The value of lung ultrasound score on evaluating clinical severity and prognosis in patients with acute respiratory distress syndrome

Abstract

To evaluate the value of lung ultrasound score (LUS) on assessing the severity and prognosis in patients with acute respiratory distress syndrome (ARDS), and to investigate its correlation with oxygenation index, acute physiology and chronic health evaluationII (APACHEII) score, sequential organ failure assessment (SOFA) score, and clinical pulmonary infection score (CPIS), and other traditional parameters. A prospective double-blind cohort study was conducted. Sixty-two ARDS patients conformed to the Berlin diagnostic criteria admitted to intensive care unit (ICU) of Beijing Huaxin Hospital from October 2013 to December 2014 were enrolled, including 14 cases with mild, 18 moderate, and 30 severe ARDS; among them 37 cases were of ARDS with pulmonary origin, and 25 non-pulmonary ARDS; 35 patients survived, and 27 died. The clinical data and scores of all patients were recorded by one specialized observer, including baseline data, hemodynamic parameters, lactate, respiratory parameters, and APACHEII, SOFA and CPIS scores. Another observer of recording was responsible for the results of lung ultrasound, LUS, and echocardiogram. The correlation between LUS and oxygenation index as well as APACHEII, SOFA and CPIS scores was analyzed by bivariate correlation analysis. Receiver operator characteristic curve (ROC) was plotted, and the predictive value, sensitivity and specificity of mild ARDS, moderate ARDS, severe ARDS and mortality by LUS were calculated. LUS had a negative correlation with oxygenation index (r=-0.755, P<0.001), a good positive correlation with APACHEII (r=0.504, P<0.001), SOFA (r=0.461, P<0.001) and CPIS (r=0.571, P<0.001) was found. LUS in the pulmonary ARDS group had a positive correlation with CPIS (r=0.399, P<0.05), and a positive correlation was found in non-pulmonary ARDS group (r=0.350, P<0.05), which indicated that the correlation in pulmonary ARDS was more satisfactory than that in non-pulmonary ARDS. LUS in the pulmonary ARDS group was significantly higher than that in non-pulmonary ARDS group (22.1±4.9 vs. 11.3±2.1, t=11.667, P<0.001); LUS in mild, moderate, severe ARDS groups was 9.9±1.7, 14.0±1.4, 23.6±4.1. The predictive value for mild ARDS by LUS was 7.0, sensitivity of 87.0%, specificity of 89.0%; that for moderate ARDS was 11.0, sensitivity of 89.0%, specificity of 87.0%; that for severe ARDS was 8.0, sensitivity of 90.0%, specificity of 88.5%. LUS was 24.3±3.8 in the death group, and 12.7±2.9 in the survival group. Area under ROC curve (AUC) was calculated, and the patients with LUS>19.0 had a high mortality, sensitivity for predicting death was 84.0%, and specificity of 89.0%. Bedside LUS, which is simple and easily available, could evaluate the changes in pulmonary ventilation area of ARDS, and its degree of severity, and prognosis including prediction of mortality of the patients.