Changes in Gastrointestinal Microbiome Composition in PD: A Pivotal Role of Covariates Original paper

What was studied?

The focus keyphrase “gut microbiome in Parkinson’s disease” anchors this study’s central aim: to determine how gut microbial composition differs between individuals with Parkinson’s disease (PD) and healthy controls, while rigorously accounting for major PD-associated covariates. This original research article examined fecal microbiota in a southern German cohort, applying strict exclusion and matching criteria to reduce confounding. Investigators evaluated how disease duration, disease stage, levodopa dose, non-motor symptoms, constipation, and coffee intake shaped microbial signatures. By sequencing the V3–V4 region of the 16S rRNA gene, the authors sought to disentangle PD-specific microbial shifts from those driven by lifestyle, motor impairment, or gastrointestinal dysfunction. They used classical statistics, random forest classification, and SparCC compositional methods to identify taxa consistently associated with PD after adjusting for relevant covariates.

Who was studied?

The study analyzed 101 individuals from an outpatient neurology center in Erlangen, Germany: 71 with clinically diagnosed PD and 30 healthy controls, mostly spouses or relatives. Participants ranged broadly in age but were well-matched demographically. Strict criteria excluded those with gastrointestinal disease, recent antibiotics, abdominal surgery, opioid use, or other confounders known to influence microbial composition. PD participants represented Hoehn & Yahr stages 1–4, allowing evaluation of disease-stage effects. After propensity score matching for constipation severity, non-motor burden, and coffee intake—three variables strongly differing between groups—the matched subset included 28 PD patients and 19 controls. This refined comparison enabled clearer attribution of observed microbial signatures to PD itself rather than secondary influences.

Most important findings

The most robust PD-associated microbial features were decreased Faecalibacterium and Ruminococcus, two genera within Firmicutes that produce anti-inflammatory short-chain fatty acids. Before covariate matching, several taxa differed between groups, including alterations within Clostridia, Lachnospiraceae, and Sutterella. Random forest analysis ranked Faecalibacterium, Sutterella, Oscillospira, Ruminococcus, and Blautia as the most discriminative genera. SparCC correlation pinpointed Faecalibacterium as negatively correlated with PD and Bacteroides as positively correlated. Disease stage and duration influenced microbial patterns, with Faecalibacterium abundance declining alongside disease progression. Levodopa dose—but not motor severity—affected correlations between duration and key genera, suggesting medication-microbiome interactions. After rigorous matching, Ruminococcus remained significantly reduced, and SparCC identified only Bacteroides and Faecalibacterium as correlating with PD, underscoring their potential diagnostic value.

Key implications

This study emphasizes that many microbial differences attributed to PD may arise from covariates such as constipation, coffee intake, and dopaminergic medication rather than the disease itself. However, the persistent reduction of Faecalibacterium and Ruminococcus—both butyrate-producing taxa—supports the hypothesis that impaired gut barrier integrity and reduced anti-inflammatory fermentation may contribute to PD pathophysiology. These taxa may represent actionable microbial biomarkers for earlier detection or therapeutic intervention. The findings reinforce the need for covariate-aware designs in microbiome research and suggest that future PD microbiome studies should incorporate drug-naïve cohorts, mucosal sampling, and longitudinal analysis for clearer mechanistic insight.

Citation

Cosma-Grigorov A, Meixner H, Mrochen A, Wirtz S, Winkler J, Marxreiter F. Changes in gastrointestinal microbiome composition in PD: a pivotal role of covariates. Front Neurol. 2020;11:1041. doi:10.3389/fneur.2020.01041