Risk Factor Analysis and Prediction Model of Secondary Epilepsy After Cerebral Infarction
Objective To analyze the risk factors of secondary epilepsy after cerebral infarction in 84 patients,and establish the XGBoost model to analyze the predictive value of this model for secondary epilepsy.Methods A total of 84 patients with secondary epilepsy after cerebral infarction from June 2019 to August 2022 were selected as the study group,and 84 patients with no secondary epilepsy after cerebral infarction during the same period were selected as the control group.Seizures were measured in the study group.Univariate and multivariate logistic regression analysis of the influencing factors of secondary epilepsy after cerebral infarction.The prediction effect of XGBoost model on secondary epilepsy was evaluated by receiver operating characteristic curve and calibration curve.Results In the study group,there were 51 cases of early onset epilepsy,33 cases of delayed onset epilepsy,56 cases of single seizure,28 cases of repeated seizure.Large infarct size,carotid circulatory system,increased National Institute of Health stroke scale(NIHSS)score,mechanical thrombectomy,bleeding transformation and elevated serum levels of homocysteine(Hey)A neuron specific enolase(NSE),soluble interleukin-2 receptor(sIL-2R),lipoprotein associated phospholipase A2(Lp-PLA2)were the risk factors for secondary epilepsy(P＜0.05).XGBoost model ranked the importance scores of different risk factors,which were infarct size,Lp-PLA2 level,NIHSS score,sIL-2R level,hemorrhage transformation,NSE level,mechanical thromboltomy,Hcy level,and involved arterial system in order.XGBoost model predicted an area under curve of 0.826 for secondary epilepsy after cerebral infarction,and Hosmer-Lemeshow goodness of fit test(x2=7.015,P=0.541).Conclusion The increase of infarct size,carotid circulation system,NIHSS score,mechanical thrombolysis,hemorrhage transformation and serum Hcy,NSE,sIL-2R and Lp-PLA2 levels are risk factors for secondary epilepsy after cerebral infarction.The establishment of XGBoost model is helpful for screening high-risk groups,and the model has certain predictive value for secondary epilepsy.It is helpful to guide clinical prevention and treatment of secondary epilepsy after cerebral infarction.
cerebral infarctionsecondary epilepsyinfluencing factorsXGBoost modelpredictive value
万方数据
维普
