Linear Chart Model for Adverse Prognosis within one Year in Acute Ischemic Stroke Patients

Objective The aim of this study was to explore the risk factors influencing adverse outcomes in patients with acute ischemic stroke (AIS)within one year and establish a linear prediction model based on them. Methods We conducted a retrospective analysis of 600 AIS patients treated at our hospital from January 2019 to June 2023. They were divided into an observation group (n=100, adverse prognosis) and a control group (n=500, good prognosis) based on the occurrence of adverse events within one year. Statistical analysis of intergroup differences was performed using the chi-square test, independent sample t-test, and Mann-Whitney U test. Single-factor, multiple-factor logistic regression, and Lasso regression analyses were conducted using the glmnet package to identify independent risk factors affecting AIS. Risk factors influencing adverse outcomes in AIS were depicted using column charts with the "rms" package.Bootstrap method was used for internal validation of the model. Results Single-factor logistic regression showed that age, admission NIHSS score, blood sugar, creatinine, blood urea nitrogen, white blood cell count, smoking history, stroke history, concurrent pneumonia, inability to walk within 48 hours of admission, and atrial fibrillation were the main risk factors ( P <0.05). Multiple-factor logistic regression revealed that age, admission NIHSS score, concurrent pneumonia, inability to walk within 48 hours of admission, and atrial fibrillation were independent risk factors influencing adverse outcomes in AIS patients within one year ( P <0.05). The ROC curve for the AIS adverse prognosis column chart model within one year showed high credibility, with a training set AUC of 0.993 (0.988-0.998) and a validation set AUC of 0.987 (0.969-1.000). Conclusion We has successfully constructed a risk prediction model based on a linear chart, which can be used to predict adverse outcomes in AIS patients within one year with high reliability.