Postoperative serum amyloid A as a primary marker in a predictive model for ventilator-associated pneumonia in elderly patients with acute ischaemic stroke undergoing endovascular therapy with general anaesthesia.

Background: The risk factors associated with ventilator-associated pneumonia (VAP) in acute ischaemic stroke (AIS) patients who have undergone endovascular therapy have been primarily reported as clinical-related parameters.

Aim: This study aims to combine clinical parameters with inflammatory biomarkers to identify VAP-related risk factors and develop a predictive model.

Methods: A total of 564 AIS patients were recruited and divided into the training set (n = 395) and the validation set (n = 169). The least absolute shrinkage and selection operator (LASSO), univariate and multivariate logistic regression analyses were utilized to examine the independent risk factors or biomarkers associated with VAP.

Findings: We identified four VAP-associated risk factors or biomarker in AIS patients, consisting of thrombolysis in cerebral infarction (TICI) score (0-IIa) (OR = 4.528; 95% CI: 2.249-9.119; P < 0.001), admission national Institute of Health stroke scale (NIHSS) (OR=1.330; 95% CI: 1.217-1.453; P<0.001), neutrophil lymphocyte ratio (NLR) (OR=2.179; 95% CI: 1.312-3.618; P=0.003), and postoperative serum amyloid A (SAA) (OR=1.194; 95% CI: 1.146-1.244; P<0.001). This predictive model demonstrated robust performance and stability, with an AUC of 0.926 (95% CI: 0.899-0.953) in the training set and 0.937 (95% CI: 0.897-0.977) in the validation set. Notably, using the machine learning algorithm Random Forest for feature importance ranking, postoperative SAA emerged as the most critical predictor of VAP.

Conclusion: The predictive model has good predictive value for VAP. Postoperative SAA may serve as a rapid diagnostic biomarker for predicting VAP.