Authors: Jesse O. Wrenn (Department of Emergency Medicine, Vanderbilt University Medical Center, Nashville, Tennessee, USA; Division of Emergency Medicine, Tennessee Valley Healthcare System VA, Nashville, Tennessee, USA), Christopher Kabrhel (Center for Vascular Emergencies, Department of Emergency Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA)
Categories: In A Nutshell, emergency medicine, pulmonary embolism, venous thromboembolism
Source: British Journal of Haematology
Doi: 10.1111/bjh.19725
Authors: Jesse O. Wrenn, Christopher Kabrhel
Despite advances in clinical decision support, the diagnosis, prognostic risk stratification, treatment and disposition of emergency department patients with pulmonary embolism remain challenging. The use of diagnostic risk stratification tools and D‐dimer can decrease unnecessary exposure to radiation and intravenous contrast; however, D‐dimer is elevated in many conditions including normal pregnancy, so imaging is often indicated. Once diagnosed, prognostic risk stratification tools can inform treatment decisions across the risk spectrum, including identifying low‐risk patients with pulmonary embolism who can safely be treated at home. For patients requiring hospitalization, alternatives to unfractionated heparin can improve time to therapeutic anticoagulation and reduce treatment‐related bleeding risk.
The diagnosis of acute pulmonary embolism (PE) is challenging to even the most seasoned physician. Although dyspnoea, pleuritic chest pain, tachycardia and haemoptysis are hallmark features, PE also presents with non‐specific symptoms like fever, syncope or cough. While hormone use, recent surgery or trauma, malignancy and prior venous thromboembolism (VTE) are well‐known predisposing factors, recent infection, autoimmune or inflammatory bowel disease and recent admission for heart failure or myocardial infarction are also risks. Pregnancy increases risk, particularly in the third trimester until 6 weeks post‐partum, though evidence for increased risk in the first two trimesters is less convincing. Advanced age, obesity and smoking amplify these risks. ^1^ PE should be considered in any patient with common presenting symptoms and pervasive risk factors.
Diagnostic algorithms have been developed to risk stratify patients based on pretest probability (PTP) of PE. These are designed to identify patients with PE while avoiding unnecessary exposure to radiation and intravenous contrast associated with CT‐pulmonary angiography (CTPA). The PE rule‐out criteria (PERC) effectively rule out PE in patients with low PTP (<15%) with no formal testing. For patients who need testing, the Wells criteria, Geneva score or the gestalt of an experienced clinician can be used to stratify patients into low‐, medium‐ or high‐PTP categories. Most guidelines recommend D‐dimer testing for low‐ and intermediate‐risk patients (PTP <30%) ^2^ and CTPA for high‐PTP; however, there is emerging evidence that D‐dimer‐only diagnostic algorithms may also be safe in high‐PTP patients. ^3^ The cut‐off for D‐dimer positivity is typically 500 μg/L, though the specificity using this cut‐off is low (about 45%). Using an age‐adjusted D‐dimer cut‐off ^4^ or a PTP‐adjusted cut‐off, ^5^ such as the YEARS criteria, improves specificity and can safely rule out PE in approximately 14% more patients. For patients requiring imaging, most guidelines recommend CTPA. Ventilation‐perfusion (V/Q) scanning can be considered for patients who cannot undergo CTPA. However, access to V/Q scanning is limited and V/Q scans are more likely to yield indeterminate results compared to CTPA (42% vs. 7%). For patients who are too unstable for imaging or otherwise unable to undergo CTPA, echocardiography demonstrating right ventricular (RV) dysfunction can increase the clinical suspicion of PE. A positive venous ultrasound can also establish a diagnosis of VTE when pulmonary imaging cannot be performed.
PE is often suspected in pregnant patients due to the overlap in symptoms of PE and normal pregnancy and concerns about pregnancy hormones increasing risk. However, the prevalence of PE in pregnant patients who undergo testing is low (2%–7%), and their diagnostic evaluation is complicated. Societal guidelines differ on the appropriate use of risk stratification tools, D‐dimer testing and imaging in pregnant women. ^6^ Studies have demonstrated decreased sensitivity and specificity of D‐dimer in pregnant patients. ^7^ Most clinical risk prediction scores, including those that allow for D‐dimer adjustment, have not been studied in pregnancy. One exception is the pregnancy‐adapted YEARS algorithm, which can decrease CTPA use. ^8^ If imaging is indicated, the risk of radiation exposure to breast and fetal tissue must be weighed against the risk of haemodynamic compromise to the mother and fetus. Fortunately, radiation doses from V/Q scanning and CTPA are far below the threshold associated with teratogenicity or fetal complications. ^1^ However, starting with a non‐ionizing approach that combines D‐dimer testing and lower extremity venous ultrasound, accompanied by shared decision‐making with the pregnant patient regarding further testing, is also reasonable.
Some patients, such as those with haemodynamic instability, require inpatient treatment. Patients with RV dysfunction, which is associated with short‐term mortality, or high bleeding risk usually also require admission to monitor for haemodynamic decompensation. However, some low‐risk patients may be treated as outpatients. To assign the appropriate disposition, clinicians may use validated scores that categorize PE patients according to the risk of 30‐day death, like the PE Severity Index (PESI) and simplified PESI (sPESI), or according to their need for admission, like the Hestia criteria. Despite the availability of these tools, and contrary to most guidelines, two‐thirds of patients with low‐risk PE are still admitted. ^9^
Patients with PE should receive therapeutic anticoagulation as soon as possible, barring contraindications. Options for initial anticoagulation include unfractionated heparin (UH) (e.g. 80 U/kg as an initial bolus followed by 18 U/kg/hour continuous infusion), low molecular weight heparin (LMWH) (e.g. enoxaparin 1 mg/kg SQ BID) and direct oral anticoagulants (DOACs) (e.g. apixaban 10 mg PO BID or rivaroxaban 15 mg PO BID). There is growing evidence that UH is associated with delayed therapeutic anticoagulation. ^10^ LMWH and DOACs are more likely than UH to achieve therapeutic anticoagulation at 24 h, with simpler dosing. Despite these benefits, UH is still used regularly. ^11^
The benefit of anticoagulation, and the choice of anticoagulant, must be considered alongside the risk of bleeding. As a class, DOACs are associated with a lower risk of major bleeding and intracranial haemorrhage than heparin and warfarin. ^12^ While scores like HAS‐BLED ^13^ are useful in assessing bleeding risk associated with long‐term anticoagulation, there are no validated scores that assess short‐term bleeding risk and guide initial therapy. In general, inferior vena cava (IVC) filters do not remove the need for anticoagulation, but for patients with absolute contraindications to anticoagulation, IVC filter placement may be considered. IVC filters are thrombogenic over time and should be removed as soon as clinically possible.
Advanced therapy is often indicated for haemodynamically unstable patients with PE, and some with RV dysfunction. Critically ill patients with PE are increasingly being evaluated by multidisciplinary PE response teams (PERTs) ^14^ that facilitate rapid evaluation and treatment. Systemic thrombolysis is typically reserved for patients with haemodynamic instability that prevents alternative therapies. Catheter‐directed therapy (CDT), including thrombolysis or embolectomy, can be considered for high‐risk patients in lieu of systemic thrombolysis, or for patients for whom systemic thrombolysis is contraindicated. If available, patients with haemodynamic instability may be temporized with veno‐arterial extracorporeal membrane oxygenation (ECMO) prior to definitive treatment. However, evidence supporting CDT (Class 2a, Level C) and ECMO (Class 2b, Level C) is limited. ^1^
Jesse O. Wrenn was primarily responsible for writing the manuscript and designing the figure. Christopher Kabrhel was primarily responsible for conceptualizing and revising the manuscript.
Jesse O. Wrenn was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health under award number T32HL170986. Christopher Kabrhel was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health under award number R01HL168040.
Christopher Kabrhel has received grants (paid to his institution) from Grifols and Diagnostica Stago. He is a paid consultant for Siemens Healthineers and has served on advisory boards for BMS/Pfizer and Abbot. He holds stock in Insera Therapeutics. The remaining authors report no conflict of interest.