儿童难治性癫痫对患儿的身体精神发育危害巨大，所以越早进行干预治疗和发作控制对儿童发育越有益。迷走神经刺激（VNS）疗法作为一种神经调控技术，已被批准作为难治性癫痫的辅助治疗，并在临床上取得广泛应用。然而对于低龄儿童，VNS的有效性和安全性缺乏高等级证据的支持。难治性癫痫以及VNS疗法对儿童心脏自主神经影响的相关研究存在研究数量目少，结论不一致的问题，其中低龄儿童的研究依旧空白。VNS总体上约50%的患者的发作频率可降低50%以上，但其疗效个体差异极大。术前筛选适合VNS疗法的患者可提高VNS疗效，但已发现的疗效预测生物标记物及预测模型离临床应用还很远。本文通过一项随机对照临床试验验证VNS治疗低龄儿童难治性癫痫的效果，通过基于心率的心脏自主神经功能评估方法，研究难治性癫痫和VNS治疗对心脏自主神经功能的效应，并探索VNS术前的临床因素和心脏自主神经功能指标预测长期疗效的可能性。本文的主要工作包括以下三部分： 1、通过心率变异性和复杂度方法，分析了93例儿童难治性癫痫和46例健康对照的24小时动态心电，结果表明了低龄儿童难治性癫痫的心脏自主神经调节功能整体受到抑制，包括迷走神经和交感神经调节能力以及心率复杂度的降低。此外表征迷走神经活性的心率变异性指标可有效区分难治性癫痫和健康状态，心率变异性指标结合心率复杂度指标可提高区分能力。2、基于世界上首个VNS治疗3-6岁低龄难治性癫痫的随机对照临床试验，观测了93例患儿术后疗效，比较术前基线和植入手术后、治疗24周后的心率变异性和复杂度变化。结果表明，VNS疗法对于低龄儿童有效且足够安全，长期刺激后发作控制更佳。24周的VNS治疗可增强迷走神经与交感神经调节能力，但植入手术对心脏自主神经功能无明显的急性影响。此外，还发现了低龄儿童植入VNS术后存在“蜜月期效应”。3、基于93例低龄儿童难治性癫痫术前临床因素和心脏自主神经功能评估指标，发现了心率变异性、心率复杂度以及部分临床因素与长期疗效之间的相关性，说明了VNS治疗时间、疗效分组对预测指标的影响，得出了建立多变量预测模型以及稳定发作控制下的预测模型效果更佳，达到临床上认可预测效果的中等评价水平，为适合VNS疗法的患者筛选提供了客观量化的依据。
Drug-resistant epilepsy in children cause great harm to their physical and mental development, while the earlier intervention and seizure control is more beneficial to children's development. Vagus nerve stimulation(VNS)，a neuroregulatory technique，has been approved as an adjunctive therapy for drug-resistant epilepsy and has been widely used clinically. However, for younger children, the efficacy and safety of VNS therapy lacks the support of high-level clinical evidence. The studies on the cardiac autonomic function of drug-resistant epilepsy and VNS therapy in children have few and inconsistent conclusions. Altogether, about 50% patients can be seizures reduced by more than 50%, but the individual curative effect varies greatly. Preoperative screening of patients suitable for VNS therapy can improve VNS efficacy, however, the found efficacy prediction biomarkers and prediction models are far from clinical application. This dissertation verified the effect of VNS therapy on younger children with drug-resistant epilepsy through a randomized controlled clinical trial. By heart rate-based assessment of cardiac autonomic nervous function, the effects of drug-resistant epilepsy and VNS therapy on cardiac autonomic nervous function were studied. The possibility of predicting long-term efficacy was explored through preoperative clinical factors and cardiac autonomic nervous function indicators. The main work of this dissertation includes three parts:1. The 24-hour holter ECG of 93 children with drug-resistant epilepsy and 46 healthy controls were analyzed by heart rate variability and heart rate complexity. The results showed that the whole cardiac autonomic regulation function was inhibited in younger children with drug-resistant epilepsy, including the decrease of vagus nerve and sympathetic tension and the reduction of heart rate complexity. Moreover, the HRV index characterizing vagus nerve activity can effectively distinguish drug-resistant epilepsy from healthy state, and the combination of HRV index and heart rate complexity index can improve the ability of differentiation.2. Based on the world's first randomized controlled clinical trial of VNS for children aged 3-6 years with drug-resistant epilepsy, the postoperative efficacy of 93 children was observed, and the measurements of heart rate variability and complexity were compared between baseline and after implantation surgery, 24 weeks after treatment. The results showed that VNS therapy was effective and safe enough for younger children, and the seizure control improved better after long-term stimulation. 24 weeks of VNS therapy improved vagus and sympathetic regulation, but implantation surgery has no significant acute effect on cardiac autonomic function. Besides, there was a "honeymoon effect" after VNS implantation in younger children.3. Based on the preoperative clinical factors and cardiac autonomic function evaluation index of 93 younger children with drug-resistant epilepsy, the relationship between heart rate variability, heart rate complexity, some clinical factors and long-term efficacy was found. The effects of VNS treatment duration and efficacy grouping on prediction indexes were explained. The results showed that the multivariable prediction model and prediction model for stable seizure control were more effective and reached the middle evaluation level of clinical recognition prediction effect, which provided an objective and quantitative basis for the patient screening of VNS therapy.
