基于组分解析分类的屋尘螨过敏原免疫治疗对过敏性鼻炎的疗效：一项网络荟萃分析

文献信息

中文摘要

背景： 屋尘螨过敏原免疫治疗（HDM-AIT）在过敏性鼻炎中的试验显示出显著的疗效异质性，传统上按给药途径进行评估。不同过敏原组分广度的HDM-AIT制剂之间临床疗效是否存在差异仍不清楚。
目的： 采用组分解析框架，按过敏原组分广度对制剂进行分层，以比较HDM-AIT的疗效。
方法： 我们检索了PubMed、Embase和CENTRAL，纳入持续时间至少12个月的HDM-AIT双盲、安慰剂对照随机试验。使用预先设定的组分解析框架将干预措施分类为综合组分皮下免疫治疗（CC-SCIT）、主要组分优势皮下免疫治疗（MCD-SCIT）或主要组分优势舌下免疫治疗片剂（MCD-SLIT-tablet），其中以治疗诱导的组分特异性IgG4反应广度作为主要节点定义标准，并以组分特异性IgG和蛋白质组学证据作为支持性证据。进行了频率学配对和贝叶斯网络荟萃分析。主要结局为症状评分；次要结局为药物评分以及症状与药物联合评分。
结果： 共纳入16项试验，涉及7,329名参与者。对于症状评分，CC-SCIT显示出相对于安慰剂最有利的估计治疗效果（标准化均数差[SMD]，-0.40；95%可信区间[CrI]，-0.73至-0.11；累积排名曲线下面积[SUCRA]，82.6%），而MCD-SLIT-tablet显示出更精确的估计，并有更大的证据基础支持（SMD，-0.30；95% CrI，-0.41至-0.20；SUCRA，60.7%）。对于药物评分，CC-SCIT（SMD，-0.40；95% CrI，-0.81至0.00；SUCRA，79.8%）和MCD-SCIT（SMD，-0.38；95% CrI，-0.72至-0.05；SUCRA，78.5%）显示出相似的估计效应，而MCD-SLIT-tablet显示出较小但更精确的效应（SMD，-0.15；95% CrI，-0.28至0.00；SUCRA，39.9%）。对于症状与药物联合评分，CC-SCIT显示出相对于安慰剂最大的估计效应（SMD，-0.84；95% CrI，-1.53至-0.37；SUCRA，99.5%）。网络呈星形，非安慰剂治疗节点之间的比较为间接比较。后验排名分布也表明治疗节点相对排序存在不确定性，尤其是那些由较少试验支持的节点。
结论： HDM-AIT制剂所报告的免疫学和组分特征与试验层面疗效估计的变异性相关。这些发现支持将产品层面的分子表征和直接比较研究作为未来精准AIT的优先事项。
系统评价注册： https://www.crd.york.ac.uk/PROSPERO/view/，CRD420261307571。

英文摘要

BACKGROUND: House dust mite allergen immunotherapy (HDM-AIT) trials in allergic rhinitis show substantial efficacy heterogeneity, traditionally evaluated by administration route. Whether clinical efficacy differs among HDM-AIT preparations with different allergen compositional breadth remains unclear.
OBJECTIVE: To compare HDM-AIT efficacy using a component-resolved framework that stratifies preparations by allergen compositional breadth.
METHODS: We searched PubMed, Embase, and CENTRAL for double-blind, placebo-controlled randomized trials of HDM-AIT lasting at least 12 months. Interventions were classified as comprehensive-component subcutaneous immunotherapy (CC-SCIT), major-component-dominant subcutaneous immunotherapy (MCD-SCIT), or major-component-dominant sublingual immunotherapy tablet (MCD-SLIT-tablet) using a prespecified component-resolved framework in which the breadth of treatment-induced component-specific IgG4 responses served as the primary node-defining criterion, with component-specific IgG and proteomic evidence used as supportive evidence. Frequentist pairwise and Bayesian network meta-analyses were performed. The primary outcome was symptom score; secondary outcomes were medication score and combined symptom and medication score.
RESULTS: Sixteen trials involving 7,329 participants were included. For symptom score, CC-SCIT showed the most favorable estimated treatment effect versus placebo (standardized mean difference [SMD], -0.40; 95% credible interval [CrI], -0.73 to -0.11; surface under the cumulative ranking curve [SUCRA], 82.6%), while MCD-SLIT-tablet showed a more precise estimate supported by a larger evidence base (SMD, -0.30; 95% CrI, -0.41 to -0.20; SUCRA, 60.7%). For medication score, CC-SCIT (SMD, -0.40; 95% CrI, -0.81 to 0.00; SUCRA, 79.8%) and MCD-SCIT (SMD, -0.38; 95% CrI, -0.72 to -0.05; SUCRA, 78.5%) showed similar estimated effects, whereas MCD-SLIT-tablet showed a smaller but more precise effect (SMD, -0.15; 95% CrI, -0.28 to 0.00; SUCRA, 39.9%). For combined symptom and medication score, CC-SCIT showed the largest estimated effect versus placebo (SMD, -0.84; 95% CrI, -1.53 to -0.37; SUCRA, 99.5%). The network was star-shaped, and comparisons among non-placebo treatment nodes were indirect. Posterior rank distributions also indicated uncertainty in the relative ordering of the treatment nodes, particularly those supported by fewer trials.
CONCLUSION: Reported immunologic and compositional profiles of HDM-AIT preparations were associated with variability in trial-level efficacy estimates. These findings support product-level molecular characterization and direct comparative studies as priorities for future precision AIT.
SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/PROSPERO/view/, CRD420261307571.