Study quality among included studies.

Intracerebral hemorrhage (ICH) is a devastating stroke with many mechanisms of injury. Edema worsens outcome and can lead to mortality after ICH. Glibenclamide (GLC), a sulfonylurea 1- transient receptor potential melastatin 4 (Sur1-Trpm4) channel blocker, has been shown to attenuate edema in ischemic stroke models, raising the possibility of benefit in ICH. This meta-analysis synthesizes current pre-clinical (rodent) literature regarding the efficacy of post-ICH GLC administration (vs. vehicle controls) on behaviour (i.e., neurological deficit, motor, and memory outcomes), edema, hematoma volume, and injury volume. Six studies (5 in rats and 1 in mice) were included in our meta-analysis (PROSPERO registration = CRD42021283614). GLC significantly improved behaviour (standardized mean difference (SMD) = −0.63, [−1.16, −0.09], n = 70–74) and reduced edema (SMD = −0.91, [−1.64, −0.18], n = 70), but did not affect hematoma volume (SMD = 0.0788, [−0.5631, 0.7207], n = 18–20), or injury volume (SMD = 0.2892, [−0.4950, 1.0734], n = 24). However, these results should be interpreted cautiously. Findings were conflicted with 2 negative and 4 positive reports, and Egger regressions indicated missing negative edema data (p = 0.0001), and possible missing negative behavioural data (p = 0.0766). Experimental quality assessed via the SYRCLE and CAMARADES checklists was concerning, as most studies demonstrated high risks of bias. Studies were generally low-powered (e.g., average n = 14.4 for behaviour), and future studies should employ sample sizes of 41 to detect our observed effect size in behaviour and 33 to detect our observed effect in edema. Overall, missing negative studies, low study quality, high risk of bias, and incomplete attention to key recommendations (e.g., investigating female, aged, and co-morbid animals) suggest that further high-powered confirmatory studies are needed before conclusive statements about GLC’s efficacy in ICH can be made, and before further clinical trials are performed.

脑出血（Intracerebral hemorrhage, ICH）是一种具有多种损伤机制的致死性卒中。脑出血后水肿会恶化患者预后，甚至导致死亡。格列本脲（Glibenclamide, GLC）是一种磺酰脲类Sur1-Trpm4通道阻滞剂，已有研究证实其可减轻缺血性卒中模型中的脑水肿，这提示其或许能在脑出血治疗中发挥获益作用。本项荟萃分析整合了当前有关脑出血后给予格列本脲（相较于赋形剂对照组）对行为学（即神经功能缺损、运动及记忆结局）、水肿、血肿体积及损伤体积影响的临床前啮齿类动物研究文献。本项荟萃分析共纳入6项研究（5项大鼠研究、1项小鼠研究），其PROSPERO注册编号为CRD42021283614。格列本脲可显著改善行为学结局（标准化均数差（standardized mean difference, SMD）=−0.63，95%置信区间[−1.16, −0.09]，n=70~74）并减轻水肿（SMD=−0.91，95%置信区间[−1.64, −0.18]，n=70），但对血肿体积（SMD=0.0788，95%置信区间[−0.5631, 0.7207]，n=18~20）及损伤体积无显著影响（SMD=0.2892，95%置信区间[−0.4950, 1.0734]，n=24）。但本研究结果需谨慎解读。纳入研究的结果存在分歧：2项为阴性结果，4项为阳性结果；Egger回归分析提示存在阴性水肿数据的发表偏倚（p=0.0001），同时可能存在阴性行为学数据的发表偏倚（p=0.0766）。通过SYRCLE及CAMARADES量表评估的实验质量令人担忧，多数研究存在较高的偏倚风险。多数研究的统计效力不足（如行为学结局研究的平均样本量仅为14.4）；未来的研究若要检测到本研究观察到的行为学效应量，所需样本量应为41，检测水肿效应量则需33例样本。综上，阴性研究缺失、研究质量偏低、偏倚风险较高，且未充分遵循关键实验建议（如纳入雌性、老年及合并基础疾病的动物），这提示在得出格列本脲对脑出血的疗效结论及开展进一步临床试验前，仍需开展高质量、高统计效力的验证性研究。