This meta-analysis was performed in strict accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and was registered in PROSPERO (CRD42023427591).

Both English databases (including EMBASE, PubMed, Cochrane library, and Web of Science) and Chinese databases (including CNKI, VIP, Wanfang, and Chinese Biomedical Literature Database) were searched comprehensively from the date of their respective inception to May 2023 for the identification of eligible data. The following terms used in the search are a combination of MESH terms and free-text words: (“Anisodine hydrobromide injection” (Text word) OR “Anisodine hydrobromide” (Text word) OR “Anisodine” (Text word) AND [“Acute ischemic stroke” (Text Word) OR “ischemic stroke” (MESH) OR “brain ischemia” (MESH) OR “stroke” (MESH) OR “cerebral infarction” (MESH) OR “Cerebrovascular ischemia” (Text word) OR “Infarction, Anterior Cerebral Artery” (Text word) OR “Infarction, Middle Cerebral Artery” (Text word) OR “Infarction, Posterior Cerebral Artery” (Text word) OR “Apoplexy” (Text word)]. Potential studies in the reference lists of valid studies were also considered as information sources.

The inclusion criteria were as follows: 1) patients with AIS, regardless of age, gender, and disease stage; 2) parallel RCTs of Ani injection for AIS patients published in English or Chinese databases; 3) control group treated with regular therapies, while Ani injection was not applied in the control group; and 4) the trial groups were treated with Ani injection, used alone or in combination with the same regular therapies used in the control groups, regardless of the dose or duration of administration.

The exclusion criteria were as follows: 1) cerebral hemorrhage in patients with AIS; 2) reviews, letters, conference reports, cohort studies, case reports, cross-over studies, and animal studies; 3) duplicate studies or those with no comparison group; 4) literature without essential information or unable to obtain the related data; and 5) the trial group underwent acupuncture.

In this systematic review and meta-analysis, the primary outcomes were as follows: National Institutes of Health Stroke Scale (NIHSS), modified Rankin Scale (mRS), and Barthel Index (BI). The secondary outcomes were computed tomography parameters (CTP), effective rate, and adverse events.

All electronic bibliographic databases mentioned above were scanned with a pre-designed search strategy. Duplicate articles were removed first. Next, two independent reviewers reviewed the titles and abstracts of the studies to select appropriate studies according to the eligibility criteria. The full texts of the selected studies were downloaded for further assessment. Three initial articles were used as a pilot to establish a standard extraction form, which contains the following domains: study information (title, first author, language, magazine, and year of publication), participant information (e.g., age, sex ratio, sample size, and disease course), intervention information (e.g., type, duration, frequency, and dose of treatment in the trial and control groups), and outcome indexes (primary outcomes and secondary outcomes). Reasons for the exclusion of ineligible studies were identified and recorded. Available data were extracted by two independent reviewers from the full texts. The two reviewers addressed disagreements through discussion or via consultation with a third reviewer.

The Cochrane Collaboration Risk of Bias Tool in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins et al., 2022) was used by two independent researchers to evaluate the methodological quality. According to the Cochrane Handbook, the risk of bias assessment was divided into seven domains: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias. Each domain in the included RCTs was marked according to a low risk of bias, high risk of bias, or an unclear risk of bias. Disagreement between the two researchers was arbitrated by a third researcher.

The RevMan 5.3 software was used for meta-analysis. The relative risk (RR) was used as the effect index for the dichotomous variables, and the mean differences (MD) or standardized mean difference (SMD) were used as the effect index for continuous variables. The confidence interval (CI) of each effect index was set to 95%. The I ² statistic was adopted to assess the heterogeneity. If I ² > 50%, there was heterogeneity between the studies, and the random-effect model was selected; otherwise, the fixed-effect model was utilized. The heterogeneity was explained by sensitivity analysis or subgroup analysis. In addition, descriptive analysis was performed if the clinical data provided by the included studies were incomplete and could not be systematically evaluated. Potential publication bias was evaluated through a funnel plot.

The PRISMA flowchart of the literature screening process is presented in Figure 1. Initially, according to the search strategy, a total of 179 studies were obtained through retrieval from multiple databases; after the deletion of 85 duplicate publications, the remaining 94 articles were screened. After examination of the titles and abstracts, 58 irrelevant studies were removed. The full texts of the 36 remaining articles were assessed for eligibility. Preclinical studies (n = 10), non-RCTs (n = 3), and studies that did not meet the inclusion or met the exclusion criteria (n = 12) were excluded. Finally, 11 studies were included in the meta-analysis (Zou et al., 2018; Wang et al., 2020; Dong et al., 2021; Li et al., 2021; Zhang et al., 2021; Jiang et al., 2022; Kang, 2022; Zhang, 2022; Zhang et al., 2022; Zhou, 2022; Yan et al., 2023).

A total of 1,337 patients with AIS were included in the meta-analysis, including 668 patients in the trial group and 669 patients in the control group. All included studies were conducted in China. The sample size in each included study ranged from 21 to 193. The shortest treatment duration of an Ani injection was 7 days, and the longest was 30 days. The time period from AIS symptom onset to hospital admission was ≤72 h in all studies. Regarding the outcome measurements, the NIHSS was adopted in all studies (Zou et al., 2018; Wang et al., 2020; Dong et al., 2021; Li et al., 2021; Zhang et al., 2021; Jiang et al., 2022; Kang, 2022; Zhang, 2022; Zhang et al., 2022; Zhou, 2022; Yan et al., 2023), five studies used the mRS (Wang et al., 2020; Dong et al., 2021; Kang, 2022; Zhang, 2022; Zhang et al., 2022), four studies used the BI (Zou et al., 2018; Jiang et al., 2022; Zhang et al., 2022; Zhou, 2022), three studies reported the CTP (Zou et al., 2018; Zhang et al., 2022; Zhou, 2022), two studies mentioned the effective rate (Kang, 2022; Yan et al., 2023), and adverse events were described in seven studies (Zou et al., 2018; Dong et al., 2021; Li et al., 2021; Zhang et al., 2021; Jiang et al., 2022; Zhang, 2022; Zhang et al., 2022), while three studies reported the adverse rate (Dong et al., 2021; Li et al., 2021; Zhang et al., 2021). The characteristics of the included studies are presented in Table 1.

The assessment of bias risk of each eligible study was performed according to the Cochrane bias risk tool. Nine of the included studies mentioned grouping by a random method, three of which specified that a random sequence was generated through the random number table method (Dong et al., 2021; Li et al., 2021; Zhou, 2022). None of these studies referred to information on allocation concealment, blinding of participants and personnel, or blinding of the outcome assessment; therefore, all the studies were rated as having an unclear risk of bias in these three sections. All other bias evaluation risks were unclear. The results of the risk of bias assessment are presented in Figures 2, 3. See the supplementary document of Supplementary Material for rating bias (Supplementary Table S1).

A funnel plot was conducted to assess the publication bias of 10 trials or more. Thus, the 10 included studies that included available NIHSS score data were used for publication bias assessment, as shown in Figure 13. The shape of the funnel plot of the NIHSS showed the moderate symmetry between the included studies, which indicated that the potential of publication bias was low.

A total of 11 RCTs were included in this meta-analysis. Combined with conventional therapy, Ani injection was used to treat 1,337 patients with AIS. The NIHSS, mRS, BI, CTP, effective rate, and adverse events were evaluated in the analysis. According to the findings, the NIHSS score of Ani injection therapy was much lower than that of conventional therapy alone. Other primary indicators showed that Ani injection significantly reduced the mRS score and increased the BI score. The secondary outcome indicators revealed that treatment with Ani injection increased the rCBV, reduced the rTTP, and improved the clinical efficacy, with significant differences observed. The pooled analysis of the included studies failed to identify a significant change in the rCBF, rMTT, and rate of adverse reactions.

Subgroup analyses indicated that at the different time periods of 7 days, 14 days, and ≥30 days for implementing the NIHSS assessment, the NIHSS score of the Ani treatment group was considerably decreased. Furthermore, the subgrouping factor might be a source of significant heterogeneity for the NIHSS. Subgroup analyses on the BI score based on the BI assessment method (original BI and modified BI) suggest that regardless of the BI assessment method used, Ani treatment exhibits an advantage in significantly increasing the BI score. However, the subgroup analysis of BI did not identify the source of heterogeneity; significant heterogeneity may be associated with factors such as small sample sizes and few included studies.

A sensitivity analysis of rCBF and rMTT suggested that the pooled results are not robust and that the study of Zhang et al. (2022) might be the source of rCBF and rMTT heterogeneity.

In the risk of bias section, the quality assessment of the current included studies showed that allocation concealment, blinding of participants and employees, and blinding of the outcome assessment, as well as other forms of bias, were not disclosed in any of the included studies, which suggested that the certainty of evidence in the included RCTs was not high. Consequently, the results of the meta-analysis may be influenced, and our findings based on the current evidence should be considered carefully in the clinic. More precise RCTs are needed to further validate the curative effect of Ani injection in patients with AIS.

Traditional Chinese herbal medicines have a long history of clinical application in treating various vascular diseases, with distinctive theories and rich practices (Hung et al., 2015; Hao et al., 2017). Products from traditional Chinese medicinal herbs have been widely described in various ancient medicine systems for treating ischemic stroke, myocardial infarction, and so on (Hung et al., 2015). Anisodine is one of the most important ingredients of the tropane-type alkaloids extracted from the traditional folk medicinal herb A. tanguticus, with significant biological activities for promoting blood circulation and removing blood stasis (Meng et al., 2023). Pharmaceutical products containing anisodine are frequently used in the clinic for the treatment of vascular diseases, including ischemic stroke (Zou et al., 2018), retinal artery occlusion (Wu et al., 2016), ischemic optic neuropathy (Zhang et al., 2019), and cerebral small vessel disease (Gui et al., 2019). Ani injection has been developed for improved chemical stability and is a promising treatment for AIS.

Poor perfusion of brain tissue caused by the abrupt interruption or reduction of cerebral blood flow is the etiology of AIS, which can then induce ischemic hypoxic necrosis, clinically manifesting as different degrees of neurological impairment (Brott and Bogousslavsky, 2000). The molecular mechanism of AIS can be summarized as a complex series of ischemic cascades, characterized by cellular bioenergetic failure, excitotoxicity, excessive intraneuronal accumulation of Na⁺, Cl⁻, and Ca²⁺, oxidative damage, inflammatory reaction, mitochondrial injury, and, finally, cell death (Rosenblum, 1997; Brouns and De Deyn, 2009). Guidelines for the management of AIS have been reported by various countries (Swain et al., 2008; Di et al., 2019; Powers et al., 2019); the fundamental goals of the intervention have been focused on restoring or increasing the blood supply to the brain and blockading or slowing of the cerebral ischemic cascade (Olsen et al., 1983; Escuret, 1995; Brott and Bogousslavsky, 2000). The conventional therapy adopted in the 11 included studies varied across different care settings, including general management (such as respiratory and oxygen intake, cardiac monitoring and cardiac disease management, temperature control, blood pressure control, blood sugar control, and nutritional support, etc.) and specific treatment (thrombolysis, antiplatelet drugs, anticoagulants, statins, defibrase, and diuretics, etc.). Nevertheless, the narrow treatment window and hemorrhagic complications have limited the utilization and therapeutic effect of conventional therapy.

Recent studies have revealed that in addition to the recanalization of the large cerebral vessels, the restoration of normal vasomotor function around the ischemic area, the improvement in micro-perfusion, and neuronal cell protection are crucial for the treatment of cerebral infarction and are closely related to the prognosis of AIS (Tuttolomondo et al., 2009; Shuaib et al., 2011; Bang et al., 2015). As a central muscarinic cholinergic antagonist, Ani can effectively relieve vasospasm, open closed arterioles and the anterior capillary sphincter, and restore the perfusion of brain tissue (Wang et al., 2018). Research has demonstrated that following Ani administration, the microvascular autonomic motion reappears in the small intestinal wall micro-artery ischemia model, with a significant increase in microvascular amplitude, blood velocity, and flow (Zhang et al., 2019). Through the establishment of a hypoxia/reoxygenation (H/R)-induced brain microvascular endothelial cell injury model, Ani injection has been shown to suppress H/R-induced hypoxia-inducible transcription factor 1(HIF-α) over-expression, nitric oxide (NO), and reactive oxygen species (ROS) production, and all these effects were dependent on M4-AchR (Zeng et al., 2021). Additionally, Ani has multiple non-cholinergic effects, including cell protective effects, autophagy (Chen et al., 2017b), attenuating neuronal cell death and apoptosis (Chen et al., 2017d), alleviating oxidative stress damage and decreasing Ca²⁺ accumulation (Chen et al., 2017; Wang et al., 2017; Chen et al., 2017e), and inhibiting membrane lipid peroxidation (Zhao and Chen, 2010), thereby alleviating cell damage caused by ischemia and hypoxia. Further studies have found that Ani can decrease the Longa rodent stroke scores and cerebral infarction area in middle cerebral artery occlusion (MCAO) rats (Chen et al., 2017c). Moreover, the underlying mechanism of the effect of Ani on AIS can also be attributed to the ability to improve hemorheology and resist platelet aggregation so as to improve cerebral microcirculation disorders (Xu et al., 2020).

In short, based on current evidence, Ani injection therapy was found to be effective and safe in patients with AIS. The positive effect of Ani injection may be attributed to the ability of Ani to penetrate the blood–brain barrier and act as a non-specific muscarinic cholinergic receptor antagonist, competing with acetylcholine in the central nervous system, resulting in increased cerebral blood supply and neuroprotection effects (Liu et al., 2020; Jiang et al., 2022). The results of this meta-analysis demonstrated that the efficacy of Ani injection in the treatment of AIS was superior to that of conventional therapy; however, several limitations still exist. First, the sample size of the included RCTs was small, so the subgroup analysis and publication bias assessment could not be conducted for all indicators, which may have affected the accuracy and reliability of the results. Second, all trials lacked a precise description of the allocation concealment and blinding methods (for participants, personnel, and outcome assessments); the description of random sequence generation was also missed in some studies. As a consequence, the general methodological quality of the studies was not satisfactory. Third, the longest intervention duration in all the articles was 90 days, and there was a lack of long-term follow-up visits for more than 90 days after treatment, which was insufficient to assess the long-term impact of Ani therapy on the health of patients. Additionally, all the included RCTs were conducted in China; therefore, it is necessary to utilize multi-regional clinical trials for Ani treatment evaluation in different regions of the world in the future to provide strong evidence for the efficacy and safety of Ani treatment in patients with AIS.