PDF(Pubmed)
Abstract：
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.
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), HbA1c 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.
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 HbA1c (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.
Overall, higher levels of HOMA-IR, CRP and HbA1c 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.
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), HbA1c 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.
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 HbA1c (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.
Overall, higher levels of HOMA-IR, CRP and HbA1c 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.
摘要：
目的：假设肠道菌群与胰岛素抵抗和其他代谢变量有关。然而,基于人群的研究数据有限。我们调查了北芬兰出生队列1966（NFBC1966）和TwinsUK队列中代谢健康的血清学指标与肠道微生物组之间的关联。
方法：在来自NFBC1966的506名具有可用粪便微生物组（16SrRNA基因序列）数据的个体中，我们估计了肠道微生物组多样性指标与HOMA胰岛素抵抗（HOMA-IR）血清学水平之间的关联，HbA1c和C反应蛋白（CRP）使用多变量线性回归模型调整性别，吸烟状况和BMI。肠道微生物组多样性测量与HOMA-IR和CRP之间的关联在来自TwinsUK的1140名成年参与者中被复制，可用的粪便微生物组(16SrRNA基因序列)数据。对于这两个队列，我们使用具有准泊松分布和基于微生物组回归的核关联检验（MiRKAT）的一般线性模型来估计代谢变量与α和β多样性指标的关联，分别,以及适用于零膨胀β分布的位置尺度和形状（GAMLSS）的广义加性模型，以识别与代谢标记相关的分类单元。
结果：在NFBC1966中，HOMA-IR较高的个体的α多样性较低，HOMA-IR的第一个四分位数平均为74.4（95％CI70.7，78.3）扩增子序列变体（ASV），HOMA-IR的第四个四分位数为66.6（95％CI62.9，70.4）。HbA1c越高，α多样性也越低(ASV数量和香农多样性，p分别<0.001和p=0.003)和更高的CRP(ASV的数量，p=0.025),即使在调整了BMI和其他潜在的混杂因素之后。在TwinsUK,在HOMA-IR和CRP测量值较高的参与者中,α多样性测量值也较低.当考虑β多样性措施时，我们发现NFBC1966和TwinsUK中的微生物群落分布与HOMA-IR相关，使用多变量MiRKAT模型，二项式偏差差异p值<0.001。在GAMLSS模型中，在两个队列中,普雷沃氏菌属和布劳特氏菌属的相对丰度与HOMA-IR相关.
结论：总体而言，更高水平的HOMA-IR，在NFBC1966和TwinsUK队列中，CRP和HbA1c与较低的微生物组多样性相关，即使在调整了BMI和其他变量之后。来自两个不同的基于人群的队列的这些结果为代谢变量与肠道微生物多样性之间的关联提供了证据。现在需要进一步的实验和机械见解，以提供对可能将肠道微生物群与代谢健康联系起来的潜在因果机制的理解。
方法：在来自NFBC1966的506名具有可用粪便微生物组（16SrRNA基因序列）数据的个体中，我们估计了肠道微生物组多样性指标与HOMA胰岛素抵抗（HOMA-IR）血清学水平之间的关联，HbA1c和C反应蛋白（CRP）使用多变量线性回归模型调整性别，吸烟状况和BMI。肠道微生物组多样性测量与HOMA-IR和CRP之间的关联在来自TwinsUK的1140名成年参与者中被复制，可用的粪便微生物组(16SrRNA基因序列)数据。对于这两个队列，我们使用具有准泊松分布和基于微生物组回归的核关联检验（MiRKAT）的一般线性模型来估计代谢变量与α和β多样性指标的关联，分别,以及适用于零膨胀β分布的位置尺度和形状（GAMLSS）的广义加性模型，以识别与代谢标记相关的分类单元。
结果：在NFBC1966中，HOMA-IR较高的个体的α多样性较低，HOMA-IR的第一个四分位数平均为74.4（95％CI70.7，78.3）扩增子序列变体（ASV），HOMA-IR的第四个四分位数为66.6（95％CI62.9，70.4）。HbA1c越高，α多样性也越低(ASV数量和香农多样性，p分别<0.001和p=0.003)和更高的CRP(ASV的数量，p=0.025),即使在调整了BMI和其他潜在的混杂因素之后。在TwinsUK,在HOMA-IR和CRP测量值较高的参与者中,α多样性测量值也较低.当考虑β多样性措施时，我们发现NFBC1966和TwinsUK中的微生物群落分布与HOMA-IR相关，使用多变量MiRKAT模型，二项式偏差差异p值<0.001。在GAMLSS模型中，在两个队列中,普雷沃氏菌属和布劳特氏菌属的相对丰度与HOMA-IR相关.
结论：总体而言，更高水平的HOMA-IR，在NFBC1966和TwinsUK队列中，CRP和HbA1c与较低的微生物组多样性相关，即使在调整了BMI和其他变量之后。来自两个不同的基于人群的队列的这些结果为代谢变量与肠道微生物多样性之间的关联提供了证据。现在需要进一步的实验和机械见解，以提供对可能将肠道微生物群与代谢健康联系起来的潜在因果机制的理解。
