Predictive validity of the simplified Radiographic Assessment of Lung Edema score for the mortality in critically ill COVID-19 patients with the B.1.617.2 (Delta) variant in Vietnam: a single-centre, cross-sectional study

Abstract

Background Evaluating the prognosis of COVID-19 patients who may be at risk of mortality using the simple chest X-ray (CXR) severity scoring systems provides valuable insights for treatment decisions. This study aimed to assess how well the simplified Radiographic Assessment of Lung Edema (RALE) score could predict the death of critically ill COVID-19 patients in Vietnam. Methods From July 30 to October 15, 2021, we conducted a cross-sectional study on critically ill COVID-19 adult patients at an intensive care centre in Vietnam. We calculated the areas under the receiver operator characteristic (ROC) curve (AUROC) to determine how well the simplified RALE score could predict hospital mortality. In a frontal CXR, the simplified RALE score assigns a score to each lung, ranging from 0 to 4. The overall severity score is the sum of points from both lungs, with a maximum possible score of 8. We also utilized ROC curve analysis to find the best cut-off value for this score. Finally, we utilized logistic regression to identify the association of simplified RALE score with hospital mortality. Results Of 105 patients, 40.0% were men, the median age was 61.0 years (Q1-Q3: 52.0-71.0), and 79.0% of patients died in the hospital. Most patients exhibited bilateral lung opacities on their admission CXRs (99.0%; 100/102), with the highest occurrence of opacity distribution spanning three (18.3%; 19/104) to four quadrants of the lungs (74.0%; 77/104) and a high median simplified RALE score of 8.0 (Q1-Q3: 6.0-8.0). The simplified RALE score (AUROC: 0.747 [95% CI: 0.617-0.877]; cut-off value >=5.5; sensitivity: 93.9%; specificity: 45.5%; PAUROC <0.001) demonstrated a good discriminatory ability in predicting hospital mortality. After adjusting for confounding factors such as age, gender, Charlson Comorbidity Index, serum interleukin-6 level upon admission, and admission severity scoring systems, the simplified RALE score of >=5.5 (adjusted OR: 18.437; 95% CI: 3.215-105.741; p =0.001) was independently associated with an increased risk of hospital mortality. Conclusions This study focused on a highly selected cohort of critically ill COVID-19 patients with a high simplified RALE score and a high mortality rate. Beyond its good discriminatory ability in predicting hospital mortality, the simplified RALE score also emerged as an independent predictor of hospital mortality.