口服第二代H1抗组胺药方案治疗过敏性鼻炎：疗效与可估计性的系统评价和网络荟萃分析

文献信息

中文摘要

背景/目的： 口服第二代H1抗组胺药的比较效应因鼻炎类别、结局指标和治疗时期而异。我们评估了各方案之间的证据结构、可估计性和区分度。方法：检索了PubMed、Embase、Web of Science核心合集、CENTRAL、WHO ICTRP、ClinicalTrials.gov和EU临床试验注册库，检索日期为2026年7月1日至3日，纳入截至2026年6月30日的记录。合格研究为自然暴露随机试验。连续结局采用协方差感知广义最小二乘法分析，二分类伤害采用贝叶斯臂水平模型分析。RoB 2、CINeMA和ROB-MEN由两人重复应用。注册为回顾性。结果：61篇报告从64个队列（34,831名参与者）中产生66条报告-随机化条目；35个模型就绪随机化提供了9种结构的信息。在季节性过敏性鼻炎中，7个TNSS估计中的6个、3个眼部估计中的2个以及4个RQLQ估计均支持积极治疗。任何不良事件估计不精确。在主要先验下，西替利嗪的WDAE OR为0.25（95% CrI 0.08-0.66），但预测分类对先验敏感。在34个连续活性-活性比较中，7个有直接信息，27个仅间接。所有147个可估计的二分类活性-活性对比均作为次要推导报告：141个CrI包含1，6个排除1，均为仅间接且未用CINeMA正式评价。14个方案水平节点中有6个仅对耐受性结构有贡献，妨碍了平衡的方案特异性获益-伤害评估。在定义的CINeMA评价范围内，所有89个可估计比较均评为极低质量；两个不可估计效应无法评价。结论：以安慰剂锚定为主且稀疏的证据既未建立可靠的方案层级，也未建立治疗等效性。研究结果具有假设生成性，嗜睡仍是一个证据缺口。无外部资金支持实施；NEXUS PEOPLE, S.L.资助了APC。OSF注册为回顾性：doi:10.17605/OSF.IO/DPT5N。

英文摘要

Background/Objectives: Comparative effects of oral second-generation H1-antihistamines vary by rhinitis category, outcome, and treatment period. We assessed evidence structure, estimability, and differentiation among regimens. Methods: PubMed, Embase, Web of Science Core Collection, CENTRAL, WHO ICTRP, ClinicalTrials.gov, and the EU Clinical Trials Register were searched on 1-3 July 2026 for records through 30 June 2026. Eligible studies were natural-exposure randomized trials. Continuous outcomes were analyzed using covariance-aware generalized least squares, and binary harms were analyzed using Bayesian arm-level models. RoB 2, CINeMA, and ROB-MEN were applied in duplicate. Registration was retrospective. Results: Sixty-one reports generated 66 report-randomization entries from 64 cohorts (34,831 participants); 35 model-ready randomizations informed nine structures. In seasonal allergic rhinitis, six of seven TNSS estimates, two of three ocular estimates, and four RQLQ estimates favored active treatment. Any-adverse-event estimates were imprecise. Under the primary prior, cetirizine yielded a WDAE OR of 0.25 (95% CrI 0.08-0.66), but predictive classification was prior-sensitive. Of 34 continuous active-active comparisons, seven were directly informed and 27 indirect-only. All 147 estimable binary active-active contrasts were reported as secondary derivations: 141 CrIs included one and six excluded one, all indirect-only and not formally appraised with CINeMA. Six of 14 regimen-level nodes contributed only to tolerability structures, precluding balanced regimen-specific benefit-harm assessment. All 89 estimable comparisons within the defined CINeMA appraisal scope were rated very low; two non-estimable effects were not evaluable. Conclusions: Predominantly placebo-anchored and sparse evidence established neither a reliable regimen hierarchy nor therapeutic equivalence. Findings were hypothesis-generating, and somnolence remained an evidence gap. No external funding supported conduct; NEXUS PEOPLE, S.L. funded the APC. OSF registration was retrospective: doi:10.17605/OSF.IO/DPT5N.