Diagnostic accuracy of point-of-care lung ultrasound in COVID-19.

ABSTRACT

BACKGROUND: A promising modality for diagnosing pulmonary manifestations of COVID-19 in the emergency department (ED) is point-of-care ultrasound (POCUS) of the lungs. The currently used PCR as well as chest X-ray and CT scanning have important disadvantages. The aim of this study is to evaluate the diagnostic accuracy of POCUS in patients with suspected pulmonary manifestations of COVID-19 in the ED. METHODS: This prospective diagnostic accuracy study was conducted at the ED of our non-academic level 1 trauma centre (Isala, Zwolle, the Netherlands). Patients were enrolled between 14 April and 22 April 2020. Patients (aged ≥16 years) with suspected COVID-19 presenting to the ED underwent POCUS. All patients received current standard of care, including PCR (naso-oropharyngeal swab). Outcome of POCUS was compared with PCR or CT scan outcome to determine diagnostic accuracy. Diagnostic accuracy measures were calculated using 2×2 contingency tables. RESULTS: 100 patients were eligible to participate in this study, data of 93 patients were analysed. 27 (29%) patients were found positive for COVID-19 by PCR or CT. POCUS had a sensitivity of 89% (95% CI 70% to 97%), specificity of 59% (95% CI 46% to 71%), negative predictive value of 93% (95% CI 79% to 98%) and positive predictive value of 47% (95% CI 33% to 61%). In a subgroup of patients without previous cardiopulmonary disease (n=37), POCUS had a sensitivity of 100% (95% CI 70% to 100%), specificity of 76% (95% CI 54% to 90%), negative predictive value of 100% (95% CI 79% to 100%) and positive predictive value of 67% (95% CI 41% to 86%). CONCLUSION: POCUS of the lungs could serve as a valuable, radiation-free tool for excluding pulmonary manifestations of COVID-19 in patients in the ED at the point of assessment, especially in patients without previous cardiopulmonary disease. TRIAL REGISTRATION Dutch Trial Register, No NTR8544.