围生期脑损伤高危新生儿神经行为异常早期预测模型构建及验证OA

Objective To establish an early prediction model of neurobehavioral abnormalities in neonates at high risk of brain injury based on amplitude integrated electroencephalogram(aEEG)and perinatal data.Methods The aEEG and perinatal data of 246 high-risk neonates with brain injury admitted to Zigong Hospital of Woman and Child Health Care from September 2021 to June 2023 were retrospectively analyzed within 72 h after birth,and were divided into normal group(n=186)and abnormal group(n=60)according to the neurobehavioral screening at 12 months of age.Perinatal characteristics and aEEG parameters were compared between the two groups.Multivariate logistic regression analysis was performed to identify independent influencing factors for neurobehavioral abnormalities.A nomogram prediction model was constructed based on the regression results and validated using receiver operating characteristic(ROC)curves and the area under the curve(AUC).Clinical utility was evaluated by decision curve analysis(DCA).Results The proportion of very low birth weight,forceps assisted delivery,NBNA score≤35,non-continuous,non-periodic aEEG background activity and seizure wave in abnormal group was higher than that in normal group,and the 1 min Apgar score,aEEG broadband voltage lower boundary,narrowband voltage upper boundary,narrowband voltage lower boundary and total score were lower than those in normal group(t/χ2 values ranged from 2.874 to 57.970,P<0.05).Multivariate Logistic regression analysis showed that preterm birth,very low birth weight,forceps delivery and NBNA score≤35 were independent risk factors for neurobehavioral abnormalities in neonates at high risk of brain injury(OR values ranged from 0.446 to 0.722,P<0.05),and 1 min Apgar score and aEEG total score were independent protective factors for neurobehavioral abnormalities in neonates at high risk of brain injury(OR values ranged from 0.446 to 0.722,P<0.05).Based on the above regression model,a nomogram prediction model for neurobehavioral abnormalities in neonates at high risk of brain injury was established.The model yielded a C-index of 0.928 and an AUC of 0.928,with a nomogram calibration degree of 0.876.Decision curve analysis showed that applying this nomogram to predict neurobehavioral abnormalities in neonates at high risk of brain injury could achieve a positive clinical net benefit.Conclusion Neonatal amplitude integrated EEG combined with preterm birth,very low birth weight,forceps assisted delivery,NBNA score,1 min Apgar score can be used for early prediction and evaluation of neurobehavioral abnormalities in neonates at high risk of brain injury,and provide reference for early clinical prediction and formulation of corresponding intervention programs.