. However, there was a high positive predictive value in those with intermediate and high pretest probability, 92 and 96%, respectively.

Radiation exposure between the different imaging modalities varies. ²⁶ Pulmonary angiography has an effective radiation dose of 10 to 20 millisieverts (mSv), CTPA 3 to 10 mSv, whereas both V/Q imaging (planar and SPECT) modalities use approximately 2 mSv. Exposure during angiography can be decreased with the use of biplane imaging. Magnetic resonance angiography is an alternative that has shown promise, mainly in that it limits radiation exposure. However, its limited availability, relatively low sensitivity (84.5%), and 30% inconclusive results, has slowed broad implantation of its use. ³⁷

The use of transthoracic echocardiography (TTE) is helpful for risk stratification for PE, but often does not aid in the diagnosis of PE. A normal TTE does not exclude a PE and a TTE demonstrating signs of RV pressure or volume overload may occur in the absence of PE. ^{38 39} RV signs of pressure and volume overload on TTE can be seen with enlargement of the RV in parasternal long view on 2D imaging, an apical 4-chamber view with a basal RV/left ventricular (LV) ratio > 1.0. In the short axis view, intraventricular septal flattening can be seen. There frequently is inferior vena cava plethora with minimal respiratory collapsibility. The “60/60” and “McConnell's” signs are not sensitive but are highly specific for PE. The 60/60 sign is demonstrated by the coexistence of the acceleration time of pulmonary ejection (< 60 ms) and midsystolic “notch” with mildly elevated (< 60 mm Hg) peak systolic gradient at the tricuspid valve. The McConnell sign is defined by RV free wall akinesis with sparing of the apex. RV dysfunction may also be estimated with a decrease in tricuspid annular plane systolic excursion (TAPSE) of < 16 mm and a decreased peak systolic (S') velocity of the tricuspid annulus (< 9.5 cm/s) on tissue Doppler imaging. Patients with more of these findings are at a higher risk of morbidity and mortality, thereby it is important to stratify patients so that the highest risk patients are treated with advanced therapies.

Risk Stratification

Current international guidelines suggest early discharge and treatment at home in patients at low risk of PE if proper outpatient care and anticoagulation treatment can be provided. By using risk stratification with both triaging strategies, more than a third of patients with PE could be managed at home with low rates of adverse events. Risk stratification also plays a role in the initial management of patients selecting out those who will need closer monitoring and possible rescue reperfusion either with thrombolytics or catheter-based thrombectomy.

The Pulmonary Embolism Severity Index (PESI) is a well-validated risk score that serves a prognostic model for PE that is helpful for guiding therapy. ⁴⁰ Furthermore, the simplified version, sPESI, also serves an excellent tool that has a high NPV for ruling out severe adverse outcomes. ⁴¹ The components of the PESI risk score include age, sex, history of cancer, heart failure, chronic lung disease, heart rate > 110 beats per minute, systolic blood pressure < 100 mm Hg, respiratory rate > 30 breaths per minute, temperature < 36°C, altered mental status, and oxygen saturation (< 90%). Altered mental status, hypotension, and history of cancer are more heavily weighted compared with the other components.

The latest European Society of Cardiology (ESC) guidelines include both the PESI and sPESI scores in the recommendations for risk stratification of PE. ²⁶ Per the ESC guidelines, a PESI class III to IV or sPESI ≥ I should prompt analysis of troponin level. If the troponin level is elevated in the setting of echocardiographic evidence of RV dysfunction, then the patient has an intermediate to high risk. Therefore, the patient must be hospitalized and may require hemodynamic support as well as rescue reperfusion such as with mechanical thrombectomy.

The Hestia rule is another simple score that can be used to risk stratify patients with PE. It was constructed to provide a simple set of readily available clinical criteria to triage patients who could be discharged home for outpatient treatment. ⁴² The Hestia rule is considered low-risk if none of the following apply: hemodynamic instability, thrombolysis or embolectomy, active bleeding or high-risk of bleeding, supplemental oxygen, PE diagnosed during anticoagulation treatment, severe pain requiring intravenous medication, medical or social reason for hospitalization, severe renal or hepatic impairment, or pregnancy. Per ESC guidelines, if any of the components of the Hestia rule are positive than the patient should be admitted to the hospital and a cardiac enzyme level should be checked. ²⁶

The HOME-PE trial, a randomized open label study, showed that risk stratification can identify approximately one-third of PE patients who can be safely treated at home. ⁴³ It compared the sPESI score and the HESTIA rule, showing that among patients with PE, risk stratification with the HESTIA rule was noninferior to the sPESI score on all-cause death, recurrent venous thromboembolism (VTE), or major bleeding. The two strategies were similar regarding the proportion of patients treated at home. By using risk stratification, approximately one-third of low-risk patients with PE could be safely managed at home. No scores are necessary to risk stratify patients with evidence of RV dysfunction on echo or CTPA. They should proceed with the evaluation of cardiac markers—if elevated, would mean that the patient is intermediate to high risk.

RV dysfunction and failure is a key determinant of adverse clinical outcomes. Massive PE is defined by the presence of PE resulting in sustained hypotension (not due to other causes of hypotension), obstructive shock requiring inotropes and vasopressors, and pulselessness. In these clinical scenarios, emergent reperfusion treatment is needed to prohibit further hemodynamic deterioration. ⁴⁴ In the large international prospective registry—Registro Informatizado de la Enfermedad TromboEmbolica venosa (RIETE)—of the 15,520 consecutive subjects who presented with VTE, 248 (1.6%) presented with symptomatic massive PE. ⁴⁵

As such, it is much more common to present with hemodynamic stability and nonmassive PE. This cohort can be divided into low-risk PE and submassive (intermediate risk) PE. This group requires further risk stratification to predict severity. Low-risk PE is defined by the presence of PE without hypotension, shock, RV dysfunction, or myocardial ischemia (elevated troponin). These patients typically do not require inpatient hospitalization and have a < 1% early mortality. Submassive PE is defined by a PE with RV dysfunction or myocardial ischemia without hypotension.

Besides clinical scores, multimodality imaging plays a major role in risk stratification. On TTE, normotensive patients with acute PE with a TAPSE ≤ 15 mm and a RV/LV ratio > 1.0 are at increased risk for early mortality. ⁴⁶ Those with TAPSE ≥ 20 mm had improved mortality outcomes and can be characterized as a low-risk cohort.

CT with contrast is the main imaging tool that helps stratify patients in a single examination-based approach. As the initial test to detect PE, it reveals the burden of PE and how central the thrombi are located. This has implications for possible mechanical or aspiration thrombectomy in high-risk patients.

Furthermore, features of the thrombus on CT can help to understand the acuity of the PE. Eccentric, calcified thrombi with bands or webs associated with enlargement of main pulmonary arteries, atherosclerotic calcification, tortuous vessels, or RV hypertrophy are likely to be chronic and associated with a more insidious presentation of symptoms. Patients with chronic PE would not benefit from rescue reperfusion but can be treated medically and with pulmonary thromboendarterectomy once stable. ^{47 48}

Risk stratification scores based on CT findings have also been developed. One such score is the CT obstruction index (CTOI). ⁴⁹ The CTOI is a thrombus burden score providing information regarding blood flow in the parenchyma distal to the clot. Results in its use have been controversial, especially in predicting risk of mortality. This is likely due to the fact that most studies that used CTOI were small. ^{50 51 52 53}

The development of multiple scores speaks to the importance of risk stratification in the management of PE. At the cornerstone of all the scores is clinical assessment and imaging. The scores provide a way to summarize the data into a concise system to identify the sickest patients that will need to be hospitalized and to be treated aggressively to avoid poor outcomes. The development of the PE response team (PERT) has helped to support rapid clinical decision-making in the setting of acute PE. The PERT team draws on experiences across multiple disciplines and is able to synthesize the clinical and anatomical data to allow for a timely intervention. ^{54 55}

Conclusion

Early diagnosis of PE is key to successful management as it is in stroke and myocardial infarction. Similar to the latter conditions, the spectrum of the presentation of PE dictates the need for specific criteria to risk stratify patients. The multiple scores that have been validated along with multimodality imaging are the cornerstones of diagnosis and risk stratification. As the diagnostic and prognostic tools evolve, so will the tools for the management of patients with various risk profiles.