Stroke remains a severe health problem that causes death and long-term disability worldwide, resulting in increased economic and social burden (Herpich and Rincon, 2020). More importantly, such negative impact of stroke will increase as people age across the globe (Murray and Lopez, 2013). Stroke is usually classified into two types: ischemic and hemorrhagic. Ischemic stroke (IS) is the most prevalent one, accounting for approximately 70%−85% of all stroke cases globally (Pluta et al., 2021; DeLong et al., 2022). At present, effective treatments targeting IS, such as thrombolytic therapy (Li et al., 2021; Zhao et al., 2022), thrombectomy (Winkelmeier et al., 2023), neuroprotective agents (Paul and Candelario-Jalil, 2021), and early rehabilitation (Gittler and Davis, 2018; Geng et al., 2022), seem to be related to a better functional outcome.

Recently, increasing evidence indicated that the outcome of IS can also be influenced by gut microbiota (GM). For example, animal model-based studies found that altering the gut microbiome of aged mice after experimental stroke through transplanting youthful microbiota can reverse the poor recovery in aged stroke mice (Spychala et al., 2018; Lee et al., 2020). Another animal study also proved that the cerebral infarct size and post-stroke outcomes were impacted by transplantation of GM, which was associated with the trimethylamine-N-oxide (TMAO) pathway (Zhu et al., 2021). Furthermore, some studies reported the relationship between the composition of GM and the prognosis of human stroke. An observational study reported that Christensenellaceae_R-7_group and norank_f_Ruminococcaceae were positively correlated with the modified Rankin scale (mRS) at 1 month, which was used to evaluate the stroke outcome, while genus Enterobacter was negatively correlated with the mRS (Li et al., 2019). Another case–control study which defined the mRS score of ≥ 3 at 3 months as a poor functional outcome also analyzed the differences of microbiota composition between the outcome groups after stroke, and the results showed that the group with poor outcomes had a higher abundance of Ruminococcaceae and Prevotella and a lower abundance of Anaerococcus, Blautia, Dialister, Aerococcaceae, Propionibacterium, Microbacteriaceae, and Rothia compared with the group with good outcomes (Chang et al., 2021). Although the above two research studies confirmed the association between GM and IS functional outcomes, the results did not seem to be entirely consistent, maybe because of the non-uniform functional outcome indicators. In addition, it is uncertain whether these associations are causal, given that the evidence obtained mainly from observational studies may result in potential bias in results due to reverse causation and residual confounding.

Mendelian randomization (MR) is an analytic method based on the summary data of genome-wide association study (GWAS) to establish the causality between exposure and outcome (Sleiman and Grant, 2010; Sekula et al., 2016). Given that genetic variants, such as single-nucleotide polymorphisms (SNPs), are used as proxies for the modifiable environmental factors (exposure) under investigation, the MR method has the advantage of reducing potential bias from confounding factors and reverse causation because of the random assignment of alleles during human gamete formation and the perpetually immutable genotype determined before birth (Sekula et al., 2016). Consequently, with the available GWAS data over the last decade, MR studies have been widely used to analyze the causal association between GM and diseases (Smith and Ebrahim, 2003; Wang et al., 2018; Kurilshikov et al., 2021).

A newly published report has used MR to reveal the causal effect of GM on the risk of IS subtypes (Meng et al., 2023). However, to date, MR studies about the causal inference between GM and IS prognosis remain unavailable. Furthermore, conducting large-scale longitudinal cohort studies or randomized controlled trials (RCTs) currently is unfeasible. In this context, we performed this MR study using the genetic variation associated with GM to assess their possible causal relationship with IS functional outcomes.

To our knowledge, this is the first MR analysis, overcoming environmental confounding and reverse causation, to investigate the causal effect of genetically determined gut microbiota on functional outcomes after ischemic stroke. Evidence from this study supported a causal association between the abundance of specific bacterial traits and the prognosis of IS. However, positive findings in this study were mainly at the genus level, and no causal associations between GM and IS prognosis at the level of the phylum, class, and order were found. Strikingly, genus Ruminococcaceae UCG005 and the genus Eubacterium oxidoreducens group showed significantly negative effects on stroke prognosis, while genus Lachnospiraceae NK4A136 group and genus Lachnospiraceae UCG004 showed protective effects against stroke prognosis in both main and comparative IVW analysis. Additionally, although the results of the main and comparative weighted median analyses were inconsistent in terms of protection factors, genus Ruminococcaceae UCG005 and genus Eubacterium oxidoreducens group as risk factors for poor prognosis of IS were indisputable. We also performed a reverse MR analysis of these positive taxa, and the results did not support a causal effect of post-stroke prognosis on these taxa. These findings may provide important implications for the discovery of novel biomarkers in future IS experiments and provide prevention and therapeutic strategies targeting dysbiosis of specific GM taxa.

Previous research found that Ruminococcaceae UCG005 increased in severe stroke patients who usually have a worse stroke outcome (Li et al., 2019). This is consistent with our research. A network analysis suggested that Ruminococcaceae UCG005 was one of the genera driving the progress of the type 2 diabetes in a Mexican cohort (Esquivel-Hernandez et al., 2023). An animal study reported a significantly negative correlation of Ruminococcaceae UCG005 with high-density lipoprotein cholesterol (HDL-C) and a positive correlation with body weight (Qin et al., 2022). It is well known that abnormal glucose and lipid metabolism are the dominant causes of cerebrovascular diseases. Genus Ruminococcaceae UCG005 may influence the stroke prognosis by the pathway of glucose and lipid metabolism. Furthermore, ample evidence shows that imbalance in GM contributes to the neuroinflammation and worse stroke outcomes (Huang et al., 2023; Park et al., 2023). An animal-based study found that maternal sleep deprivation (MSD) caused high expression of pro-inflammatory cytokines in offspring rats Moreover, pro-inflammatory cytokines were positively associated with Ruminococcaceae UCG005 (Yao et al., 2022). Therefore, it is speculated that Ruminococcaceae UCG005 may also be involved in the inflammatory response after stroke, thereby adversely affecting the prognosis of stroke. These results may support the conclusion of Ruminococcaceae UCG005 as a risk factor for stroke prognosis.

For the Eubacterium oxidoreducens group, we failed to find some direct evidence of its role in stroke prognosis from existing research. Research on the role of the Eubacterium oxidoreducens group in other diseases is also rare. However, a previous study reported a significantly elevated relative abundance of Eubacterium oxidoreducens group which was positively correlated with the levels of serum and fecal lipopolysaccharide (LPS) in high-fat diet (HFD)-fed mice (Zhang X. Y. et al., 2020). Increased LPS contents could impair the intestinal epithelial barrier (IEB) (Guo et al., 2015) and blood–brain barrier (BBB) (Peng et al., 2021). Recent studies have revealed that gut barrier integrity and BBB are involved in the influence of GM on IS (Gwak and Chang, 2021; Zeng et al., 2023). A new study found that intraperitoneal injection of LPS after stroke exhibited intestinal morphology damage, decreased expression of tight-junction proteins associated with the BBB, and more neuronal loss, and these changes were consistent with stroke mice transplanted with gut microbiota associated with post-stroke cognitive impairment (Wang et al., 2022). Therefore, we speculate that the Eubacterium oxidoreducens group, as a risk factor, may have adverse effects on stroke prognosis by affecting intestinal epithelial integrity and blood–brain barrier.

The results of another MR study evaluating the causal effect of GM on cardioembolic IS support the protective effect of the genus Lachnospiraceae NK4A136 group on IS prognosis in this MR analysis (Dai et al., 2024). A previous study reported a significant negative association of the genus Lachnospiraceae NK4A136 group with intestinal permeability and the plasma LPS level (Ma et al., 2020), which means that genus Lachnospiraceae NK4A136 group is beneficial for protecting the intestinal barrier. The intestinal barrier acts as the first barrier to prevent harmful substances from penetrating the intestinal mucosa and damaging other tissues of the body. The disruption of IEB after stroke contributes to the microbial translocation which will increase the risk of post-stroke infections (Zhao et al., 2023). Therefore, the protective effect of the genus Lachnospiraceae NK4A136 group on IEB may be helpful in IS prognosis.

High abundance of genus Lachnospiraceae UCG004 was also identified useful for IS prognosis. As probiotics in the body, it is beneficial for reducing obesity (Xu et al., 2024). Evidence from a clinical study also showed a negative correlation between Lachnospiraceae UCG004 and cardiovascular disease risk factors (Tindall et al., 2020). Another case–control study found a low abundance of Lachnospiraceae UCG004 in lacunar cerebral infarction patients compared with healthy controls (Ma et al., 2023). It is worth noting that both the genus Lachnospiraceae NK4A136 group and Lachnospiraceae UCG004 belong to family Lachnospiraceae which is recognized as short chain fatty acid (SCFA)-producer. In most studies, increased inflammation response and decreased SCFAs could be observed in stroke individuals and were significantly related to poor IS outcomes (Spychala et al., 2018; Tan et al., 2021). Animal-based studies proved that transplantation of SCFAs-rich gut microbiota or SCFA supplementation in drinking water can effectively promote the recovery following ischemic stroke (Chen et al., 2019; Lee et al., 2020; Sadler et al., 2020). Studies showed that SCFAs participate in the regulatory process of inflammatory responses, including promoting the anti-inflammatory cytokines and suppressing the pro-inflammatory cytokines (Maslowski et al., 2009; Vinolo et al., 2011). This regulatory effect of SCFAs on inflammation may be one of the mechanisms by which GM affects the prognosis of stroke (Iadecola and Anrather, 2011; Chamorro et al., 2012).

The underlying mechanisms involved in the influence of GM on outcomes following IS are multifaceted. In addition to the above-mentioned content, neurotransmitters and TMAO pathway (Zhu et al., 2021), among others are involved in the regulation of GM on stroke. In conclusion, more research is needed to confirm whether and how the four taxa identified in this study are involved in these mechanisms.

To date, most studies evaluating the role of gut microbiome in stroke have focused on bacteria due to the overwhelming abundance of bacteria. Research on viruses, fungi, and archaea is scarce. However, these non-bacteria gut microbes play an important role in human health and diseases (Goralska et al., 2018; Mukhopadhya et al., 2019; Coker, 2022; Ezzatpour et al., 2023). There is a growing body of evidence that suggests non-bacteria gut microbes are associated with neurological diseases, including stroke (Forbes et al., 2018). A new study found that fecal viral taxa was altered significantly after stroke and dissimilar phage protein networks in mice (Chelluboina et al., 2022). Phages are regarded as the most identified human gut virus components with significant roles. Phages can infect diverse bacterial phyla in the gut, such as Firmicutes, Bacteroidetes, Proteobacteria, and Actinobacteria (Mirzaei and Maurice, 2017). When healthy people took phage dietary supplements orally, there was an increase in the butyrate-producing genus, Eubacterium (Febvre et al., 2019). In addition to the viruses, some fungi housed in the gastrointestinal tract are considered to be pathogenic and can destroy CNS astrocytes, leading to BBB disruption and central infection. Colonizing the gut by archaea has also been demonstrated to decrease levels of trimethylamine (Forbes et al., 2018), a compound produced by intestinal bacteria that is linked to an increased risk of atherosclerosis as well as cardiovascular and cerebrovascular diseases. Hence, the role of non-bacterial microorganisms in stroke is also one of the important directions of future stroke research. Unfortunately, there is no GWAS data for non-bacterial microorganisms, and the causal relationship between stroke and stroke prognosis cannot be further explored in this study. Nevertheless, we are eagerly looking forward to the improvement of the GWAS data that will enable us to delve deeper into this fascinating area of research in the future.

The MR analysis we designed is unlikely to be biased because of confounding factors and reverse causality. We also drew a relatively consistent and reliable conclusion from the data obtained from two sets of outcomes which were with or without adjustment for the baseline NIHSS. Although we have tried our best to make the conclusions accurate and robust, there are still some limiting factors that need to be considered in our research. First, the study's limitation in not being able to mine data below the genus level indeed restricts the depth of understanding of the potential links between the gut microbiota and outcomes of ischemic stroke. Second, a test about the associations between GM and post-stroke outcomes for different stroke subtypes was not performed because of the lack of available GWAS data in the GISCOME database. Different types, localization of stroke have a decisive influence on recovery in survivors (Biffi et al., 2015; Zhang K. et al., 2020), which may contribute to some bias in our results. Third, we did not adequately control for age and sex among the patients included in the study. Thus, age and sex need to be taken into account in further investigations on the relationship between GM and post-stroke outcomes because the formation and shaping of the GM are easily influenced by sex and lifespan (Snigdha et al., 2022). Fourth, a previous MR analysis has indicated that genetically determined GM was associated with the onset of ischemic stroke (Meng et al., 2023), which made the collider bias a non-negligible issue in this MR investigation of IS prognosis (Paternoster et al., 2017). Fifth, a small sample size of subjects (n = 6,021) included in this study were only of European ancestry, which may restrict the generalizability of our findings to other populations. Thus, future studies with a larger sample size of other ancestries are needed to explore the associations between GM and ischemic stroke outcomes.

In summary, this MR analysis demonstrated that genetically predicted gut microbiota is causally associated with worse post-stroke outcomes. Our results suggest that interventions addressing particular GM taxa, such as Lachnospiraceae NK4A136, Lachnospiraceae UCG004, Ruminococcaceae UCG005, and Eubacterium oxidoreducens groups, may provide new opportunities to improve recovery after ischemic stroke.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

AZ: Conceptualization, Writing – review & editing. PL: Software, Writing – review & editing. YC: Software, Writing – review & editing. XW: Data curation, Writing – review & editing. JZ: Data curation, Writing – review & editing. LL: Methodology, Writing – review & editing. TZ: Conceptualization, Writing – original draft. JY: Funding acquisition, Writing – original draft.