振幅整合脑电图特征与癫痫儿童神经预后的相关性分析

Correlation between amplitude-integrated electroencephalogram features and neurological prognosis in children with epilepsy

摘要:

目的 分析振幅整合脑电图（aEEG）特征与癫痫儿童神经预后的相关性。方法 选取2019年1月—2021年7月海南省妇女儿童医学中心儿童神经内科收治的127例癫痫儿童，按1年后神经系统发育结局分为异常组与正常组。收集两组临床资料，并于治疗前后分别行aEEG监测，观察其图形连续性（Co）、下界振幅（LB）、睡眠周期（Cy）、宽带（B）表现；采用多因素一般Logistic回归模型分析患儿神经发育结局的影响因素，Pearson法分析aEEG各指标与神经发育结局的相关性，绘制受试者工作特征（ROC）曲线分析aEEG监测评分对患儿神经发育结局的预测价值。结果 经aEEG监测结果显示，治疗1年后神经系统发育正常患儿86例，异常患儿41例。两组治疗前后Co、LB、Cy、B评分及总分（T）的差值比较，差异有统计学意义（P <0.05），正常组Co、LB、Cy、B、T评分高于异常组。多因素一般Logistic回归分析结果显示，Co评分［O^R=1.619（95% CI：1.255，2.123）］、Cy评分［O^R=1.461（95% CI：1.087，1.911）］、T评分［O^R=1.548（95% CI：1.177，2.053）］为患儿神经系统发育结局的影响因素（P <0.05）。Pearson法分析结果显示，Co、Cy、T评分与神经发育结局正常呈正相关（P <0.05）。ROC曲线分析结果显示，aEEG监测中Co、Cy、T评分对神经发育结局预测的截断值分别为1.39分、2.59分和7.96分；3者联合的ROC曲线下面积为0.873（95% CI：0.805，0.941），敏感性与特异性分别为75.6%（95% CI：0.639，0.825）、88.4%（95% CI：0.794，0.972）。结论 aEEG监测中的指标与癫痫患儿神经系统发育密切相关，且对其预后具有预测价值。

Abstract:

Objective To analyze the correlation between amplitude-integrated electroencephalogram (aEEG) features and neurological prognosis in children with epilepsy.Methods One hundred and twenty-seven children with epilepsy in our hospital from January 2019 to July 2021 were enrolled, and classified into abnormal group and normal group according to the one-year neurodevelopment outcome. Clinical data were collected in both groups, and aEEG monitoring was performed before and after treatment. The aEEG features including continuity (Co), sleep-wake cycle (Cy), amplitude of the lower border (LB), bandwidth span (B), and total scores (T) were recorded. Logistic regression analysis was used to screen the factors affecting the neurodevelopmental outcomes. Pearson method was used to discuss the correlation between aEEG indicators and neurodevelopmental outcomes. Then receiver operating characteristic (ROC) curve was plotted to evaluate the diagnostic value of aEEG monitoring in predicting the neurodevelopmental outcomes in children with epilepsy.Results After one year of treatment, aEEG monitoring results showed that 86 children had normal neurodevelopment and 41 children had abnormal neurodevelopment. The difference of Co, LB, Cy, B, and T scores between the two groups before and after treatment was statistically significant (P < 0.05) by t - test. The score of the normal group was higher than that of the abnormal group. Logistic regression analysis showed that Co, Cy, and T scores were the influencing factors of neurodevelopmental outcome in children (P < 0.05). Pearson analysis showed that Co, Cy, and T scores were positively correlated with normal neurodevelopmental outcomes (P < 0.05). ROC curve showed that the cutoff values of Co, Cy, and T scores in aEEG monitoring for predicting neurodevelopmental outcomes were 1.39, 2.59, and 7.96, respectively. The AUC, sensitivity, and specificity of the combined detection of three indicators in predicting the neurodevelopmental outcomes were 0.873 (95% CI: 0.805, 0.941), 75.6% (95% CI: 0.639, 0.825), and 88.4% (95% CI: 0.794, 0.972), which were all larger than those of the separate detection.Conclusion The aEEG monitoring indicators are closely related to the neurodevelopmental outcomes in children with epilepsy, and are of certain predictive value for their prognosis.