The intestinal microbiota, also commonly known as the ‘‘gut microbiome’’ is integral to human physiology and has wide-ranging effects on the development and function of the immune system, energy metabolism and even nervous system activity. There is a lot of excitement around the potential of targeting the microbiome therapeutically to promote health and to prevent or treat medical conditions. Further, as DNA sequencing technologies and computational methods continue to improve, there is significant interest in developing microbiome-based diagnostics for clinical applications.

The industry has recognized the consumer interest in microbiome-based diagnostics as an opportunity, and a number of commercial laboratories are marketing tests directly to patients. Health care providers are increasingly being asked by their patients to help interpret such test reports; in some cases, the patient may even request a physician order to purchase the tests for insurance coverage or other reasons.

Providers need a basic understanding of the current state of microbiome science, so my colleagues and I recently published a primer for clinicians on microbiome testing. Other colleagues have recently done the same. Below are the key takeaways from our article, which can help you in your interactions with patients.

Limitations of microbiome sequencing. Microbiome datasets have the same limitations as any other sample-dependent dataset. First and foremost, a single stool sample will tell you something about a person’s microbiome profile only at the time and location that the sample was collected. How the sample was collected and how it was stored may significantly impact the analysis. The analysis generally provides a relatively low-resolution overview of bacteria present, mostly at a family and genus level, and little information about the viruses, protozoa, and fungi. Furthermore, stool analysis may not reflect well the microbiome composition at the mucosal surface in the intestine. As a result, a single analysis of an individual stool sample merely provides a snapshot of the fecal microbiome that is incomplete and extremely limited in what we can learn from it.

“Good” vs. “bad.” The reports resulting from microbiome-based tests often describe the patient’s microbiome profile in terms of relative abundance or how much “good” and “bad” bacteria are present. This kind of a classification framework represents a naïve and cartoonish view of the microbial world. Instead, it is important to appreciate microbial communities as functional networks, and that their functionality cannot be defined as a mere summation of individual microorganisms. Microbes, just like people, vary their behavior in accordance with the context that may be provided by the activity of other microbes and the host. Whether a particular species or strain is helpful or harmful depends on what other bacteria are present, their density, how they interact with each other (e.g., are they mutually beneficial or competitive?), and factors from the human host, such as their diet or immune system activity. For example, Clostridioides difficile (C. difficile) is a potential pathogen, yet it also naturally exists in the intestines of many people as a non-harmful, commensal species. Its pathogenic potential depends on the state of the other intestinal microbes and host factors, such as presence of anti-C. difficile toxin antibodies.

Importantly, microbiome tests, which generally provide only a low-resolution microbial community overview, are not designed for pathogen identification. That is best done with targeted diagnostics. Even then, as well illustrated by the C. difficile example, diagnosis of an infection cannot be made on the basis of laboratory testing alone and requires clinical information.

Taxonomy vs. function. Current technology allows a fairly inexpensive characterization of most bacterial taxa (at family and genus levels). However, taxonomy is not easily translated into functional information. Different taxa of microbes may be able to execute the same chemical transformations. In contrast, functional information depends on the genes present and how much are these genes expressed. However, obtaining this kind of information is much more resource intensive. Measurements of metabolites may also provide very valuable functional information, but proper sample collection for metabolomics is much more difficult, and interpretation is complicated by multiple factors that influence fecal levels of individual metabolites, including production, absorption and intestinal transit time.

Inter-individual variability. The consistent lesson we’ve learned from the microbiome literature is that there is not a single “healthy” microbiome profile. We have not identified a particular microbiome profile that is predictive of a particular disease, though many researchers are working to develop microbiome-based indices for diseases such as inflammatory bowel disease or obesity. Crowd-sourced studies such as the American Gut Project are working to expand and diversify microbiome datasets so that we can better understand the variability and begin to identify reproducible microbiome signatures. The microbiome data is extremely multidimensional and complex. Therefore, developing predictive patterns will likely require analyses of tens of thousands of samples linked to highly granular clinical metadata.

Microbiome-based tests have potential to transform clinical care and become incorporated into the personalized medicine paradigm. However, we are at the very beginning of understanding what one’s microbiome profile means for their susceptibility to or progression of disease. As patients approach their health care providers with requests to order commercial microbiome-based tests or to help interpret a report, it is important to set the expectation that these tests are not well suited for diagnoses of infectious diseases or validated in specific diagnoses of any diseases. There are far more unknowns than knowns regarding the role of the microbiome and human health.

When talking with your patients about microbiome testing, you can emphasize:

• There is no definitive example of a “healthy” microbiome, nor any examples of a specific microbiome profile that can reliably predict whether someone will develop a particular disease.
• Current microbiome-based tests cannot diagnose disease. They may be able to detect pathogens, but that can already be done (and better) by existing diagnostic tests that health care providers already use.
• The bacteria in our microbiomes are neither “good” nor “bad,” but rather different person-to-person, based on what other microbes are present and the environment that they are in.
• We’re still in the beginning stages of determining how microbiome-based tests will impact clinical care. Currently, they do not have clinical value for patients or their health care providers.
• There is much more we need to learn regarding the role of the microbiome in human health before microbiome profiles become a meaningful part of clinical diagnosis and care.

Alexander Khoruts is a gastroenterologist and member, AGA Center for Gut Microbiome Research & Education scientific advisory board.

Image credit: Shutterstock.com