[Lung Ultrasound to Predict Unfavorable Progress in Patients Hospitalized for COVID-19]

Objective: Thoracic ultrasound has been shown to be useful in the diagnosis of COVID-19 pulmonary involvement. Several scores for quantifying the degree of involvement have been described, although there is no evidence to show that they have any capacity for predicting unfavorable progress.

Methodology: Prospective cohort study of patients hospitalized for COVID-19. The sample was stratified according to clinical course, and patients requiring invasive or non-invasive respiratory support were classified as having unfavorable progress. Biomarkers were analyzed at admission and on the same day that thoracic ultrasound was performed. Prognostic scales were also determined at admission. The ultrasound score was obtained in 8 or 14 areas, depending on the patient's ability to sit.

Results: We included 44 patients, 13 (29,5%) of whom subsequently needed ventilatory support. Eight areas were explored in all patients and 14 areas in 35 (79.5%). The most affected areas were the posterior lower lobes. Significant differences were found between the 2 groups on the SOFA and quick SOFA multidimensional scales, and PCR and LDH on the same day as thoracic ultrasound, and the ultrasound scores. The best area under the ROC curve (AUC) was obtained with the 14-area score, with a result of 0.88 (95% CI: 0.75-0.99). Its sensitivity and specificity for a cut-off score of 13.5 were 100% and 61.5%, respectively.

Conclusions: The use of scores to quantify lung involvement measured by thoracic ultrasound provides useful information, facilitating risk stratification in patients hospitalized with COVID-19.