Predicting Adverse Recanalization Therapy Outcomes in Acute Ischemic Stroke Patients Using Characteristic Gut Microbiota

Recanalization therapy is the most effective treatment for eligible patients with acute ischemic stroke (AIS). Gut microbiota are involved in the pathological mechanisms and outcomes of AIS. However, the association of gut microbiota features with adverse recanalization therapy outcomes remains unclear. Herein, we investigated gut microbiota features associated with neurological deficits in patients with AIS after recanalization therapy and whether they predict the patients’ functional outcomes. We collected fecal samples from 51 patients with AIS who received recanalization therapy and performed 16S rRNA gene sequencing (V3–V4). We compared the gut microbiota diversity and community composition between mild to moderate and severe disability groups. Next, the characteristic gut microbiota was compared between groups, and we noted that the characteristic gut microbiota in patients with mild to moderate disability included Bilophila, Butyricimonas, Oscillospiraceae_UCG-003, and Megamonas. Moreover, the relative abundance of Bacteroides fragilis, Fusobacterium sp., and Parabacteroides gordonii was high in patients with severe disability. The characteristic gut microbiota was correlated with neurological deficits, and areas under the receiver operating characteristic curves confirmed that the characteristic microbiota predicted adverse recanalization therapy outcomes. In conclusion, gut microbiota characteristics are correlated with recanalization therapy outcomes in patients with AIS. Gut microbiota may thus be a promising biomarker associated with early neurological deficits and predict recanalization therapy outcomes.


Introduction
Stroke is a brain injury caused by a disruption in the blood supply to a specific region of the brain, resulting in permanent neurological deficits or even death. Stroke is not only This was a single-center, prospective cohort study, and it was conducted at a hospital in the southern part of Taiwan. As indicated in Figure 1, patients who were admitted to the hospital within 6 h of symptom onset and given a diagnosis of AIS were enrolled. We included only patients who received recanalization therapy (i.e., IVT, EVT, or both) as well as brain magnetic resonance imaging to detect acute ischemic lesions. We excluded patients who were younger than 20 years, had intracerebral hemorrhage on initial radiological examination, had contraindications to IVT or EVT, had received probiotics or antibiotics within 1 week prior to admission, or had received antibiotic therapy prior to fecal sample collection after admission. Each patient was comprehensively evaluated for demographics, medical history, physical and neurological examinations, and blood biochemistry analysis.

Patients
This was a single-center, prospective cohort study, and it was conducted at a hospital in the southern part of Taiwan. As indicated in Figure 1, patients who were admitted to the hospital within 6 h of symptom onset and given a diagnosis of AIS were enrolled. We included only patients who received recanalization therapy (i.e., IVT, EVT, or both) as well as brain magnetic resonance imaging to detect acute ischemic lesions. We excluded patients who were younger than 20 years, had intracerebral hemorrhage on initial radiological examination, had contraindications to IVT or EVT, had received probiotics or antibiotics within 1 week prior to admission, or had received antibiotic therapy prior to fecal sample collection after admission. Each patient was comprehensively evaluated for demographics, medical history, physical and neurological examinations, and blood biochemistry analysis.

Stroke Severity, Functional Outcomes, and Reperfusion Assessment
Stroke severity was evaluated on the basis of each patient's National Institutes of Health Stroke Scale (NIHSS) score, whereas functional outcomes were assessed using each patient's modified Rankin Scale (mRS) score. A neurologist blinded to the results of the gut microbiota analyses assessed both the NIHSS and mRS scores. The NIHSS score (range of 0-42) indicates the degree of a patient's neurological impairment; the higher the NIHSS score, the more severe the patient's neurological deficit. The NIHSS score was obtained

Stroke Severity, Functional Outcomes, and Reperfusion Assessment
Stroke severity was evaluated on the basis of each patient's National Institutes of Health Stroke Scale (NIHSS) score, whereas functional outcomes were assessed using each patient's modified Rankin Scale (mRS) score. A neurologist blinded to the results of the gut microbiota analyses assessed both the NIHSS and mRS scores. The NIHSS score (range of 0-42) indicates the degree of a patient's neurological impairment; the higher the NIHSS score, the more severe the patient's neurological deficit. The NIHSS score was obtained before treatment and at discharge. The mRS score was assessed at baseline and at discharge; an mRS score of 0-3 at discharge was considered to indicate mild to moderate disability. Post-EVT reperfusion status was assessed using the modified treatment in cerebral ischemia (mTICI) score. The mTICI score was determined by the EVT operator based on each patient's final angiogram.

Fecal Sample Collection, Bacterial DNA Extraction, and 16S rRNA Gene Sequencing
Fecal samples were collected from each participant before their first meal during hospitalization. The samples were frozen immediately after collection and delivered to the laboratory in a cooler bag within 24 h. The fecal samples were stored at −80 • C for up to 3 days prior to processing.
Bacterial deoxyribonucleic acid (DNA) was extracted from the fecal samples using a stool DNA extraction kit (Topgen Biotechnology, Kaohsiung, Taiwan). After its quality and concentration were assessed on a Colibri Microvolume spectrophotometer (Titertek Berthold, Pforzheim, Germany), the extracted DNA was immediately frozen at −20 • C.
We outsourced our DNA samples to Welgene Biotech (Taipei, Taiwan) for 16S rRNA gene sequencing. Each bacterial DNA sample was subjected to 16S rDNA amplicon sequencing using Illumina Sequencing-by-Synthesis technology on an Illumina MiSeq sequencer to produce 2 × 300 bp paired-end reads. The primers for the 16S rRNA gene (V3-V4 region) were as follows: forward, TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGC-CTACGGGNGGCWGCAG; and reverse, GTCTCGTGGGCTCGGAGATGT GTATAAGA-GACAGGACTACHVGGGTATCTAATCC.

Statistical and Bioinformatic Analyses of Microbiota
The patients' demographic, medical history, and blood biochemistry data were classified as either categorical or continuous variables. In order to compare the data between patients with mild to moderate disability (mRS score = 0-3) and severe disability (mRS score = 4-6), we utilized the two-tailed independent t-test to analyze continuous variables, while categorical variables were assessed using the chi-square test. The NIHSS scores, type of recanalization therapy, onset to treatment time, presence of good angiographic reperfusion (mTICI score ≥ 2b), and occluded artery were also compared between the two groups.
The raw sequencing data were imported into QIIME2 [19] and processed using the DADA2 plugin [20] to merge and denoise paired-end reads into amplicon sequence variants (ASVs). The median number (interquartile range (IQR)) of reads filtered through each quality control step was 105,487 (95,432, 117,353). To avoid false conclusions due to an uneven sampling depth in the microbiome diversity assessment, we standardized the sampling depth of each sample by rarefying it to 66,649 reads, which corresponded to the lowest number of reads detected in all samples and the point at which the rarefaction curves of both groups leveled off.
We compared the alpha diversity indexes using pairwise Kruskal-Wallis tests. To assess beta diversity, we performed a pairwise analysis of similarities (ANOSIM) and permutational multivariate analyses of variance (PERMANOVA) with 999 permutations as well as a principal coordinate analysis (PCoA) based on various distance matrixes. All p values were adjusted using the Benjamini-Hochberg procedure (to obtain q values).
The ASV taxonomy was classified using a SciKit Learn-based approach and by searching in the SILVA reference database (version 138; trimmed to the V3-V4 region; L7 taxonomy) [21]. We analyzed the relative abundance of taxa using linear discriminant analysis (LDA) effect size (LEfSe) [22]. Next, we identified differential taxa features between the groups, which were identified on the basis of a log LDA score for discriminative features of >2 and p < 0.05 in the factorial Kruskal-Wallis test.
The Spearman correlation was used to analyze the associations between gut microbiota and stroke severity, as determined by NIHSS. To explore the value of gut microbiota features in the prediction of post-recanalization therapy outcomes, we plotted receiver operating characteristic (ROC) curves based on the relative abundance of the bacteria in the two patient groups and calculated the areas under ROC curves (AUCs).

Ethics Approval
This study was conducted in accordance with the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board of Kaohsiung Medical University Hospital [KMUHIRB-E(I)-20200424]. Informed consent was obtained from all patients involved in this study or their legal representatives.
Compared with the severe disability group, the mild to moderate disability group had a significantly lower atrial fibrillation or flutter prevalence and a significantly higher rate of only IVT use. However, the between-group differences in the median time from stroke onset to IVT, groin puncture, or reperfusion and in the rates of successful reperfusion (mTICI score ≥ 2b) were nonsignificant. Most (84.3%) of the occluded vessels were in the anterior circulation; the proportions of the patients in each group with such vessels were similar. Table 1 presents an overview of the patient characteristics, vascular risk factor, and administration details of the recanalization therapy.

Characterization of Gut Microbiota Based on 16S rRNA Gene Sequencing
As presented in Figure 2, intragroup diversity tended to be lower in the severe disability group than in the mild to moderate disability group, as indicated by the Chao1 index (586.69 ± 101.67 vs. 534.00 ± 114.07, p = 0.091) and Shannon's index (5.9 ± 0.6 vs. 5.7 ± 0.7, p = 0.391) values; however, these differences were nonsignificant. Regarding the intergroup diversity, PCoA based on Bray-Curtis dissimilarity (ANOSIM: R = −0.02, p = 0.74; PERMANOVA: pseudo-F = 1.00, p = 0.46) and weighted unnormalized UniFrac (ANOSIM: R = −0.03, p = 0.90; PERMANOVA: pseudo-F = 0.59, p = 0.87) demonstrated an absence of significant microbial clustering differences between the mild to moderate and severe disability groups. These results reveal no significant differences in the intragroup and intergroup diversity between our mild to moderate and severe disability groups.

Characterization of Gut Microbiota Based on 16S rRNA Gene Sequencing
As presented in Figure 2, intragroup diversity tended to be lower in the severe disability group than in the mild to moderate disability group, as indicated by the Chao1 index (586.69 ± 101.67 vs. 534.00 ± 114.07, p = 0.091) and Shannon's index (5.9 ± 0.6 vs. 5.7 ± 0.7, p = 0.391) values; however, these differences were nonsignificant. Regarding the intergroup diversity, PCoA based on Bray-Curtis dissimilarity (ANOSIM: R = −0.02, p = 0.74; PER-MANOVA: pseudo-F = 1.00, p = 0.46) and weighted unnormalized UniFrac (ANOSIM: R = −0.03, p = 0.90; PERMANOVA: pseudo-F = 0.59, p = 0.87) demonstrated an absence of significant microbial clustering differences between the mild to moderate and severe disability groups. These results reveal no significant differences in the intragroup and intergroup diversity between our mild to moderate and severe disability groups.

Relative Abundance of Discriminative Taxa between Mild to Moderate Disability and Severe Disability Groups
Between-group differences in the relative abundance of gut microbiota were estimated using LEfSe based on a log LDA score of >2. The differences in the abundance of the phyla Firmicutes (49.69% in the mild to moderate disability group vs. 50.16% in the severe disability group), Bacteroidetes (29.72% in the mild to moderate disability group vs. 29.41% in the severe disability group), and the Firmicutes/Bacteroidetes ratio (2.68 ±

Relative Abundance of Discriminative Taxa between Mild to Moderate Disability and Severe Disability Groups
Between-group differences in the relative abundance of gut microbiota were estimated using LEfSe based on a log LDA score of >2. The differences in the abundance of the phyla Firmicutes (49.69% in the mild to moderate disability group vs. 50.16% in the severe disability group), Bacteroidetes (29.72% in the mild to moderate disability group vs. 29.41% in the severe disability group), and the Firmicutes/Bacteroidetes ratio (2.68 ± 3.43 in the mild to moderate disability group vs. 2.25 ± 1.68 in the severe disability group) were nonsignificant. Figure 3 presents the abundant taxa among the patient groups. The Oscillospiraceae_ UCG-003 and Megamonas, as well as its family Selenomonadaceae in the phyla Firmicutes, Butyricimonas and its family Marinifilaceae, Bacteroides fluxus and Alistipes shahii in the phyla Bacteroidetes, Bifidobacterium sp., and Bilophila were significantly enriched in the mild to moderate disability group. In contrast, Bacteroides fragilis and Parabacteroides gordonii in the phyla Bacteroidetes and Fusobacterium sp. were significantly enriched in the severe disability group. Random forest models were used for taxonomy prediction, and four genera could be used to discriminate the mild to moderate disability group from the severe disability group: Bilophila, Butyricimonas, Oscillospiraceae_UCG-003, and Megamonas. in the phyla Bacteroidetes, Bifidobacterium sp., and Bilophila were significantly enriched in the mild to moderate disability group. In contrast, Bacteroides fragilis and Parabacteroides gordonii in the phyla Bacteroidetes and Fusobacterium sp. were significantly enriched in the severe disability group. Random forest models were used for taxonomy prediction, and four genera could be used to discriminate the mild to moderate disability group from the severe disability group: Bilophila, Butyricimonas, Oscillospiraceae_UCG-003, and Megamonas.

Analysis of Association between Gut Microbiota, NIHSS Scores, and Functional Outcomes
We selected Bilophila, Butyricimonas, Oscillospiraceae_UCG-003, Megamonas, Bacteroides fragilis, Fusobacterium sp., and Parabacteroides gordonii for further analysis of the associations among gut microbiota, stroke severity, and functional outcomes based on LDA values and random forest models. As indicated in our Spearman correlation heatmap (Figure 4), the discharge NIHSS was correlated negatively with Bilophila (p = 0.004) and Megamonas (p = 0.011) but positively correlated with Bacteroides fragilis (p = 0.037). Moreover, Bilophila was correlated with neurological improvement, as indicated by NIHSS score changes between discharge and admission (p = 0.032).

Analysis of Association between Gut Microbiota, NIHSS Scores, and Functional Outcomes
We selected Bilophila, Butyricimonas, Oscillospiraceae_UCG-003, Megamonas, Bacteroides fragilis, Fusobacterium sp., and Parabacteroides gordonii for further analysis of the associations among gut microbiota, stroke severity, and functional outcomes based on LDA values and random forest models. As indicated in our Spearman correlation heatmap (Figure 4), the discharge NIHSS was correlated negatively with Bilophila (p = 0.004) and Megamonas (p = 0.011) but positively correlated with Bacteroides fragilis (p = 0.037). Moreover, Bilophila was correlated with neurological improvement, as indicated by NIHSS score changes between discharge and admission (p = 0.032).
Microorganisms 2023, 11, x FOR PEER REVIEW 8 of 13 Figure 4. Heatmap of the correlation of the differential bacteria with NIHSS scores at admission and discharge and NIHSS score changes during hospitalization. Green grids represent positive Spearman's rank correlation coefficients, whereas red grids represent negative Spearman's rank correlation coefficients. The deeper green or red indicate higher correlation values. * and ** denote p < 0.05 and p < 0.01, respectively.
We subsequently assessed the potential of using gut microbiota as a biomarker for predicting recanalization therapy outcomes. As presented in Figure 5, Bilophila and Butyricimonas have good predictive power for mild to moderate disability (AUCs = 0.713 and 0.741, respectively), and Bacteroides fragilis and Parabacteroides gordonii have a good predic- Figure 4. Heatmap of the correlation of the differential bacteria with NIHSS scores at admission and discharge and NIHSS score changes during hospitalization. Green grids represent positive Spearman's rank correlation coefficients, whereas red grids represent negative Spearman's rank correlation coefficients. The deeper green or red indicate higher correlation values. * and ** denote p < 0.05 and p < 0.01, respectively. We subsequently assessed the potential of using gut microbiota as a biomarker for predicting recanalization therapy outcomes. As presented in Figure 5, Bilophila and Butyricimonas have good predictive power for mild to moderate disability (AUCs = 0.713 and 0.741, respectively), and Bacteroides fragilis and Parabacteroides gordonii have a good predictive power for severe disability (AUCs = 0.712 and 0.679, respectively). Therefore, the identified bacteria could be potential biomarkers for recanalization therapy outcomes in patients with AIS. Figure 4. Heatmap of the correlation of the differential bacteria with NIHSS scores at admission and discharge and NIHSS score changes during hospitalization. Green grids represent positive Spearman's rank correlation coefficients, whereas red grids represent negative Spearman's rank correlation coefficients. The deeper green or red indicate higher correlation values. * and ** denote p < 0.05 and p < 0.01, respectively.
We subsequently assessed the potential of using gut microbiota as a biomarker for predicting recanalization therapy outcomes. As presented in Figure 5, Bilophila and Butyricimonas have good predictive power for mild to moderate disability (AUCs = 0.713 and 0.741, respectively), and Bacteroides fragilis and Parabacteroides gordonii have a good predictive power for severe disability (AUCs = 0.712 and 0.679, respectively). Therefore, the identified bacteria could be potential biomarkers for recanalization therapy outcomes in patients with AIS.

Discussion
To the best of our knowledge, this is the first study to delineate gut microbiota features and their associations with recanalization therapy outcomes in patients with AIS. Our results demonstrate that after recanalization therapy, gut microbiota composition differs between patients with mild to moderate and severe disability after AIS. We discovered that Bilophila, Butyricimonas, Oscillospiraceae_UCG-003, and Megamonas are enriched in patients with mild to moderate disability, whereas Bacteroides fragilis, Fusobacterium sp., and Parabacteroides gordonii are enriched in patients with severe disability. The richness of specific gut microbiota was noted to be correlated with neurological deficits post-recanalization therapy. Thus, Bilophila and Butyricimonas may predict mild to moderate disability,

Discussion
To the best of our knowledge, this is the first study to delineate gut microbiota features and their associations with recanalization therapy outcomes in patients with AIS. Our results demonstrate that after recanalization therapy, gut microbiota composition differs between patients with mild to moderate and severe disability after AIS. We discovered that Bilophila, Butyricimonas, Oscillospiraceae_UCG-003, and Megamonas are enriched in patients with mild to moderate disability, whereas Bacteroides fragilis, Fusobacterium sp., and Parabacteroides gordonii are enriched in patients with severe disability. The richness of specific gut microbiota was noted to be correlated with neurological deficits post-recanalization therapy. Thus, Bilophila and Butyricimonas may predict mild to moderate disability, whereas Bacteroides fragilis and Parabacteroides gordonii may predict severe disability. Taken together, these findings indicate that gut microbiota are ideal, noninvasive fecal biomarkers for the early prediction of neurological deficits and functional outcomes in patients with AIS after recanalization therapy.
Bilophila, which is enriched in patients with stroke [23] and acute coronary syndromes [24], is associated with the consumption of animal protein and a lack of plant-based protein sources [25]. However, the pathological mechanism underlying the association between Bilophila and AIS outcomes has not been established. The genera Butyricimonas and Megamonas and the family Oscillospiraceae can improve stroke outcomes through several mechanisms. Butyricimonas and Oscillospiraceae both produce butyrate [26][27][28], which has been noted to reduce neuronal apoptosis occurrence and cerebral infarction volume and to improve neurological function in animal stroke models [29,30]. In addition, Megamonas ferments glucose into short-chain fatty acids, mostly acetate and propionate [31], both of which are beneficial for stroke recovery. Reduced acetate and propionate levels were associated with an increased risk of poor functional outcomes in patients after stroke [16]. The results of an animal experiment demonstrated that supplementation with a mix of acetate, butyrate, and propionate improves poststroke recovery and cortical reorganization [32].
Butyricimonas can activate glucagon-like peptide-1 receptor and peroxisome proliferatoractivated receptor α, which can alleviate diabetes and metabolic disorders induced by a high-fat diet [33]. Oscillospiraceae was reported to be correlated with adiponectin in neurodegeneration disease [34], and an abundance of Oscillospiraceae is associated with decreased insulin resistance [35]. The abundance of Megamonas is higher in individuals with normal glucose tolerance than in those with type 2 diabetes mellitus [36]. In general, the presence of Butyricimonas, Oscillospiraceae, and Megamonas can help stabilize glucose metabolism and increase short-chain fatty acid levels, potentially leading to improved stroke recovery and decreased disability.
On the other hand, Bacteroides fragilis can biosynthesize and secrete pathogenic and proinflammatory neurotoxins, namely LPS and Bacteroides fragilis toxins [37]. Bacteroides fragilis negatively affects the biophysiological barrier structure and function, and thus disrupts the normal blood-brain barrier and elicits inflammatory neuronal dysfunction [38]. Additionally, Bacteroides fragilis deteriorates glucose and lipid metabolism, activates an inflammatory response, and promotes atherosclerosis progression in animal models [39].
Fusobacterium generates a proinflammatory microenvironment in the gut [40], induces immune cell death [41], alters vascular endothelial integrity, and passes through the bloodbrain barrier [42], eventually impairing stroke outcomes through the microbiota-gut-brain axis. Numerous studies have indicated that an increase in the number of Fusobacterium is associated with hypertension [43,44] and is positively correlated with homocysteine levels [45]-both of which are well-known risk factors for stroke. Therefore, increased Fusobacterium is considered to be strongly associated with unfavorable stroke outcomes [46].
Parabacteroides, a large artery atherosclerotic stroke biomarker [47], is more abundant in patients with ischemic stroke than in healthy individuals [15,48,49]. In addition, Parabacteroides is associated with vascular risk factors and stroke severity-as reflected by its positive correlation with infract volume and its negative correlation with poststroke daily function [47].
Taken together, our results indicate that in patients with AIS, a significant abundance of Bilophila, Butyricimonas, Oscillospiraceae, and Megamonas, which produce short-chain fatty acids and contribute to glucose homeostasis, possibly contributes to the beneficial effects of recanalization therapy. By contrast, Bacteroides fragilis, Fusobacterium, and Parabacteroides are associated with vascular risk factors, gut integrity disruption, blood-brain barrier impairment, and neuroinflammation induction, thereby increasing the likelihood of severe post-AIS disability.
There is a higher prevalence of atrial fibrillation/flutter observed in the severe disability group compared to the mild to moderate disability group among the demographics and medical history analyzed. Studies have shown that dysbiosis is linked to atrial fibrillation, possibly due to dietary habits, bacterial LPS, and microbial metabolites. These mediators are suggested to increase inflammation and contribute to atrial arrhythmogenesis, thereby affecting susceptibility to atrial fibrillation [50]. Furthermore, AIS patients with atrial fibrillation have reported unfavorable functional outcomes following IVT and EVT [51,52]. Hence, it is possible that in the severe disability group, gut microbiota may contribute to adverse functional recovery post-recanalization therapy via atrial fibrillation.
The novelty of the current study lies in its recruitment of patients with AIS who received recanalization therapy; thus far, this patient group has rarely been studied. This is because this population is small among AIS patients, as recanalization therapy is an urgent treatment strategy. We thus obtained a newer understanding of the use of gut microbiota as a prognostic biomarker of recanalization therapy outcomes in patients with AIS than other studies have.
Furthermore, targeting dysbiosis of the gut microbiota can potentially serve as a therapeutic intervention to alleviate poststroke neuroinflammation and to enhance stroke outcomes following recanalization therapy. An animal model revealed that the transplantation of healthy and SCFAs-producing microbiota notably improved stroke outcomes [11,29], while the modulation of the microbiota has been linked to a decrease in LPS and strokerelated neuroinflammation [53]. Additionally, a cerebral ischemia reperfusion model, similarly to recanalization therapy, demonstrated that microbiota from young mice may inhibit interleukin-17 production and lower reperfusion injury in aged mice [54].
In our view, although the majority of evidence comes from animal studies, microbiotatargeted therapy presents a promising potential for the treatment of AIS, particularly in patients undergoing recanalization therapy. It is conceivable that a therapy to modify the microbiota composition, such as dietary regulation, the administration of probiotics or prebiotics, and fecal microbiota transplantation, could be combined with recanalization therapy to mitigate the extent of reperfusion injury and neuroinflammation in the acute phase of AIS [55]. Thus, future clinical investigations are needed to explore the feasibility of targeting the gut microbiota as an innovative therapeutic approach that can improve functional outcomes in individuals with AIS.
The current study, however, has several limitations. First, no information regarding the patients' dietary habits and lifestyles was collected. Patients who had used probiotics or antibiotics in the week before AIS diagnosis were excluded, and the fecal samples were collected before their first meal after receiving an AIS diagnosis. Through this design, we minimized the effects of diet and antibiotics on the gut microbiota after AIS. Second, we collected fecal samples at a single timepoint; this limited our ability to assess dynamic changes in the association of gut microbiota with functional outcomes after recanalization therapy. Third, because of our limited sample size, we could not determine the association of characteristic microbiota with their metabolites, such as trimethylamine-N-oxide and short-chain fatty acids. Thus, we could not investigate the causal association between gut microbiota and functional outcomes after recanalization therapy. Finally, our follow-up period was short. We determined functional outcomes on the basis of the mRS scores at discharge, mostly within 30 days. Future larger-scale and longer-term studies addressing these limitations and investigating the influence of gut microbiota on recanalization therapy outcomes by assessing the effects of metabolic products and pathways are warranted.

Conclusions
Our results confirmed the associations between gut microbiota characteristics and recanalization therapy outcomes in patients with AIS. Gut microbiota could be a pertinent biomarker for predicting adverse recanalization therapy outcomes in patients with AIS. Given the growing preclinical evidence suggesting that modulation of the gut microbiota is a promising therapeutic target for AIS, the translation of these results into clinical practice may represent a major breakthrough in the treatment of AIS and merits further research.