Bedside Transthoracic Sonography in Suspected Pulmonary Embolism: A New Tool for Emergency Physicians

Pulmonary embolism (PE) has an incidence rate up to 69 per 100,000 persons per year¹ and a mortality rate estimated at 9 per 100,000 per year.² Because signs and symptoms are nonspecific, the diagnosis of PE may be difficult and requires a high index of suspicion. Recent technical advances, including D-dimer and multidetector-row computed tomography (MDCT) scan, have had a significant effect on PE diagnosis, increasing detection rates especially for segmental and subsegmental PE, but not without a decrease in specificity.^3–7 However, MDCT is still imperfect in both sensitivity and specificity and is neither universally available nor feasible for all patients requiring a work-up for PE.^4–6,8 MDCT may also increase cancer risk through exposure to radiation,^9–11 may not be suitable for unstable patients, and cannot be done as rapidly as bedside ultrasound (US).

Although bedside cardiac US has been reported in the emergency department (ED) diagnosis of critically ill patients with PE by imaging of the right ventricle,¹² we are not aware of any report of the ED use of bedside pulmonary US for the diagnosis of PE. Lung US is an emerging area of interest for emergency and critical care physicians. It can be performed at the bedside and has been used successfully to diagnose pneumothorax, pleural effusion, pneumonia, and lung edema, as well as PE.^13–17 Surprisingly, however, little focus has been on emergency US evaluation for PE.^18–21

Here we report three cases in which the diagnosis of PE was provisionally made by chest/pulmonary US and confirmed by MDCT. In all three patients, US for deep vein thrombosis (DVT) of the lower extremities was negative, and no acute right heart strain was found on echocardiogram. The literature evaluating classic sonomorphology for PE and its accuracy will be discussed.

Case 1

A 21-year-old woman presented to the ED without significant past medical history and with several days of severe and worsening chest pain that ultimately woke her from sleep. The symptoms had begun about 24 hours after an elective surgical abortion at 8 and 2/7 weeks’ gestational age. She was a smoker, was not recently on birth control, and had no other known risk factors for PE. The pain was inferior to the right breast. She also reported palpitations, episodes of shortness of breath, and cough with a small amount of hemoptysis on the day of presentation. Her physical exam showed a blood pressure of 127/81 mm Hg, heart rate of 109 beats/min, respiratory rate of 18 breaths/min, oxygen saturation of 100% on room air, and a temperature of 38.2°C. When speaking, her respiratory rate increased to the mid-30s and her pain worsened. On auscultation she had normal heart tones and coarse ronchi over her right lung, with decreased breath sounds over her right lung base. Her chest pain was reproducible with chest wall palpation at her right lower chest.

A focused bedside US was performed in the ED using a curvilinear transducer (3–5 MHz) and a high-frequency probe (7.5–10 MHz, Sonosite M-Turbo, Bothell, WA). The left lung appeared normal (Figure 1A), but the right lower lung showed multiple wedge-shaped hypodense parenchyma defects located at the lung periphery with localized and basal effusion (Figure 1B, Figure 2, and Video Clip S1). No perfusion of the edematous lung tissue was detected using Doppler US. From these observations, PE with lung infarction was diagnosed. Subsequent chest x-ray (CXR) showed right lower lobe consolidation with effusion (Figure 3). Because of the concerning US findings, emergent contrast-enhanced MDCT of the chest (Siemens Somatom Sensation 16, Siemens Medical Systems, Erlangen, Germany) was performed and confirmed the diagnosis of right segmental PE with lung infarction and pleural effusion (Figure 4). The patient was admitted to the hospital for further medical work-up and remained stable. A work-up for lower-extremity DVT showed negative results.

Figure 1.  (A) TUS showing normal left lung. The transducer is placed at the left mid-axillary line above the diaphragm parallel to the fourth intercostal space. The pleura is identified as a bright hyperechoic line. (B) The transducer is placed in the right coronal plane at the level of the diaphragm. The lung parenchyma is seen with a pleural-based, wedge-shaped, parenchymal hypodense area (white arrowheads) marking the visible area of lung infarction. There is strong posterior enhancement and parenchymal consolidation (white hyperdense area central to the peripheral defect). *Pleural effusion between liver (L) and right lung. CW = chest wall; TUS = transthoracic ultrasound.

Figure 2.  TUS showing a larger area of two adjacent wedge-shaped defects. The transducer is placed slightly more posterior than shown in Figure 1B. *Pleural effusion. CW = chest wall; L = liver; TUS = transthoracic ultrasound.

Figure 3.  CXR showing right lower lung consolidation and effusion (*). CXR = chest x-ray.

Figure 4.  Two CT images of the mid-chest (A) and lower chest (B) showing lung infarction (arrows) and pleural effusion (*).

Case 2

A 65-year-old woman with past medical history of hypertension and stroke presented several hours after a sudden onset of sharp chest pain that worsened with deep inspiration. No additional risk factors for PE were known. Her physical exam revealed mostly normal vital signs, including a blood pressure of 168/69 mm Hg, heart rate of 69 beats/min, respiratory rate of 18 breaths/min, and oxygen saturation of 99% on 2 L of oxygen by nasal cannula. Her lungs were clear and heart auscultation showed a regular rate and rhythm with no murmurs or rubs. A CXR was obtained and was unremarkable (Figure 5). On sonographic evaluation of her lungs using curvilinear (3–5 MHz) and linear high-frequency (7.5- to 10-MHz) probes, several subtle wedge- or rounded-shaped hypodense parenchymal areas were detected adjacent to the pleural space, raising the concern for PE (Figure 6A). Large areas of the lung appeared normal on transthoracic US (Figure 6B). Given her sonographic findings, a contrast-enhanced MDCT with PE protocol was obtained. The MDCT revealed bilateral pulmonary emboli (Figure 7). The patient was admitted to the medical service in stable condition for further management. A vascular US for DVT was negative, and the patient was discharged home after an uneventful hospital stay without clear etiology of her PE.

Figure 5.  CXR showing no signs of infiltrate, wedge-shaped infarct, or effusions. CXR = chest x-ray.

Figure 6.  (A) A subtle, pleural-based, hypodense parenchymal defect (white arrowheads) representing an area affected by PE is shown on TUS. The transducer is placed at the right posterior axillary line in a coronal plane. (B) TUS of an area of normal lung. CW = chest wall; PE = pulmonary embolism; TUS = transthoracic ultrasound.

Figure 7.  CT images depicting wedge-shaped changes of the peripheral lung parenchyma (A, black arrows), caused by bilateral pulmonary emboli seen on CT angiogram (B, white arrows).

Case 3

A 52-year-old male patient without significant past medical history presented with chest pain and shortness of breath while walking to work. He felt diaphoretic and febrile, and upon arriving at his work place, a hospital, he came to the ED to be “checked out.” He had no known risk factors for PE. At triage he presented with a blood pressure of 153/87 mm Hg, but was both tachycardic (132 beats/min) and mildly tachypneic (20 breaths per minute). Initial room air oxygen saturation was 83%; this improved to 96% on 3 L of oxygen by nasal cannula. On auscultation he had a regular heart rate without murmurs, coarse breath sounds bilaterally, and chest pain with deep inspiration. Focused bedside sonography using a curvilinear 3- to 5-MHz probe and a linear 7.5- to 10-MHz transducer revealed no pneumothorax, but multiple small wedge-shaped parenchymal defects were noted at the lung periphery with localized pleural effusion, highly suspicious for multiple PE (Figure 8, Figure 9, Video Clip S2). The patient was started on an intravenous heparin infusion. A bedside US in the ED for DVT of the lower extremities was negative. Immediate CXR was interpreted as revealing no significant infiltrates, minimal cardiomegaly, and hypoinflation (Figure 10). A subsequent contrast-enhanced MDCT of the chest with PE protocol confirmed the diagnosis of PE with multiple segmental and subsegmental bilateral emboli (Figure 11). The patient was admitted to the hospital in stable condition and all further medical diagnostics, including venous duplex US of the lower extremities and sepsis work-up, were negative. He was discharged home on hospital day 7 in stable condition.

Figure 8.  TUS depicting a wedge-shaped parenchymal defect (white arrowheads) and localized pleural effusion. CW = chest wall; TUS = transthoracic ultrasound.

Figure 9.  Small parenchymal defect (white arrow) with irregular or abolished pleural line and posterior enhancement and lung consolidation.

Figure 10.  CXR with mild cardiomegaly. No significant infiltrates, atelectasis, or effusions were seen. CXR = chest x-ray.

Figure 11.  CT scan showing parenchymal defects caused by PE (A, black arrows) corresponding with US findings. (B) Multiple emboli on CT angiogram (white arrows). PE = pulmonary embolism; US = ultrasound.

Discussion

Sonographic findings of PE on transthoracic US (TUS) are well described^{13–16,22–29} and correlate with the pathophysiologic results found on autopsy in patients diagnosed with PE and on helical CT.^28,30,31 Sonomorphology can be divided into parenchymal and pleural pathology. Typical parenchymal pathology includes multiple hypoechoic areas of lung tissue extending to the pleura. These lesions commonly present as wedge-shaped (85%), round, or polygonal; are hypoechoic parenchymal areas; and are pleural based. Their size can vary significantly and can range from 5 mm up to several centimeters. Occasionally, larger lesions have a central enhancement.^22,24,25 A posterior enhancement artifact with hyperechoic pulmonary consolidation can be seen central to the pleural-based parenchymal lesion. If acute and complete pulmonary infarction is present, pulmonary vascular flow might not be detectable using color Doppler US within the affected lung area.^{13,14,28–30} Pleural findings consist of localized or basal pleural effusions and can be found in up to two-thirds of patients with PE.^22–27 They correspond to the site of the parenchymal lesion.^14,22–29 Interestingly, several investigators have found that the majority of detected PEs are found in the lower lung lobes, also noting that the localization of the parenchymal and pleural defects correlate with the area of pleuritic chest pain.^22,24,25

Given the rising concern over radiation exposure and projected increased cancer risk,^8,10,11 it is not surprising that investigators have sought to determine if TUS can be used as an alternative or complementary imaging modality for PE. Mathis et al.²² substantiated results from several small European studies in a prospective multicenter trial of 352 adult patients. He reported an overall sensitivity of TUS of 74%, a specificity of 95%, positive and negative predictive values of 95 and 75%, respectively, and accuracy of 84% when compared to helical CT.²² The specificity of TUS increased to 99% if the criterion for sonographic diagnosis of PE was two or more subpleural parenchymal lesions. The overall number of lesions found with sonography was higher than by CT, but TUS measured them significantly smaller.²² Results from studies for TUS for PE in children were recently reported by Kosiak et al.²⁶ and also showed promising results. All investigators concluded that TUS is a highly specific complementary or alternate tool and should be considered in the emergency setting for patients unable to receive a CT or in the pediatric setting.

The efficiency of TUS when compared with newer generation 64 or higher MDCT still needs to be determined, but might be somewhat less than for helical CT.²⁷ This would not be surprising, as MDCT technology is known to detect higher rates of seventh-generation subsegmental PEs^3–5 and could also find indirect signs of PE more frequently, such as wedge-shaped subpleural lesions, which only have been reported by helical CT in 25% to 62% of PE patients.^31,32

Echocardiography is another indirect but more thoroughly studied application of US for evaluating PE. Most patients with PE have a normal cardiac US, making this modality unsuitable for routine detection and diagnosis of PE.^32,33 However, if signs of acute right ventricular strain and dysfunction are present on transthoracic echo, it can be an independent predictor of death.³²

Incorporating bedside US of the thorax and heart into risk-stratifying algorithms for patients with suspected PE could further improve sensitivity and specificity of these clinical decision rules, especially when applying the criteria established by Mathis et al.²² This could be especially useful for patients in the intermediate risk group or when D-dimer and chest CT are not immediately available or feasible. Large, multicenter studies will be necessary to compare the effectiveness of these imaging modalities for both adult and pediatric populations and clinical decision algorithms incorporating TUS. In the emergency setting, however, and especially in patients presenting with contraindications for CT radiation exposure or contrast application, TUS for PE diagnosis should be considered.