Do oral microbiome changes have enough human evidence for clinical recommendations?

How strong is the evidence that oral microbiome changes can detect disease?

The strongest human evidence comes from studies that built and validated diagnostic classifiers—essentially, algorithms that use the pattern of bacteria in your mouth to predict whether you have a disease. A 2025 study of saliva samples from patients with upper gastrointestinal and pancreaticobiliary cancers found that oral microbial composition was significantly different between cancer patients and healthy controls (p = 0.001 to 0.004) [1]. The classifier for gastric cancer achieved an area under the curve (AUC) of 0.961, meaning it correctly distinguished cancer from non-cancer about 96% of the time, using just five bacterial genera and three species [1]. For esophageal cancer, the classifier was still good but less accurate (AUC = 0.791) [1]. This tells us that oral microbiome signatures can be powerful for some cancers but not equally for all.

Similarly, a 2021 study on COVID-19 collected 392 tongue-coating samples and built an eight-oral-microbial-marker classifier that achieved 87.2% diagnostic accuracy in a cross-regional validation cohort, and 92.1% in suspected patients who were IgG-positive [3]. That study also showed that butyric acid-producing bacteria decreased and lipopolysaccharide-producing bacteria increased in COVID-19 patients, linking microbial shifts to immune and metabolic changes [3]. These two studies—one on cancer, one on an infectious disease—demonstrate that oral microbiome changes can be robust enough to serve as non-invasive diagnostic tools, at least in research settings.

What about other diseases—and do we know if oral microbiome changes cause disease or just correlate?

Beyond cancer and COVID-19, human studies have linked oral microbiome changes to preterm birth, cholangiocarcinoma (bile duct cancer), and systemic diseases like diabetes and rheumatoid arthritis. A 2023 case-control study of 152 pregnant women found that the genera Veillonella, Prevotella, and Capnocytophaga were enriched in the oral microbiome of women who delivered preterm, independent of other pregnancy complications [6]. This suggests a potential causal role, but the authors note that no specific bacterial species has been proven causative yet [6]. A 2022 study on cholangiocarcinoma found 36 genera significantly different between patients and healthy controls, with Streptococcus, Veillonella, and Haemophilus enriched in cancer patients [4]. These associations are consistent across multiple studies, which strengthens the case that oral microbiome changes are not random noise.

However, most of this evidence is associative, not causal. A 2023 review on the oral microbiome in health and disease emphasizes that while oral dysbiosis is linked to cancer, cardiovascular disease, and respiratory disease, the direction of causality is often unclear—disease can worsen microbiome imbalance, creating a vicious cycle [9]. The same review notes that understanding the heterogeneity of both the oral microbiome and individual humans is critical for establishing causal links [9]. A 2023 perspective article adds that factors like diet, smoking, and medications can disrupt the oral microbiome, making it hard to separate cause from effect [7]. So while the evidence is strong enough to say 'oral microbiome changes are associated with disease,' we are not yet at the point where we can say 'changing your oral microbiome will prevent disease' for most conditions.

What would it take for oral microbiome testing to become a clinical recommendation?

For a test to move from research to clinical guidelines, it needs standardized methods, validation in diverse populations, and proof that it improves patient outcomes. A 2025 review on microbiome translation notes that significant challenges remain, including interindividual variability, incomplete functional annotation of microbial 'dark matter,' and the absence of validated biomarkers [8]. The same review calls for standardized methodologies, harmonized regulatory frameworks, and longitudinal studies across diverse populations [8]. The studies we have are promising but mostly small: the cancer classifier study had a case-control design (not a prospective screening trial) [1], and the COVID-19 study, while cross-regional, was still a case-control design [3].

Another key gap is transmission and stability. A 2023 Nature study of over 9,700 human metagenomes found that oral microbiome strain sharing is largely horizontal (between cohabiting adults) and increases with time spent together, with a median strain-sharing rate of 32% among cohabitants [2]. This means your oral microbiome is influenced by who you live with, which could complicate using it as a stable biomarker. Additionally, a 2023 study on the Planetary Health Diet showed that dietary changes can alter the gut microbiome within weeks, but the effect on the oral microbiome was not directly measured [5]. For clinical recommendations, we would need to know how stable oral microbiome signatures are over time and how much they are affected by diet, medications, and household contacts. In summary, the human evidence is strong enough to say 'oral microbiome changes are clinically relevant,' but not yet strong enough to say 'everyone should get their oral microbiome tested.'

About These Sources

This answer is built on 9 peer-reviewed studies — published from 2021 to 2025, 3 from 2024 or later, 8 in Q1 journals, collectively cited 910 times — selected as the most relevant from 13 studies that passed quality screening, drawn from 46 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions