Abstract
The gut microbiome is believed to play an important role in the development and onset of Parkinson's disease (PD). While human studies report differences in gut microbiota between PD individuals and healthy controls, it is unclear whether preclinical animal models show similar patterns. We performed a systematic review and Bayesian regularised meta-analysis of preclinical PD studies that assessed both motor function and gut microbiota. Motor deficits were consistently observed across models, but gut bacterial diversity (α-diversity) and changes in key taxa (e.g. Akkermansia, Lactobacillus, Bifidobacterium) were inconsistent and poorly aligned with human data. In contrast, short-chain fatty acids (SCFAs) showed more reproducible changes and greater translatability to human findings. Chronic toxin-based models demonstrated the highest reproducibility. Overall, gut microbiota composition in animal PD models lacks consistency and human relevance, whereas SCFAs may offer a more reliable outcome. Finally, our study makes possible recommendations for reporting to improve future studies.