Markers of metabolic health and gut microbiome diversity: findings from two population-based cohort studies

• Published in: Diabetologia
• Main Authors: Zouiouich, Semi, Loftfield, Erikka, Huybrechts, Inge, Viallon, Vivian, Louca, Panayiotis, Vogtmann, Emily, Wells, Philippa M., Steves, Claire J., Herzig, Karl-Heinz, Menni, Cristina, Jarvelin, Marjo-Riitta, Sinha, Rashmi, Gunter, Marc J.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Springer Science and Business Media LLC 2021
• Subjects: Endocrinology, Diabetes and Metabolism; Internal Medicine; Northern Finland
• Online Access: http://dx.doi.org/10.1007/s00125-021-05464-w; https://link.springer.com/content/pdf/10.1007/s00125-021-05464-w.pdf; https://link.springer.com/article/10.1007/s00125-021-05464-w/fulltext.html

Abstract Aims/hypothesis The gut microbiome is hypothesised to be related to insulin resistance and other metabolic variables. However, data from population-based studies are limited. We investigated associations between serologic measures of metabolic health and the gut microbiome in the Northern Finland Birth Cohort 1966 (NFBC1966) and the TwinsUK cohort. Methods Among 506 individuals from the NFBC1966 with available faecal microbiome (16S rRNA gene sequence) data, we estimated associations between gut microbiome diversity metrics and serologic levels of HOMA for insulin resistance (HOMA-IR), HbA 1c and C-reactive protein (CRP) using multivariable linear regression models adjusted for sex, smoking status and BMI. Associations between gut microbiome diversity measures and HOMA-IR and CRP were replicated in 1140 adult participants from TwinsUK, with available faecal microbiome (16S rRNA gene sequence) data. For both cohorts, we used general linear models with a quasi-Poisson distribution and Microbiome Regression-based Kernel Association Test (MiRKAT) to estimate associations of metabolic variables with alpha- and beta diversity metrics, respectively, and generalised additive models for location scale and shape (GAMLSS) fitted with the zero-inflated beta distribution to identify taxa associated with the metabolic markers. Results In NFBC1966, alpha diversity was lower in individuals with higher HOMA-IR with a mean of 74.4 (95% CI 70.7, 78.3) amplicon sequence variants (ASVs) for the first quartile of HOMA-IR and 66.6 (95% CI 62.9, 70.4) for the fourth quartile of HOMA-IR. Alpha diversity was also lower with higher HbA 1c (number of ASVs and Shannon’s diversity, p < 0.001 and p = 0.003, respectively) and higher CRP (number of ASVs, p = 0.025), even after adjustment for BMI and other potential confounders. In TwinsUK, alpha diversity measures were also lower among participants with higher measures of HOMA-IR and CRP. When considering beta diversity measures, we found that microbial community profiles were associated with HOMA-IR in NFBC1966 and TwinsUK, using multivariate MiRKAT models, with binomial deviance dissimilarity p values of <0.001. In GAMLSS models, the relative abundances of individual genera Prevotella and Blautia were associated with HOMA-IR in both cohorts. Conclusions/interpretation Overall, higher levels of HOMA-IR, CRP and HbA 1c were associated with lower microbiome diversity in both the NFBC1966 and TwinsUK cohorts, even after adjustment for BMI and other variables. These results from two distinct population-based cohorts provide evidence for an association between metabolic variables and gut microbial diversity. Further experimental and mechanistic insights are now needed to provide understanding of the potential causal mechanisms that may link the gut microbiota with metabolic health. Graphical abstract