神经功能损害儿童胃造瘘术的结局：一项系统评价

文献信息

中文摘要

引言： 胃造瘘喂养广泛用于存在经口摄入不足的神经功能损害儿童，但跨家庭重要结局的证据尚未被综合。既往系统评价仅限于脑性瘫痪，且未发现对照研究。本研究旨在评估胃造瘘与继续经口/鼻胃管喂养相比对营养、呼吸、神经发育和照护者报告结局的影响，并量化并发症。
方法： 按照PRISMA 2020指南开展了一项PROSPERO注册的系统评价（CRD420261483110）。检索5个数据库至2026年7月15日，纳入0-18岁接受胃造瘘的神经功能损害儿童。进行了双人筛选、数据提取和偏倚风险评估（RoB 2、ROBINS-I、JBI、CASP）。异质性使Meta分析不可行；叙述性综合优先考虑研究设计和方法学质量，而非按效应方向计票。
结果： 从5,503条记录中，120项研究（约117个队列，30个国家）符合标准。证据以观察性研究为主，多数比较性研究存在严重偏倚风险。在56项研究中，大多数研究显示营养状况改善，尽管较早置管与更显著的追赶生长相关，但这不能确立因果关系。在56项呼吸研究中，较小队列报告感染较少，但最大的倾向匹配和人群关联数据集显示肺炎住院无减少，在某些分析中反而增加。轻微并发症常见（20%-95%）。未同时行抗反流手术时，主要并发症集中在0%-3%，但增加胃底折叠术或空肠延伸后升高数倍。反流结局（59项研究）取决于术前状态；根据有限的生理学数据推测，胃造瘘可能增加既往无反流儿童的的新发反流风险，但<10%需要后续抗反流手术。全因死亡率（67项研究）较高但与疾病相关；手术归因死亡罕见，在最脆弱群体之外30天死亡率为0%-3%。生活质量数据（21项研究）显示14/17项研究中照护者结局有利，但儿童的生活质量仅被恰当测量过一次且无变化。仅11项研究涉及神经发育；唯一使用标准化工具的研究发现认知恢复与喂养途径无关。
讨论： 胃造瘘始终与体重相关营养状况改善相关，并且在未同时行抗反流手术时严重手术伤害发生率低。它不能消除误吸风险，并推测可能增加既往无反流儿童的新发反流风险。照护者生活质量在经验证的测量中改善；然而，现有证据不足以确定胃造瘘对儿童神经发育是有利、不利还是中性影响。
系统评价注册： PROSPERO：https://www.crd.york.ac.uk/PROSPERO/view/CRD420261483110。

英文摘要

INTRODUCTION: Gastrostomy feeding is widely used for children with neurological impairment and inadequate oral intake, but evidence across family-important outcomes has not been synthesised. Previous systematic reviews were restricted to cerebral palsy and found no controlled studies. The objective of this study is to assess gastrostomy vs. continued oral/nasogastric feeding on nutritional, respiratory, neurodevelopmental, and caregiver-reported outcomes, and to quantify complications.
METHODS: A PROSPERO-registered systematic review (CRD420261483110) was conducted following PRISMA 2020 guidelines. Five databases were searched to 15 July 2026 for children aged 0-18 with neurological impairment undergoing gastrostomy. Duplicate screening, data extraction, and risk-of-bias appraisal (RoB 2, ROBINS-I, JBI, CASP) were performed. Heterogeneity precluded meta-analysis; narrative synthesis prioritized study design and methodological quality over direction-of-effect vote counting.
RESULTS: From 5,503 records, 120 studies (≈117 cohorts, 30 countries) met the criteria. The evidence was predominantly observational, and most comparative studies carried a serious risk of bias. Nutritional status improved in the majority of 56 studies, though earlier placement's association with more robust catch-up growth does not establish causality. Among 56 respiratory studies, smaller cohorts reported fewer infections, but the largest propensity-matched and population-linked datasets showed no reduction, and in some analyses an increase, in pneumonia hospitalisations. Minor complications were common (20%-95%). Major complications clustered at 0%-3% without concurrent antireflux surgery, but rose several-fold when fundoplication or jejunal extensions were added. Reflux outcomes (59 studies) depended on preoperative status; it is hypothesized from limited physiological data that gastrostomy may increase de novo reflux risk in previously reflux-free children, though <10% required subsequent antireflux surgery. All-cause mortality (67 studies) was high but disease-related; procedure-attributable death was rare, and 30-day mortality was 0%-3% outside the most fragile groups. Quality-of-life data (21 studies) showed favourable caregiver outcomes in 14/17 studies, but the child's quality of life has been measured properly only once and was unchanged. Only 11 studies addressed neurodevelopment; the single study using a standardised instrument found cognitive recovery unrelated to the feeding route.
DISCUSSION: Gastrostomy is consistently associated with improved weight-based nutritional status and carries low serious surgical harm when performed without concurrent antireflux surgery. It does not eliminate aspiration risk and is hypothesized to increase de novo reflux risk in previously reflux-free children. Caregiver quality of life improves on validated measurement; however, the available evidence is insufficient to determine whether gastrostomy has a favourable, unfavourable, or neutral effect on a child's neurodevelopment.
SYSTEMATIC REVIEW REGISTRATION: PROSPERO: https://www.crd.york.ac.uk/PROSPERO/view/CRD420261483110.