Assessment of lung parenchyma recovery after antibiotic administration using lung ultrasound in critically ill patients with pneumonia

BACKGROUND: Pneumonia is a common cause of intensive care unit (ICU) admission, requiring frequent imaging for following up parenchymal lung involvement and antibiotic response. Being bedside and non-invasive technique; lung ultrasound (LUS) is increasingly used in ICU. AIM: The objectives of the study were to assess the accuracy of LUS in detecting parenchymal lung recovery following antibiotic administration in critically ill patients with pneumonia. METHODS: Fifty patients with pneumonia were included in the study with time-dependent analysis for acute physiology and chronic health evaluation II, CURB-65 and modified clinical pulmonary infection score. LUS at day 0 described basal lung condition then according to changes in lung parenchyma, ultrasound (US) score could be first calculated at day 3. At day 5, US score was calculated again and changes in score (delta score) were calculated to assess ability of US to predict early good antibiotic response and finally LUS was repeated at day 7, score calculated to detect lung parenchyma recovery and compared with follow-up computed tomography (CT) for accuracy and agreement. Air bronchogram was reported whenever seen, described as static or dynamic, and assessed in follow-up examinations to be compared with CT follow-up. RESULTS: LUS score ranged from –2 to 17 with mean value of 8.75 ± 3.88 for improving patients, while worsening patients showed LUS score of –11 to –20 with mean value of –10.08 ± 6.95 with high statistical significance (p < 0.001). The best cutoff value of LUS score changes for detecting good response to antibiotic was 2.5, detected using area under the curve (p < 0.001). US score on day 7 showed excellent sensitivity and specificity of 91.89–92.31%, respectively, when compared to CT with positive predictive value (PPV) of 97.14% and negative predictive value (NPV) 80% and accuracy 92% with strong statistical significance (p < 0.001). Air bronchogram showed sensitivity of 61.5% and specificity of 89.1% and with PPV of 66.67% and NPV of 86.84% and accuracy of 82% and moderate agreement (0.52) with CT while B-lines were significant for assessing lung reaeration with sensitivity of 69.2% and specificity of 67.5% and accuracy of 68% but with fair (0.31) agreement with CT (p < 0.027) in detecting parenchymal lung recovery. CONCLUSION: LUS is a reasonable bedside method for quantifying parenchymal lung recovery in patients with pneumonia who are successfully treated with antibiotics.