Abstract: Objective To explore the prediction of the severity of pulmonary infections in severe stroke patients by changes in serum levels of miR-497 and miR-5787 and their correlation with inflammatory response. 
Methods A retrospective study was conducted on 202 patients with severe stroke treated in  the Third Hospital of Xingtai City. The patients were classified into three groups based on the occurrence and severity of pulmonary infection: non-infection group, mild to moderate infection group, and severe infection group. Serum levels of miR-497 and miR-5787 were evaluated in patients from these groups, and inflammatory factors were collected. The correlation between serum miR-497 and miR-5787 levels and inflammatory factors was analyzed using the Pearson method. Univariate and multivariate Logistic regression analyses were performed to identify the risk factors associated with different severities of pulmonary infection. Additionally, the predictive value of serum miR-497 and miR-5787 for different severities of pulmonary infection was assessed using receiver operating characteristic (ROC) curves. 
Results The serum C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α) levels in the non-infection group were lower than those in the mild to moderate infection group and severe infection group, which were lower in the mild to moderate infection group than in the severe infection group (P＜0.05). The serum miR-497 and miR-5787 levels in the non-infection group were lower than those in the mild to moderate infection group and severe infection group, which were lower in the mild to moderate infection group than in the severe infection group. Serum miR-497 was positively correlated with serum CRP, IL-6, and TNF-α (r=0.756, 0.595, 0.718, P＜0.001), and serum miR-5787 was positively correlated with serum CRP, IL-6, and TNF-α (r=0.929, 0.827, 0.878, P＜0.001). Logistic regression analysis showed that miR-497 (OR=1.824, 95%CI: 1.473-2.258) and miR-5787 (OR=2.034, 95%CI: 1.234-3.353) were influencing factors for pulmonary infection in patients with severe stroke (P＜0.05). The ROC results showed that the critical value for miR-497 diagnosis was 3.55, with a corresponding sensitivity of 69.77%, specificity of 71.07%, and the area under the ROC curve (AUC) of 0.801 (95%CI: 0.750-0.852). The critical value for miR-5787 diagnosis was 10.27, with a corresponding sensitivity of 74.42%, specificity of 69.81%, and AUC of 0.760 (95%CI: 0.707-0.812). The sensitivity of regression analysis was 86.05%, specificity was 83.02%, and AUC was 0.869 (95%CI: 0.831-0.907). 
Conclusion Serum levels of miR-497 and miR-5787 are found to be significantly upregulated in patients with severe stroke, which have a positive correlation with the severity of pulmonary infection. They can be used to predict pulmonary infection in severe stroke patients and are worthy of further clinical research and promotion. 

Key words: stroke, respiratory tract infections, miR-497, miR-5787