The findings of a Chinese-led international study into regional variations in reference ranges and models of diseases clearly related to dysbiosis in the gut microbiome could have important public health implications for managing disorders including fatty liver and other metabolic diseases.

Defined as a departure of the gut microbiome from that seen in healthy individuals, dysbiosis has been proposed as being a powerful biomarker of disease incidence and progression, with such noninvasive techniques having potentially valuable public health applications.

Microbiota tests are now commercially available that can identify dysbiosis by comparing the relative abundance of gut microbes between patients and healthy reference populations.

However, "dysbiosis is not widely used as a biomarker for disease diagnosis and progression in China, as it has not been approved by Chinese FDA for clinical use, mainly due to insufficient clinical evidence," study leader Hong-Wei Zhou told BioWorld.

"Therefore hospitals are not permitted to use it clinically. While some small companies are now marketing such services, this is not well-known to the public," said Zhou, professor and director of the Division of Laboratory Medicine at Zhujiang Hospital, Southern Medical University, Guangzhou.

A major reason is that dysbiosis patterns vary among studies and are potentially affected regionally. For example, studies comparing Western and non-industrialized populations have found substantial variability in gut microbiota characteristics among those two groups.

Moreover, the ability to generalize healthy baseline and disease model data among different locations has been investigated in only a few meta-analyses and small-scale studies, which may not necessarily be applicable to other populations.

Microbiome diagnostic signals are also affected by a wide range of confounding factors, including drugs, and human gut microbiota are highly variable, so large sample sizes are needed to draw reliable conclusions.

Diagnoses based on microbiota differences between health and disease therefore require data from studies with a regionalized study design, extensive sampling methods, and standardized study protocols.

Such noninvasive sampling could facilitate large-scale public health applications, including early diagnosis and risk assessment in metabolic and cardiovascular diseases, the possibility of which was assessed in the Guangdong Gut Microbiome Project.

Guangdong Gut Microbiome Project

To understand the generalizability of microbiota-based diagnostic models of metabolic disease, researchers led by Zhou conducted a stratified, randomized study to characterize the gut microbiota of roughly 7,000 individuals from 14 districts in Guangdong province.

Among the different phenotypes analyzed, host locations were shown to have the strongest associations with microbiota variations, far exceeding the effects of other host phenotypes, the researchers reported in the study published in the Sept. 3, 2018, edition of Nature Medicine.

That is a significant finding, since "human gut microbiota are causally associated with many diseases and highly variable, meaning that two individuals could harbor totally different gut microbial configurations," said Zhou.

"Understanding factors shaping such variability is one of the major goals in this field. Previously, we knew that age, ... drugs, diet, and so forth were major co-variates of human gut microbiota, but our study showed geography could have the strongest phenotypic associations with microbiota variations," he said.

Major factors causing such microbiota variations were not addressed in this study, said Zhou. "Whether the geographical effect we report is driven by host-specific factors or introduced by other ecological processes, such as dispersal, drift, local diversification or host interactions with environmental microbiota, requires further investigation," he noted.

The study also showed that microbiota-based metabolic disease models developed in one location failed when used elsewhere in Guangdong Province, suggesting such models cannot be extrapolated to include general populations.

"We showed that applying [a] microbiota-based metabolic disease diagnostic model built extrapolating data from one district to other regions generally failed," said Zhou.

Interpolation works better than extrapolation

However, while extrapolations among districts did not yield useful models, interpolated models, where a value is estimated from two known values, performed much better, especially in diseases with obvious microbiota-related characteristics.

"When we tested combining data from all districts to build a province-level model, i.e. an interpolated model and applied it to Guangdong, this showed that, while it is difficult to extrapolate metabolic diseases models between locations, interpolating data collected on a larger geographic scale might work," he said.

"Results will depend on identifying consistent population signatures, which becomes more difficult as the geographic scale becomes larger, although the interpolation model can provide dysbiosis signatures for a relatively large area."

The researchers conclude those data reinforce the need for consistent sampling protocols in order to construct localized reference baselines for gut microbiota studies.

Furthermore, the applicability of disease models in different populations must be tested rather than assumed, with populations used to generate healthy reference data and disease models being clearly stated when performing gut microbiome analyses, especially for clinical use.

"In the present study, the dysbiosis-associated characteristics are obvious for fatty liver disease and to a lesser extent for type 2 diabetes mellitus and metabolic syndrome," Zhou said.

Regarding public health, "examination of human gut microbiota is non-invasive and could facilitate large-scale public health applications, including early diagnosis and risk assessment of such metabolic and cardiovascular diseases," he said.

In the future, "we will try to determine what factors shape human gut microbiota, including causes of geographical variations, and explore the associations between the gut microbiota and host metadata observed in our study."