An analysis of the 28-day mortality risk factors in acute respiratory distress syndrome patients and the establishment of prediction models

摘要

Objective: To explore the risk factors and prediction models of 28-day mortality in acute respiratory distress syndrome (ARDS) patients. Methods: A total of 215 ARDS patients treated in our hospital were enrolled in this prospective observational study, including 70 patients who died within 28 days and were placed in the death group, and the remaining 145 patients who survived and were placed in the survival group. The laboratory examination indexes and critical scoring system scores were compared between the two groups. A Cox regression analysis was used to analyze the factors associated with 28-day mortality, and a receiver operating characteristic (ROC) curve was used to analyze the performance of the prediction models. Results: The ROC curve analysis showed that the erythrocyte distribution width (RDW), the neutrophil to lymphocyte ratio (NLR), the procalcitonin to albumin ratio (PAR), and the Murray lung injury score (MLIS) were effective at diagnosing the 28-day mortality, each with an area under the curve (AUC) of 0.5 (P0.001). A multivariate Cox analysis showed that the RDW, NLR, PAR, and MLIS were independent predictors of 28-day mortality. The results of the multi-index joint prediction showed that the AUC of RDW+NLR+PAR+MLIS was 0.945 (95% CI: 0.910-0.979), and the sensitivity was as high as 94.25%. Conclusion: NLR, PAR, RDW, and MLIS are independent predictors of 28-day mortality, and their combined prediction can significantly improve the predictive ability of the 28-day mortality in ARDS patients.