Health and disease markers correlate with gut microbiome composition across thousands of people

The human gut microbiome is increasingly recognized for its role in various diseases, including inflammatory bowel disease, type 2 diabetes, hypertension, and colorectal cancer. While previous research has established links between specific gut bacteria and blood markers or lifestyle factors, large-scale studies integrating microbiome profiles, clinical data, and lifestyle information have been limited. This study addresses this gap by analyzing the relationship between the gut microbiome and a comprehensive set of host factors in a large cohort of healthy US individuals. The primary goal is to identify lifestyle and clinical factors associated with gut microbiome composition, diversity, and functionality, providing a foundation for future research on host-microbe interactions and interventions.

Existing literature highlights the gut microbiome's influence on various diseases and its responsiveness to lifestyle and dietary changes. Studies have shown associations between specific gut bacteria and biomarkers for conditions like diabetes and high cholesterol. The impact of diet and physical activity on the microbiome has also been explored, although large-scale studies integrating multiple host factors are scarce. This study builds upon previous research by employing a comprehensive analysis of host factors in a large, diverse population.

This cross-sectional study involved 3409 participants from a commercial wellness program (Arivale Inc.). Data included lifestyle questionnaires, stress assessments, digestion information, diet questionnaires, clinical blood tests, and 16S rRNA gene amplicon sequencing of stool samples to determine gut microbiome composition. Statistical analyses involved Pearson correlations, edgePCA (a microbiome-specific principal component analysis), NMDS (nonmetric multidimensional scaling), linear regression models (adjusted for age, sex, race, season, and sequencing vendor), and generalized linear models (for microbiome genera and pathways). The analysis also involved defining clusters based on Firmicutes-to-Bacteroidetes and Bacteroidetes-to-Prevotella axes to investigate cluster-specific associations. The impact of medication usage was assessed using Poisson regression models adjusted for relevant biomarkers.

The study revealed a nonlinear association between Bacteroidetes abundance and Shannon diversity, with a diversity maximum at approximately 15% Bacteroidetes and 80% Firmicutes. Principal component analysis identified Firmicutes and Bacteroidetes as major drivers of taxonomic variance, along with the genera Prevotella and Bacteroides. Seventy-five out of 148 lifestyle and clinical factors were significantly correlated with Shannon diversity, including markers for diabetes, inflammation, liver function, and cholesterol. Positive correlations were observed with omega-3 fatty acids, fish intake markers, and physical activity. Negative correlations were found with BMI, weight, blood pressure, and sugary drink consumption. Furthermore, the study identified unique associations within different microbiome compositional clusters (Firmicutes-rich, Bacteroidetes-rich, Prevotella-rich). Host factors aggregated into health-related and disease-related groups based on their microbiome association patterns. Health-related factors were positively correlated with genera like Coprococcus, Lachnospira, and Faecalibacterium, while disease-related factors showed positive correlations with Bacteroides, Ruminoccoccaceae, and Sutterella. Finally, the analysis explored the impact of medications on the microbiome, revealing associations between specific medications and changes in the abundance of genera and functional pathways.

The study's findings provide valuable insights into the complex interplay between the gut microbiome and host factors. The identification of both known and novel associations highlights the microbiome's potential as a biomarker for health and disease. The nonlinear relationship between Bacteroidetes abundance and diversity suggests the limitations of using simple ratios (like Firmicutes-to-Bacteroidetes) to capture the complexity of microbiome-host interactions. Cluster-specific associations emphasize the importance of considering microbiome heterogeneity when designing personalized interventions. The aggregation of host factors into health-related and disease-related groups based on microbiome associations underscores the potential of the microbiome to provide a holistic view of health and wellbeing. The identification of medication-induced changes in the microbiome provides further evidence for the gut's role in drug response and metabolism.

This large-scale study provides compelling evidence for the intricate relationship between the gut microbiome and a wide array of host phenotypes, including health and disease markers. The findings suggest potential avenues for targeted interventions aimed at modifying the microbiome composition to improve health outcomes. Future research should focus on mechanistic studies and randomized controlled trials to establish causality and refine personalized interventions based on microbiome profiles. Longitudinal studies are also needed to better understand the dynamic interplay between the microbiome and host factors over time.

While the study's large sample size and comprehensive phenotyping are strengths, some limitations exist. The cross-sectional design limits causal inferences, and residual confounding might exist despite adjusting for multiple covariates. The reliance on self-reported data for some lifestyle factors introduces potential biases. Furthermore, the study population primarily consisted of self-reported European-Americans, limiting the generalizability of the findings to other populations. Future research should address these limitations using prospective studies and diverse cohorts.