Causal role of gut microbiota in immune thrombocytopenia: A systematic review of mendelian randomization studies

Abstract Background Gut dysbiosis has been repeatedly linked to immune thrombocytopenia (ITP), but confounding and reverse causation hamper interpretation. Mendelian randomization (MR) leverages germ-line variants as proxies for microbial abundance, offering quasi-experimental insight into causality. We systematically reviewed MR evidence probing the microbiota–ITP axis. Methods PubMed, Embase and Web of Science were searched from inception to 1 July 2025 for two-sample or multivariable MR studies that treated gut taxa (phylum–species) as exposures and ITP as outcomes. Three reviewers extracted exposure/outcome GWAS characteristics and synthesised findings narratively because taxonomic depth and analytic pipelines differed. Results Five eligible Mendelian-randomisation studies—Guo 2023, Jiang 2024, Li 2024, Hong 2024 and Ji Gan Wang 2024—each analysed 196–473 gut microbial traits in more than 600,000 MiBioGen (±Dutch Microbiome Project) participants and harmonised these with FinnGen or UK Biobank GWAS datasets that included 810–3,804 European immune-thrombocytopenia (ITP) cases; all instruments were strong (F > 10) and sensitivity analyses revealed negligible horizontal pleiotropy. Across studies, risk-enhancing taxa replicated in at least two analyses—Alcaligenaceae, methanogens, Alistipes, Gordonibacter and Coprococcus—with Guo first identifying Alcaligenaceae (odds ratio [OR] 2.40) and Gordonibacter (OR 1.64) as hazards. Protective taxa comprised the Bacteroidales S24-7 group, Actinobacteria, Porphyromonadaceae and the Eubacterium oxidoreducens group; Li confirmed a strong inverse association for Bacteroidales S24-7 (OR 0.46). Hong 2024 showed that three serum metabolite ratios—sphingomyelin, glucose-to-mannose and bilirubin—mediated 8–21 % of the total causal effects of Intestinimonas and Senegalimassilia on ITP, while Wang 2025 demonstrated that UBA1066, Treponema and Haloplasma transmitted risk via distinct cytokines (CXCL10, VEGF-A and CXCL5, respectively), accounting for 4–17 % of their total effects, and further established five cytokines (CXCL10, CXCL5, IL-12RA, TRAIL and VEGF-A) as directly causal for ITP. Reverse-direction MR consistently indicated that genetically predicted ITP does not modify gut composition, and leave-one-out tests across all studies confirmed the robustness of these estimates. Conclusions Convergent MR evidence now implicates distinct gut microbes in the modulation of ITP susceptibility. High-risk signatures (methanogens, Alistipes, UBA1066, Treponema, Haloplasma) and protective taxa (Bacteroidales S24-7, Porphyromonadaceae, Eubacterium spp.) elucidate mechanistic pathways, while cytokine-specific mediation indicates targeted anti-inflammatory strategies. Future metagenomic research on diverse ancestries, along with strain-level metagenomics and integrated multi-omics, is essential to enhance biomarkers and create microbiome-targeted interventions for ITP.