key: cord-0759284-n42zl295
authors: Gibbons, Ryan C.; Magee, Mark; Goett, Harry; Murrett, James; Genninger, Jessica; Mendez, Kendra; Tripod, Matthew; Tyner, Nicholas; Costantino, Thomas G.
title: Lung Ultrasound vs. Chest X-ray for the Radiographic Diagnosis of COVID-19 Pneumonia in a High Prevalence Population
date: 2021-02-04
journal: J Emerg Med
DOI: 10.1016/j.jemermed.2021.01.041
sha: 70f24c1904c3fcc10a6b53e005cc9dfc04d35eb7
doc_id: 759284
cord_uid: n42zl295

Background The viral illness severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), more commonly known as coronavirus 2019 (COVID-19), has become a global pandemic, infecting over 100 million individuals worldwide. Objectives The objective of this study was to compare the test characteristics of point-of-care lung ultrasound (LUS) with chest x-ray (CXR) at radiographically detecting COVID-19 pneumonia. Methods This was a single-center, prospective, observational study at an urban university hospital with >105,000 patient visits annually. Patients >18 years old, who presented to the emergency department with predefined signs and symptoms of COVID-19, were eligible for enrollment. Each patient received a LUS using a portable, handheld ultrasound followed by a single view, portable anteroposterior CXR. Patients with an abnormal LUS or CXR underwent a non-contrast-enhanced computed tomography (NCCT). The primary outcome was the radiographic diagnosis of COVID-19 pneumonia on NCCT. Results 110 patients underwent LUS, CXR, and NCCT. 99 LUS and 73 CXRs were interpreted as positive. 81 NCCT were interpreted as positive providing a prevalence of COVID-19 pneumonia of 75% (95% CI 66-83.2) in our study population. LUS sensitivity was 97.6% (95% CI 91.6-99.7) vs 69.9% (95% CI 58.8-79.5) for CXR. LUS specificity was 33.3% (95% CI 16.5-54) vs 44.4% (95% CI 25.5-64.7) for CXR. LUS positive predictive value (PPV) and negative predictive value (NPV) were 81.8% (95% CI 72.8-88.9) and 81.8% (95% CI 48.2-97.7) vs. 79.5% (95% CI 68.4-88) and 32.4% (95% CI 18-49.8) for CXR. Conclusion LUS was more sensitive than CXR at radiographically identifying COVID-19 pneumonia.

a sensitivity of 94% for viral pneumonia. 17-18 94 Similar to the H1N1 outbreak, one of the most significant challenges during this 95 pandemic is to diagnose and disposition patients accurately and efficiently, while ensuring the 96 safety of providers and conserving limited medical resources. Portable, handheld ultrasounds are 97 a practical option in triage scenarios, outpatient clinics, and resource-limited settings. 98

To date only anecdotal case reports, editorials, and two small retrospective, descriptive 99 studies have looked at LUS in patients with COVID-19 pneumonia. [19] [20] [21] [22] To our knowledge, no 100 prospective studies have assessed the utility of LUS at diagnosing COVID-19 pneumonia. 101

The primary objective of this study is to compare the test characteristics of LUS and 102

Patients with one or more of the predefined signs and symptoms of COVID-19 were 117 eligible for enrollment. Upon enrollment, a PGY1-3 emergency medicine resident or emergency 118 medicine attending, unblinded to the LUS indication, performed the LUS using the portable, 119 handheld Butterfly iQ (Guilford, CT) transducer in the lung setting. The transducer was 120 connected to a 5 th generation Apple iPad Mini (Cupertino, CA) with the Butterfly iQ application 121 pre-installed. Prior to study commencement, our residents and faculty reviewed a two minute 122 video reviewing the LUS findings of viral pneumonia. [17] [18] [19] [20] [21] [22] No additional training was provided to 123 our emergency medicine residents or attendings prior to their participation. All of our emergency 124 medicine attendings are credentialed in the core American College of Emergency Physicians 125 (ACEP) point-of-care-ultrasound applications, including lung ultrasound. 23 Each participating 126 resident and attending had performed >25 previous LUS per ACEP and Accreditation Council 127 for Graduate Medical Education (ACGME) guidelines for emergency medicine training. 23-24 128

Each physician followed a predetermined standard LUS protocol (Figure 1) recording 6 129 seconds clips for each view and performed the LUS wearing full personal protective equipment 130 the need to disinfect the scanner following each patient. At the discretion of the emergency 140 medicine attending, high-risk patients with normal LUS and CXR findings had a NCCT done as 141 well. Per departmental and hospital guidelines, one or more of the following criteria defined 142 potentially high risk patients: temperature >101°F (38.3°C), heart rate >110, respiratory rate >20, 143 hypoxia <92%, absolute lymphocyte count <1000/mm 3 (1.0-4.8 K/mm 3 ), and systolic blood 144 pressure <100mmHg. Immunocompromised patients were considered high risk as well. 145

Prior to study commencement, study investigators defined 4 LUS findings consistent with 147 viral/atypical pneumonia: irregular pleural line, B-lines, consolidation, and pleural effusion 148 We did not mandate transducer orientation for the LUS. Physicians positioned the probe two faculty review it. A third ultrasound faculty member was available if the initial two 163 disagreed regarding an interpretation. However, this was not necessary during our study. Study 164 investigators calculated a kappa coefficient to measure the interrater reliability between the non-165 ultrasound fellowship trained providers and the fellowship trained faculty. 166

Predefined CXR findings included opacity, infiltrate, interstitial edema or markings, and 167 atelectasis. The presence of one or more of these abnormalities was considered positive for 168 viral/atypical pneumonia. Unilobar and unilateral findings were considered abnormal. 169

Abnormalities did not have to be multilobar or bilateral to mandate a NCCT. A board-certified 170 radiologist, blinded to LUS findings but not to the CXR indication, provided an official CXR 171 interpretation. We used any mention of infiltrate, pneumonia, or atelectasis as being positive. 172 A board-certified radiologist, blinded to the LUS findings but not to the CXR findings or 173 to the NCCT indication, interpreted the NCCT. Due to the current COVID-19 pandemic, our 174 radiology and pulmonary/critical care departments developed hospital guidelines for NCCT 175 interpretation in the setting of viral pneumonia. The presence of ground glass opacities was 176 defined as Category (Cat) 1 and was consistent with viral/atypical pneumonia. Similar to CXR, 177 abnormalities did not have to be multilobar or bilateral to be defined at Cat 1. Category 2 was 178 indeterminate, and category 3 was consistent with "other diagnosis." Per hospital guidelines, all 179 patients with a Cat 1 or 2 NCCT were admitted, isolated, and treated as positive for COVID-19 180 pneumonia. Therefore, all Cat 1 and 2 scans were considered positive for the purpose of this 181 study. Cat 3 was negative for viral/atypical pneumonia. 182

CoV nasopharyngeal swab. 186

Using the electronic medical record, Epic (Verona, WI), study investigators performed 187 chart abstraction on all discharged patients 7 days after initial ED presentation in order to 188 identify any patients subsequently diagnosed or admitted for COVID-19 pneumonia. Epic allows 189 providers to query participating local health systems in order to share medical records. to define viral/atypical pneumonia in an attempt to capture all cases. We chose to include both 294

Cat 1 and 2 as well. As mentioned in our results, 86% of category 1 NCCT scans had a positive 295 COVID-19 swab as well as 61% of category 2 NCCT scans. None of the category 3 NCCT scans 296 had a positive COVID-19 swab. 297

Due to previous reports suggesting diffuse skip lesions of COVID-19 pneumonia, we 298 devised an 8 view LUS protocol. [19] [20] [21] [22] Our results suggest that this is a sufficient approach. Our 2 299 field of view, and we surmise that is why they were missed initially. We did not record the scan 301 time, but the ease of portable, handheld ultrasounds allowed the EP to move efficiently through 302 subsequent patients. We estimate the entire LUS took less than 5 minutes for each patient. 303

The specificities of both LUS and CXR were low. However, our study was designed to 304 augment sensitivity of both LUS and CXR. We defined any CXR finding of infiltrate or Unfortunately, we were unable to include 33 patients in the final data. 27 patients were 330 considered low risk and discharged without a NCCT at the discretion of the attending EP, 4 were 331 admitted for alternate diagnoses, and 2 left AMA. Although we can speculate some of these 332 patients had coronavirus, we were unable to perform outpatient testing due to limited availability 333 of tests. This limited our study to more clinically sick patients. It is unknown if this would 334 change our sensitivity, but it is likely that if these patients were included, the specificity of both 335 LUS and CXR would increase. 336 COVID-19 is a highly contagious infection transmitted via contact and airborne 337 droplets. 39 We chose to utilize a portable, handheld ultrasound device to perform LUS to 338 minimize staff exposure, patient movement, and the use of PPE. Moreover, the handheld device 339 is easily disinfected, operates on battery power, and requires less supporting infrastructure. Even 340 a portable CXR requires 1-2 staff members to perform and extensive disinfection. These 341 characteristics make portable, handheld ultrasounds ideal in diverse medical environments that 342 are resource-limited: whether in newly constructed tents in large urban areas or in remote 343 villages far from modern facilities.

Our protocol took less than 5 minutes to complete. Furthermore, a diverse group of 21 PGY1-3 346 emergency medicine residents and 22 attending EPs performed the LUS without additional 347

training. If larger studies corroborate these results, LUS may be a viable choice for diagnosing 348 COVID-19 pneumonia, especially in situations where CXR and NCCT are difficult to obtain. In 349 outpatient settings, temporary surge facilities, or resource-limited areas, a negative LUS could 350 obviate the need for further imaging. Although a positive LUS lacks specificity, combining this 351 imaging technology with a rapid COVID-19 lab test would help determine which patients might 352 require further imaging and treatment. Such LUS protocols are already being used, and our data 353 lends support to these efforts. 40 354

It is important to note that we do not advocate that all patients with suspected COVID-19 355 pneumonia have a NCCT done. This was a temporary hospital policy, during the initial surge in 356 the spring of 2020. The aim was to identify patients with radiographic evidence of viral/atypical 357 pneumonia, presumed to be COVID-19, in order to isolate them to particular sections of our 358 institution. Given the similar sensitivities of LUS and NCCT, POCUS may provide a means to 359 diagnose or at least rule out COVID-19 pneumonia given the 100% NPV in our study, when 360 utilized by fellowship trained providers. Furthermore, the association of more positive LUS 361 zones with NCCT cat 1 and 2 and less positive zones with Cat 3 may allow providers to 362 distinguish the sicker individuals from those less ill and to assess disease progression or 363 improvement without the need for NCCT. 364

This study suffers from the limitations of a single center study. Furthermore, the EPs 366 participating were not blinded to the LUS indication, which may have caused providers to 367 definition of positive findings for both LUS and CXR which increases the sensitivity of both at 369 the expense of specificity. 370

Another significant limitation was that only AP CXR was performed. This is a temporary 371

institutional policy during the current pandemic in order to limit patient movement and staff 372 Why is this topic important? COVID-19 has been a global pandemic. It is paramount to 511 have an accurate means to diagnosis patients in order to expedite treatment and quarantine. 512

What does this study attempt to show? The utility of point-of-care lung ultrasound for the 513 diagnosis of COVID-19 pneumonia. 514

What are the key findings? Point-of-care lung ultrasound is more sensitive than CXR at 515 diagnosing COVID-19 pneumonia. J o u r n a l P r e -p r o o f

Stability Issues of RT-PCR Testing of SARS-CoV-2 for Hospitalized Patients Clinically Diagnosed with COVID-19

Relevance of lung ultrasound in the diagnosis of acute 438 respiratory failure: the BLUE protocol

Diagnosing Acute Heart Failure in Patients With 440

Undifferentiated Dyspnea: A Lung and Cardiac Ultrasound

Early recognition of the 2009 pandemic influenza A 443 (H1N1) pneumonia by chest ultrasound

Prospective application of clinician-performed lung 445 ultrasonography during the 2009 H1N1 influenza A pandemic: distinguishing viral from 446 bacterial pneumonia

Can Lung US Help Critical Care Clinicians in the

Early Diagnosis of Novel Coronavirus (COVID-19) Pneumonia?

Point-of-Care Lung Ultrasound findings in novel 450 coronavirus disease-19 pnemoniae: a case report and potential applications during

Accreditation Council of Graduate Medical Education Emergency Medicine Defined Key 463 Index Procedure Minimums

American College of Emergency Physician Guideline on COVID-19: Ultrasound Machine 467 and Transducer Cleaning

Butterfly iQ Cleaning and Disinfection Guidelines

Accessed April 14, 2020. the diagnosis of adult community-acquired pneumonia: review of the literature and meta-483 analysis

Lung ultrasound for the diagnosis of pneumonia 485 in adults: a systematic review and meta-analysis

Lung ultrasound for the diagnosis of 487 pneumonia in children: a meta-analysis

Lung ultrasound in the diagnosis and follow-up of 489 community-acquired pneumonia: a prospective, multicenter, diagnostic accuracy study

Usefulness of lung ultrasound B-lines in connective 492 tissue disease-associated interstitial lung disease: a literature review

A simplified lung ultrasound for the diagnosis of 495 interstitial lung disease in connective tissue disease: a meta-analysis

Table 1 Patient Characteristics. NCCT+ Pneumonia (n=83) NCCT-Pneumonia (n=27)

Heart Rate, Median (IQR), Beats/min 98

Median (IQR), mm

Respiratory Rate, Median (IQR)