Managing Pulmonary Embolism

INTRODUCTION

Pulmonary embolism (PE) is commonly encountered in the Emergency Department (ED), with an estimated annual incidence of 39–115 per 100,000 individuals.¹ Classic risk factors include major trauma, lower extremity surgery, prior venous thromboembolism, recent hospitalization, oral contraception, postpartum period, malignancy, and thrombophilias.^2–4 However, numerous other risk factors exist.^5,6

Historically, mortality after PE was reported as 18–30%.⁷ However, more recent studies have found that PE-related mortality is estimated at 1–3%.^8,9 In the decades following the introduction of computed tomography pulmonary angiography (CTPA), an increasing incidence of PE has been accompanied by a lower case-fatality rate coupled with an increase in complications of anticoagulation for PE.^10,11 Together, these features suggest overdiagnosis of PE, warranting reconsideration of the evaluation and management of PE to minimize the harms from over-evaluation and over-treatment. There is a need to better understand the nuanced evaluation and treatment of PE. This paper does not intend to be a comprehensive review of all aspects pertaining to PE, but rather seeks to provide the key tenets of management based on the current literature and years of practice.

ASSESSMENT

The signs and symptoms of PE are non-specific and can include chest pain, dyspnea, cough, syncope, or hemoptysis. Consequently, the diagnostic approach for suspected PE typically involves pre-test probability assessment and potentially D-dimer testing and/or imaging (Figure 1).

Figure 1.

Possible diagnostic algorithm for non-pregnant patients with suspected pulmonary embolism

Elevated D-dimer levels are non-specific; however, the D-dimer assay plays an important role in risk stratification for suspected PE, as a normal D-dimer in non-high-risk patients can safely exclude PE.^12–14 D-dimer values rise with age, which has led to research on age-adjusted D-dimer levels for assays using fibrin equivalent units (FEU). In patients older than 50 years with a low or intermediate risk of PE, an age-adjusted cutoff of age x10 μg/L can be used.^12,15–17 Although research evaluating some assays reporting D-dimer units (DDU) demonstrates good performance of an age-adjusted dimer using age x 5 μg/L, it is not clear if this threshold can be applied to all DDU assays.^12,18,19

CTPA is the imaging modality of choice for PE in high-risk patients or those with clinical suspicion for PE and elevated D-dimer levels, as it is highly sensitive and specific for diagnosis.²⁰ However, radiologist agreement on the presence or absence of a PE varies. Up to 25% of CTPAs interpreted as diagnostic for PE, particularly at the subsegmental and segmental level, are false positives.^21–24 Ventilation/perfusion (V/Q) scans can be used in patients with a normal chest radiograph but is complicated by a substantial number of non-diagnostic tests and the inability to identify an alternative diagnosis.²⁵

Risk stratification

Risk stratification can occur by clinician gestalt or a risk stratification tool (RST). The most commonly used RST for PE in the US are the Wells’ Criteria and the Pulmonary Embolism Rule out Criteria (PERC) (Tables 1 and 2).^26–28 The Wells’ criteria may be perceived as complex because several scoring systems exist.²⁹ The proportion of patients with PE in each risk group depends on the study setting. These scoring systems perform similarly with regard to diagnosis of PE; however, the three-tier Wells’ Criteria may allow more individuals to avoid imaging as only patients with a Wells’ score > 6 proceed directly to CTPA rather than the threshold of 4.5 in the dichotomized Wells’ score.³⁰

Table 1.

Wells’ Criteria for Pulmonary Embolism

Consideration		No	Yes
Clinical signs and symptoms of DVT		0	+3
PE is #1 diagnosis OR equally likely		0	+3
Heart rate > 100		0	+1.5
Immobilization at least 3 days OR surgery in the previous 4 weeks		0	+1.5
Previous, objectively diagnosed PE or DVT		0	+1.5
Hemoptysis		0	+1
Malignancy w/ treatment within 6 months or palliative		0	+1
Score:
Three-Tier Model	Risk (Prevalence of PE in US studies)
0–1	Low risk (1–3%)79–81
2–6	Intermediate risk (8.5–15%)79–81
>6	High risk (37–43%)79,81
Two Tier Model	Risk (Prevalence of PE in US studies)
≤4	PE unlikely (1.8–7.2%)79
≥5	PE likely (28%)79

DVT, deep venous thrombosis; PE, pulmonary embolism

Table 2.

PERC Rule for Pulmonary Embolism

Consideration	No	Yes
Age ≥ 50	0	+ 1
Heart Rate ≥ 100	0	+ 1
Oxygen saturation on room air < 95%	0	+ 1
Unilateral leg swelling	0	+ 1
Hemoptysis	0	+ 1
Recent surgery or trauma (≤ 4 weeks ago requiring general anesthesia)	0	+ 1
Prior PE or DVT	0	+ 1
Hormone use (e.g., oral contraceptives, hormone replacement or estrogenic hormones use in males or female patients)	0	+ 1
Score: If any criteria are positive, PE is not excluded

DVT, deep venous thrombosis; PE, pulmonary embolism

Whether to use an RST or clinical gestalt is a topic of debate.³¹ One meta-analysis of prospective studies found that the sensitivity of clinician gestalt was comparable to RSTs; however, the specificity of RSTs was higher (81% vs 52%), suggesting that RSTs may have value in decreasing imaging.^15,32–34 A randomized trial of PERC reported 9.7% fewer imaging studies compared with usual care.³²

Low Probability

Low risk patients (estimated probability <15% or Wells’ Criteria <2) should be evaluated with the PERC criteria. Those who do not meet PERC criteria can forgo further evaluation for PE.^26,32 Low-risk patients with suspected PE with ≥1 PERC criteria should undergo D-dimer testing. Importantly, PERC should not be used in higher risk patients.^35,36

Intermediate Probability

Patients with an intermediate probability of PE (estimated probability of PE 15–50% or Wells’ Criteria 2–6) require D-dimer testing. A negative D-dimer in this population safely excludes PE.^12–14,20 More recently, the introduction of risk stratification algorithms using probability-adjusted D-dimer threshold have demonstrated promise in increasing PE evaluation efficiency.^37–39 Of these, the YEARS algorithm has the most robust evidence (Table 3).^15,38,40,41 YEARS allows a D-dimer threshold of up to 1.0 μg/mL in patients without hemoptysis, deep vein thrombosis (DVT), or PE as the most likely diagnosis.³⁸ This algorithm, combined with PERC, has been validated in multiple settings and can safely exclude PE and reduce imaging in non-high risk patients.^15,40,41 An observational study found that use of YEARS in all patients undergoing evaluation of PE would have a sensitivity of 92.6% (95% CI 87.0 – 96.0%) with an overall miss rate within accepted safety margins (0.59%; 95% CI 0.3–1.1%).⁴² Importantly, YEARS cannot be used in anticoagulated or critically ill patients.

Table 3.

YEARS Algorithm for Pulmonary Embolism

Consideration	No	Yes
Clinical signs of DVT	0	+1
Hemoptysis	0	+1
PE most likely diagnosis	0	+1
If no criteria present, use FEU D-dimer threshold 1,000 ng/mL
If ≥ 1 criterion present, use FEU D-dimer threshold 500 ng/mL
If patient pregnant and clinical signs of DVT present, obtain compression ultrasonography of the symptomatic leg; if normal, obtain D-dimer

DVT, deep venous thrombosis; PE, pulmonary embolism; FEU, fibrin-equivalent units

High Probability

Patients with a high probability of PE (e.g., Wells >6) should undergo imaging. The ability of CTPA to exclude PE in patients with a high likelihood of PE is controversial.⁴³ A meta-analysis of 22 studies found 0.56% (95% CI 0.39–0.72%) of patients with a negative CTPA were diagnosed with PE on follow up; however, this proportion was higher in studies in which ≥40% of patients were diagnosed with PE (1.3%; 95% CI 0.69–2.3%).⁴⁴

Pregnancy

Recent studies have demonstrated that pregnant patients can safely undergo evaluation for PE using risk-stratification tools, D-dimer, and ultrasound for DVT, which may prevent the need for further imaging.^33,45,46 In a prospective evaluation of 498 pregnant patients using the YEARS algorithm and DVT ultrasound (Figure 2), no patients who initially had PE excluded were diagnosed with PE in the following 3 months, and a significant proportion of patients avoided imaging (65% in the first trimester, 46% in the second trimester, and 32% in the third trimester).³³ This algorithm has been retrospectively validated in other pregnant cohorts and found to be safe and associated with reduced need for imaging.^46,47 As a result, professional society guidelines are beginning to recommend YEARS in pregnant patients.²⁰

Figure 2.

Possible diagnostic algorithm for pregnant patients with suspected pulmonary embolism

In pregnant patients who warrant imaging for PE, there is debate over the use of CTPA or V/Q imaging. The estimated fetal radiation exposure is minimal and safe for both V/Q scans and CTPA.⁴⁸ Although CTPA results in higher radiation exposure to breast tissue in the pregnant patient (3–10 milligrays), this dose has negligible implications for their lifetime cancer risk.⁴⁹ CTPA may be preferred by some physicians as it may reveal an alternative diagnosis.

MANAGEMENT

Management is influenced by the severity of the PE. Although PEs were formerly described as “massive” or “submassive,” these categories have been replaced by high, intermediate, and low risk categories that predict early mortality (Figure 3).²⁰ These categories integrate clinical, imaging, and laboratory parameters and can be used to guide treatment decisions. Although laboratory tests and/or echocardiography may be helpful in patients with clinical features suggesting right ventricular strain (e.g., electrocardiogram, brain natriuretic peptide, troponin), these evaluations are not necessary for all patients with PE.

Figure 3.

Pulmonary Embolism Risk Stratification for Mortality

For patients appropriate for oral anticoagulation, direct oral anticoagulants (DOACs) (e.g., apixaban, rivaroxaban) are recommended for treatment of most patients with PE over vitamin K antagonists due to efficacy, safety, and convenience. Patients with cancer-associated thrombosis can be treated with low-molecular-weight heparin (LMWH) or, in the absence of high risk of gastrointestinal or genitourinary bleeding, an oral Xa inhibitor (e.g., apixaban, rivaroxaban).^50,51 In patients who warrant parenteral anticoagulation, LMWH is generally recommended over unfractionated heparin (UFH) in most patients due to the unpredictable pharmacokinetics of UFH and increased risk of bleeding.^20,52–54 UFH is recommended in patients with serious renal impairment (creatinine clearance ≤30 mL/min).²⁰

High-Risk PE (“Massive”)

Patients with high-risk PE and those who deteriorate despite anticoagulation warrant treatment with thrombolysis. The American College of Chest Physicians and the European Society of Cardiology recommend that patients with high-risk PE and absence of high bleeding risk receive systemic thrombolysis (Figures 4, 5).^20,50 Although an increasing amount of observational data have examined surgical embolectomy and catheter-directed treatments (CDT; including catheter-directed thrombolysis and suction embolectomy), there are minimal comparative randomized trial data.^55–58 These treatments may be associated with less bleeding than systemic thrombolysis, but selection bias complicates direct comparison. Currently, guidelines recommend that embolectomy or CDT be reserved for patients with contraindications to systemic thrombolysis or deteriorate despite thrombolysis.^20,50

Figure 4.

Algorithm for the treatment of acute pulmonary embolism

Figure 5.

Possible Thrombolytic Doses for Acute Pulmonary Embolism

Intermediate Risk PE (“Submassive”)

Patients with intermediate risk PE (“submassive”) are hemodynamically stable but have signs of right ventricular dysfunction. Patients with intermediate-low risk PE should be started on oral anticoagulation or LMWH, depending on clinical factors. Patients with intermediate-high risk PE should receive anticoagulation with LMWH or heparin for the first 2–3 days, the time frame in which patients are most likely to decompensate.^20,59 Adjunct treatment (e.g., thrombolytics or CDT) of patients with intermediate-high risk or “submassive” PE is more controversial. Overlapping meta-analyses of randomized trials comparing thrombolytic therapy compared with anticoagulation alone have found an all-cause mortality benefit coupled with a wide range of major bleeding events.⁶⁰ As a result, current guidelines recommend thrombolytics only in intermediate risk PE patients who deteriorate despite systemic anticoagulation.^20,50 In contrast, CDT has been increasingly performed in patients with intermediate risk PE and may be associated with improvement in surrogate markers (e.g., right ventricle dilation, mean pulmonary artery pressure).^61–64

Low-risk PE

Patients with low-risk PE may be treated with LMWH or oral anticoagulation depending on clinical factors.

Subsegmental

Treatment of subsegmental PEs (SSPEs) is controversial, as several guidelines recommend close observation in low-risk patients with SSPE in the absence of DVT.^12,65 This is, in part, because the interrater reliability of radiologists in diagnosing SSPEs is slight-to-moderate, and false positive SSPEs are common (5–25% of CTPAs).^21,23,24,66 However, a prospective study found that among 266 patients with isolated SSPE, 3.1% had VTE recurrence by 90 days, half of which were PEs. Recurrence was more common in those with multiple SSPEs and in those over age 65.⁶⁷ Although these data suggest that anticoagulation may be indicated in patients with SSPE, the patient population and testing pattern in this international cohort included patients with a higher probability of PE than in United States cohorts.⁶⁸ We recommend anticoagulation in patients with DVT and SSPE; however, younger patients with SSPE and no evidence of DVT may not require anticoagulation based on patient risk factors.

DISPOSITION

Historically, patients with PE have been admitted to the hospital. However, two clinical trials and several observational studies have found outpatient management of select patients with acute PE to be safe and effective.^69–74 The introduction of DOACs has increased the ease of outpatient treatment. As a result, professional society guidelines recommend that patients with low-risk acute PE be managed as outpatients.^12,20,50,75 Several risk-stratification tools are available to identify patients who can be safely managed as outpatients, including the Pulmonary Embolism Severity Index (PESI), Simplified PESI (sPESI), and Hestia criteria.^76–80 The PESI and sPESI were developed to predict 30-day mortality using comorbidities and features of clinical presentation. Patients with a PESI of I/II or a sPESI of 0 are considered low risk, with a 90-day mortality of 0.8% (95% CI 0.4–1.8%).⁸¹

Both PESI and sPESI identify patients with low risk of mortality but do not encompass all patients who may need hospitalization. Unlike the PESI and sPESI, the Hestia criteria were specifically created to identify patients who can be safely treated as outpatients and includes comorbidities and psychosocial factors. As a result, we recommend using the Hestia criteria to identify patients eligible for outpatient management (Table 4).⁷⁶ Figure 4 demonstrates one possible treatment pathway recommended by a consensus group on criteria for outpatient treatment with international treatment guidelines.^20,76

Table 4.

Hestia Criteria for Outpatient Pulmonary Embolism Treatment

Consideration	No	Yes
Hemodynamically unstable – SBP < 100 mmHg and HR > 100, requires ICU care, or clinician judgment	0	+1
Thrombolysis or embolectomy needed – For reasons other than hemodynamic instability	0	+1
Active bleeding or high risk for bleeding – GI bleeding or surgery < 2 weeks ago, stroke < 1 month ago, bleeding disorder or platelet disorder < 75 × 109/L, uncontrolled HTN (SBP > 180 or DBP > 100), or clinician judgment	0	+1
>24 hrs on supplemental oxygen required to maintain SaO2 >90%	0	+1
PE diagnosed while on anticoagulation	0	+1
Severe pain needing IV pain medication required >24 hr	0	+1
Medical or social reason for admission >24 hr (infection, malignancy, no support system)	0	+1
Creatinine clearance <30 mL/min by Cockcroft-Gault	0	+1
Severe liver impairment by clinician judgment	0	+1
Pregnant	0	+1
Documented history of heparin-induced thrombocytopenia	0	+1
Score: If ≥ 1 present, patient is not eligible for outpatient management

SBP, systolic blood pressure; HR, heart rate; ICU, intensive care unit; HTN, hypertension; DBP, diastolic blood pressure; PE, pulmonary embolism; IV, intravenous