EXPRESS: Risk factors and predictive model for non-aspiration stroke-associated infections in patients with acute ischemic stroke followed by intravenous thrombolysis.
Abstract
Objective
To analyze clinical characteristics, risk factors, and a predictive model for non-aspiration stroke-associated infections (NASAI) in acute ischemic stroke (AIS) patients receiving intravenous alteplase thrombolysis.
Methods
This prospective cohort study included AIS patients from the DATIS cohort (Jan 2018-Jul 2024) who received rt-PA thrombolysis. NASAI was defined as stroke-associated infection without evidence of aspiration. Patients were randomized 7:3 into training (n=1171) and validation (n=503) sets. LASSO and logistic regression were used to select predictors. Model performance was assessed by ROC curve, calibration curve (CC), and decision curve analysis (DCA).
Results
Among 1828 patients, NASAI prevalence was 10.72% (196/1828), accounting for 56% of all SAI cases. Independent risk factors included older age (OR 1.62), hypertension (1.64), higher admission NIHSS (4.03), heart rate (1.06), and white blood cell count (1.16). The nomogram showed AUCs of 0.80 (training) and 0.78 (validation), with good calibration and net benefit on DCA.
Conclusion
A nomogram based on five easily accessible variables demonstrated strong predictive performance for NASAI, offering a practical tool for early risk assessment in thrombolyzed AIS patients.