Analysis of risk factors for adverse outcomes and development of a predictive model in patients with acute mild ischemic stroke following intravenous thrombolytic therapy with rt-PA

To identify risk factors associated with early unfavorable outcomes in patients with mild ischemic stroke (MIS) treated with intravenous recombinant tissue plasminogen activator (rt-PA), develop a prognostic prediction model, and validate its performance externally using multi-center cohort data. This study retrospectively enrolled 231 patients with MIS who received intravenous thrombolysis with rt-PA at the Second Affiliated Hospital of Fujian Medical University, Jinjiang Hospital, and Anxi County Hospital between January 2015 and September 2023 to form the training cohort. Patients were classified into favorable prognosis (mRS 0–1) and unfavorable prognosis (mRS 2–6) groups according to their mRS scores at discharge. Independent predictors were identified through univariate and multivariate logistic regression analyses, and a predictive nomogram model was developed based on these variables. Subsequently, an independent validation cohort comprising 63 MIS patients treated with rt-PA intravenous thrombolysis at Longyan City First Hospital during the same time period was used for external validation. Model performance was comprehensively assessed in terms of discrimination, calibration, and clinical utility using receiver operating characteristic (ROC) curve analysis, calibration plot, and decision curve analysis (DCA). Multivariate logistic regression analysis identified baseline NIHSS score (OR = 1.650), age (OR = 1.056), early neurological deterioration (END) (OR = 25.242), and anterior circulation infarction (OR = 2.468) as independent predictors of unfavorable prognosis. A nomogram model was subsequently developed based on these variables, demonstrating strong predictive performance in the training cohort with an area under the receiver operating characteristic curve (AUC) of 0.828. In the external validation cohort, the model maintained excellent discrimination, achieving an AUC of 0.827, with a sensitivity of 87.5%, specificity of 72.3%, and a negative predictive value of 94.4%. Calibration plots revealed good agreement between predicted and observed probabilities of unfavorable outcomes. Decision curve analysis further demonstrated that the model provides substantial clinical net benefit across a broad range of threshold probabilities. Age, baseline NIHSS score, END, and anterior circulation stroke are independent predictors of unfavorable prognosis in patients with MIS following intravenous thrombolysis with rt-PA. The predictive nomogram model developed from these variables demonstrated excellent accuracy, generalizability, and clinical utility after external validation, supporting the identification of high-risk individuals and aiding clinical decision-making.