• 
• 

Three Part Question

[In adult patients presenting to the ED with respiratory symptoms suggestive of Covid-19][what is the sensitivity and specificity of lung ultrasound (LUSS) by Emergency Practitioners at diagnosing Covid 19 pneumonitis][as confirmed by CT]

Clinical Scenario

A 45 year-old female presents to the ED with a 3-day history of cough and breathlessness. You wonder whether this patient has COVID-19. You worry that sending this lady into the hot zone of the department may result in a potentially vulnerable patient being exposed to COVID-19. You know that a CT scan is the gold standard imaging choice for COVID-19 pneumonitis, but its practicality, cost, and irradiation cause you to wonder whether LUS would be an appropriate alternative to facilitate safe triage of this patient.

Search Strategy

Embase <1974 to 2021 May 13> , Ovid MEDLINE(R) ALL <1946 to May 13, 2021>
Other papers found via citation chaining of review articles, word of mouth from colleagues, social media and searching point-of-care ultrasound websites (eg. www.ultrasoundgel.org) and ahead-of-print databases (i.e. MedRxiv) were also included.

Inclusion criteria:
Studies comparing LUS as an index test to CT as a reference standard.
Study outcome of diagnostic accuracy in terms of sensitivity/specificity
Study outcome with recorded numbers of true and false positives, and true and false negatives from which sensitivity and specificity can be calculated.

Exclusion criteria:
Diagnostic case control studies where CT was the basis of diagnosis
Case reports
Small case series
Not English language
Descriptions of LUS techniques
Descriptions of COVID-19 pneumonitis LUS findings
Unable to access paper

1 (Ultraso* or Lung ultrasound or Thoracic ultrasound or Chest ultrasound).af. (1246516)
2 (Coronavirus or Covid* or SARS-CoV-2 or novel coronavirus or nCoV*).af. (313555)
3 (diagnostic accuracy or sensitivity or specificity).af. (3892619)
4 1 and 2 and 3 (445)
5 remove duplicates from 4 (309)
6 limit 5 to human (279)
7 limit 6 to yr="2019 -Current" (238)
8 limit 7 to covid-19 (223)

Search Outcome

The most recent search was conducted 13/05/2021 The most recent searches yielded 223 search results. 210 were eliminated based on title. Abstract and full text review eliminated a further 6 results, leaving 7 papers for inclusion. Other search methods yielded 3 papers. This culminated in 10 papers for inclusion.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Quarato et al 03/21 Switzerland	82 consecutive patients with confirmed COVID-19 using a nasopharyngeal swab PCR referred to COVID-19 centre and admitted to the research institute.	Prospective study	Estimate sensitivity of admission LUS for the detection of SARS-CoV-2 lung involvement. Lungs were divided into zones, and direct comparison was made between zones on LUS and CT for specific findings.	Key Results Global LUS sensitivity found to be 52% when detecting COVID-19 pulmonary lesions. LUS sensitivity was 8% for detecting focal opacities (mild disease), 52% for crazy-paving pattern (moderate disease) and up to 100% in extensive subpleural consolidations (severe disease).	Patients already PCR confirmed COVID-19 cases. LUS specificity cannot be calculated as all patients had an abnormal CT scan.
Haak et al 18/11/2020 Netherlands	26 patients presenting to the ED with suspected COVID-19 that underwent LUS were included in the study.	Prospective cohort study	Diagnostic accuracy (Sensitivity, Specificity, negative predictive value and positive predictive value and diagnostic accuracy of LUS compared to a laboratory PCR testing and CT scan.	Excluding patients who received a PCR reference standard, we calculated LUS sensitivity, specificity, negative predictive and positive predictive values to be 60%, 62%, 87% and 28% respectively, compared to a CT reference standard.	Only 26 patients in the study underwent both LUS and CT from which we were able to calculate diagnostic accuracy as per the inclusion criteria of this BET.
Rizzetto et al 05/2021 Italy	219 patients with PCR +ve COVID and abnormal CT findings. LUS was performed on admission to the ED.	Retrospective study	Lungs were divided into zones and their individual severity score combined. A severity score > 0 was defined as a COVID-19 positive. Diagnostic accuracy of LUS was calculated for each lung zone against CT scores.	Lung ultrasound sensitivity and specificity using CT scan as a reference standard were 75% and 66% respectively.	Selection of only COVID +ve patients could lead to overestimation of diagnostic accuracy LUS performed by different experienced clinicians introduces issues of inter operator reliability.
Schmid et al 07/12/2020 Germany	31 patients presenting to the ED with respiratory symptoms. patients underwent LUS, CT, and PCR.	Retrospective cohort study	Diagnostic accuracy (Sensitivity, Specificity, negative predictive value and positive predictive value and diagnostic accuracy of LUS compared to a laboratory PCR testing and as a subgroup study, to CT scan, for which sensitivity and specificity were calculated. Positive and negative predictive values can be calculated from the data.	LUS sensitivity and specificity when compared to a CT scan was 65.0% and 72.7% respectively. From our calculations, positive and negative predictive values are 81.3% and 53.3% respectively	Only 31 patients of the 2236 patients in the study underwent both CT and LUS and were eligible for this BET. Funded by multiple external sources. Authors claim no conflict of interest.
Tung-Chen et al 13/07/2020 Spain	51 clinically suspicious PCR negative patients presenting to ED underwent CT and LUS	Prospective Study	Assess the diagnostic accuracy of LUS using a ROC curve analysis.	ROC curve analysis described LUS as having a sensitivity, specificity, positive and negative predictive values of 100%, 78.6%, 92.5% and 100% respectively vs CT.	No blinding of clinicians to reference standard. PCR positive patients were not given CT scans, suggesting patients with only mild disease were included.
Karagöz et al 12/2020 Turkey	72 patients presenting to ED with suspected COVID-19	Prospective, cross sectional cohort study.	To assess the diagnostic accuracy of LUS by using the BLUE protocol to assess the LUS and compare these findings to those on the CT.	LUS 96.9% (95%CI 84.2-99.9) sensitive 92.3% (95%CI 79.1-98.3%) specific The positive and negative predictive values were 84.3% (95%CI 64.5%-94.1%) and 98.6% (95%CI 91.1%-99.8%) respectively.	Wide confidence intervals No measure of interrater reliability
Walsh et al 28/09/2020 USA	49 patients over 14 years of age with no relevant underlying disease retrospectively selected on the basis of having an ultrasound scan performed and reported on before CT scanning was performed.	Retrospective Study	Reviewed records of eligible patients who underwent LUS for COVID-19 diagnosis. Outcome was the corroboration of treating clinician’s LUS findings with CT scan results.	49 records met inclusion criteria, 10% had COVID-19 changes on LUS. LUS found to be 92-100% sensitive and 37-80% specific at detecting COVID-19 related lung changes compared to CT.	Patients not randomly selected Very broad estimate of specificity of LUS.
Benchoufi et al 05/2020 France	107 patients presenting to 3 emergency departments with suspected or confirmed COVID-19 were given a CT scan, and LUSs by two physicians, blinded to each other and the CT.	Multicenter, observational non-randomised study	LUS conducted on 8 lung regions scored independently between 0-3, totalling a score of 0-24. This was compared to corresponding CT scans scored based on signs of interstitial pneumonia. Some patients were dual assessed by experienced and trainee LUS operators to ascertain differences between novice and experienced LUS operators.	86 of 107 patients had positive CT findings. Scores between LUS and CT showed a strong concordance. 17 patients excluded from this due to incomplete data. LUS showed 95% sensitivity and 83% specificity as compared to CT. 48 cases were dual assessed. The weighted kappa (0.62-8.1) showed a strong concordance with results between the trainee and experienced ultrasound operators.	Not peer reviewed, ahead of print. Lead author has declared competing interests as a cofounder and stakeholder in a company producing handheld ultrasound devices for clinical use. Study not randomised
Poggiali et al 03/2020 Italy	12 COVID positive patients presenting to the emergency department. Patients were given CT and LUS scans.	Observational study	Results of CT and LUS scans were assessed for correlation.	All LUS scans showed B-lines in accordance with COVID-19. CT scans showed ‘strong correlation’, with 5/12 patients showing crazy-paving patterning.	Small sample size (12). Limited quantitative analysis No statistical analysis or calculations of diagnostic accuracy.
Yang et al 25/05/2020 China	29 patients with laboratory confirmed COVID-19 underwent simultaneous CT and ultrasound scan. Ultrasound scans were divided into 12 regions, and mapped onto the corresponding CT scans. Results were interpreted in terms of regions.	Retrospective study	Direct comparison of 12 lung zones from each patient’s LUS findings to CT findings in lab confirmed COVID-19 patients.	Total of 540 lung regions were scanned by ultrasound. 63% of regions displayed signs of COVID-19 related changes. 95% of these regions showed B-Lines. CT scan detected changes in 38.7% of the total number of regions scanned. LUS is more sensitive at detecting COVID-19 lung changes than CT scan, in laboratory confirmed cases.	Reported in terms of regions makes it difficult to know the sensitivity and specificity on a patient by patient basis.

Comment(s)

The sensitivity and specificity of LUS at detecting COVID-19 pneumonitis ranged from 60-100% and 62-92% respectively across the included studies, with most papers reporting 80-95% sensitivity and 70-80% specificity. This demonstrates that LUS has potential to be a powerful tool at detecting lung changes associated with COVID-19 pneumonitis. Across the studies included here, there is no controlling of LUS techniques. Some studies broadly defined the features they were looking for across the whole lung, whilst a number of others implemented protocols, such as the BLUE protocol. Lung changes associated with COVID-19 on lung ultrasound are now well documented - pleural lines, B lines, subpleural consolidations and small pleural effusions. The varying effectiveness of the differing techniques, and the extent of lung changes required to form a ‘positive’ diagnosis creates a threshold effect that may be responsible for the large range of diagnostic accuracies shown in these studies. Benchoufi et al8 found a strong concordance with novice and experienced operators’ findings. This indicated that inter-operator reliability may not be a significant factor in COVID-19 LUS. Inter-operator reliability was not directly commented upon in most studies, though some did comment on the experience and number of operators. LUS is described as having a high diagnostic accuracy. It could therefore be used in emergency departments alongside other clinical information to form working diagnosis or exclusion of COVID-19 whilst awaiting definitive PCR diagnosis.

Clinical Bottom Line

In adult patients presenting to the ED with respiratory symptoms suggestive of Covid-19, the sensitivity and specificity of lung ultrasound (LUSS) by Emergency Practitioners at diagnosing Covid 19 pneumonitis as confirmed by CT ranged between 60-100% and 62-92% respectively across the included studies, with most papers reporting 80-95% sensitivity and 70-80% specificity.

References

1. Quarato CMI et al Low Sensitivity of Admission Lung US Compared to Chest CT for Diagnosis of Lung Involvement in a Cohort of 82 Patients with COVID-19 Pneumonia Medicina
2. Haak SL et al Diagnostic accuracy of point-of-care lung ultrasound in COVID-19 Emergency Medicine Journal 2020;38(2)
3. Rizzetto F et al Correlation between lung ultrasound and chest CT patterns with estimation of pulmonary burden in COVID-19 patients European Journal of Radiology 10/3/2021
4. Schmid B et al Lung ultrasound in the emergency department - a valuable tool in the management of patients presenting with respiratory symptoms during the SARS-CoV-2 pandemic BMC Emergency medicine 7/12/2021
5. Tung-Chen Y et al Correlation between Chest Computed Tomography and Lung Ultrasonography in Patients with Coronavirus Disease 2019 (COVID-19) Ultrasound in Medicine and Biology 11/2020
6. Karagöz A et al Accuracy of Bedside Lung Ultrasound as a Rapid Triage Tool for Suspected Covid-19 Cases Ultrasound Quarterly 12/2020
7. Walsh P et al. Point-of-care Lung Ultrasound Is Useful to Evaluate Emergency Department Patients for COVID-19 Western Journal of Emergency Medicine 28/09/2020
8. Benchoufi M et al Lung injury in patients with or suspected COVID-19 : a comparison between lung ultrasound and chest CT-scanner severity assessments, an observational study MedRxiv 04/05/2020
9. Poggiali E et al Can Lung US Help Critical Care Clinicians in the Early Diagnosis of Novel Coronavirus (COVID-19) Pneumonia? Radiology 13/03/2020
10. Yang Y et al Lung ultrasonography versus chest CT in COVID-19 pneumonia: a two-centered retrospective comparison study from China Intensive Care Medicine 25/05/2020