A Stool Sample Is Not a Diagnosis

The human gut microbiome is one of the most active areas of biomedical research and one of the easiest to overinterpret.

A stool sample can now be sequenced cheaply enough to produce a report containing hundreds or thousands of microbial taxa, diversity metrics, colourful bar charts and personalised recommendations. The report may assign a gut-health score, identify “beneficial” and “harmful” organisms, calculate a dysbiosis index and recommend foods, prebiotics or probiotics intended to move the microbiome toward a healthier state.

The technological sophistication is real.

The clinical meaning is much less mature.

This distinction is central to diagnostic medicine. A test can measure something reproducibly without that measurement being clinically useful. It can also produce a biologically interesting signal that has no validated threshold, no agreed reference range and no demonstrated ability to improve outcomes when acted upon.

Microbiome testing currently sits inside that gap.

Analytical validity, clinical validity and clinical utility are different problems

Three questions should be separated.

Analytical validity asks whether the laboratory can measure the intended feature accurately and reproducibly.

Clinical validity asks whether the measured feature reliably distinguishes a clinically meaningful state such as disease, prognosis or treatment response.

Clinical utility asks whether using the result to guide decisions improves outcomes.

A consumer microbiome service may perform advanced DNA sequencing and still fail at the second or third level.

That is not unusual in medicine. Many biomarkers are biologically real without being clinically useful. The difference is that microbiome reports often present all three levels as if they were already solved.

An international consensus statement published in The Lancet Gastroenterology & Hepatology in 2025 was unusually direct: evidence supporting routine clinical use of microbiome testing remains limited, and widespread direct-to-consumer testing has developed ahead of demonstrated clinical value.

The panel discouraged several common practices, including reporting the Firmicutes-to-Bacteroidetes ratio as a clinical result, using dysbiosis indices as if they were established diagnostic metrics and providing strict “healthy” reference ranges for individual taxa.

That is an important contrast with the confidence of many commercial reports.

Different companies can produce materially different answers from the same sample

The most basic requirement of a laboratory test is reproducibility.

A 2026 study involving NIST evaluated seven direct-to-consumer gut microbiome testing services using standardised faecal reference material. The investigators found substantial discrepancies both within and between providers. Variation between companies was on the same scale as biological variation between different donors.

That finding matters because interpretation is downstream of measurement.

If two laboratories describe the same sample differently, recommendations based on those descriptions inherit that uncertainty.

The reasons are not mysterious. Microbiome analysis contains many methodological choices: DNA extraction, primer selection, sequencing platform, sequencing depth, quality filtering, taxonomic database and computational pipeline can all alter the final profile.

A report may therefore appear objective while containing substantial method dependence.

This does not mean sequencing is unreliable as a research technology. It means standardisation matters when results are translated into individual clinical claims.

A stool sample is a sample of stool, not a map of the entire gut

Another problem is anatomical.

Faecal samples are convenient and useful. They are also incomplete.

The gastrointestinal tract contains different microbial environments along the stomach, small intestine, colon, mucus layer and mucosal surface. Oxygen, pH, nutrient availability and transit vary by location.

Stool predominantly reflects luminal material from the distal gastrointestinal tract.

It does not perfectly represent mucosa-associated communities or microbes living elsewhere in the intestine.

Clinical reviews of microbiome methodology repeatedly emphasise this distinction. Faecal sampling is valuable precisely because it is practical and non-invasive, but the sample type defines the biological question.

The problem begins when a stool profile is described as “your gut microbiome” without qualification.

It is a useful approximation to part of the system, not a complete census of the gastrointestinal ecosystem.

One sample is also a snapshot

Human microbiomes have both stability and variability.

At broad levels, individuals often retain recognisable microbial signatures over time. But specific taxa can fluctuate substantially with diet, stool consistency, transit, illness, antibiotics and other exposures.

A densely sampled longitudinal study of healthy adults found substantial day-to-day variation in many major genera. For 78% of analysed genera, day-to-day absolute abundance variation within a person exceeded between-person variation, with some taxa showing very large shifts.

Other longitudinal studies show that broad community features can remain sufficiently stable for some epidemiological questions while lower-abundance organisms and specific taxa vary more.

These findings are not contradictory.

They show that the answer depends on the feature being measured and the question being asked.

A single stool sample can be informative for some purposes. It should not automatically be treated as a timeless description of a person's microbial state.

This is particularly important when a company recommends dietary or supplement changes based on small deviations from a reference database.

The next sample may differ even without a clinically meaningful change.

Relative abundance creates another interpretation problem

Most sequencing reports present organisms as percentages.

That sounds straightforward but introduces a compositional problem.

If one organism increases in relative abundance, another must decrease proportionally even if its absolute number has not changed. Relative data therefore describe fractions of a total rather than independent cell counts.

Suppose one taxon doubles while every other taxon remains constant. Their relative percentages will all fall even though their absolute abundance is unchanged.

This is why percentage charts can create apparent changes that do not correspond to absolute depletion.

The 2025 international consensus statement explicitly warns against interpreting sequence-based relative abundances as absolute quantities and notes that strict healthy reference ranges for individual species are not currently justified.

For consumer reports, this is a major issue because language such as “low in species X” often sounds like a deficiency diagnosis.

A low percentage is not necessarily a biological deficiency.

“Dysbiosis” does not have one accepted clinical definition

The word dysbiosis is widely used to mean an unhealthy microbial imbalance.

Its intuitive appeal exceeds its precision.

Reviews of proposed dysbiosis indices show substantial heterogeneity in how the concept is defined and calculated. Different indices use different taxa, diversity metrics, cohorts and disease contexts.

There is no universal gold-standard healthy microbiome against which every person can be scored.

This is partly because healthy people differ substantially from one another. Geography, diet, medication, age and environment all influence community structure.

The international consensus panel concluded that current evidence is insufficient to include dysbiosis indices routinely in microbiome reports and noted that there is no common clinical definition of dysbiosis.

That is very different from a commercial report giving a person a score of 42 and suggesting that the gut is “unbalanced”.

The numerical scale may belong to the company, not to clinical medicine.

The Firmicutes/Bacteroidetes ratio is a case study in premature biomarkers

Few microbiome metrics have travelled further than the Firmicutes-to-Bacteroidetes ratio.

Early studies suggested that obesity might be associated with a higher ratio. The idea was simple enough to become part of popular microbiome discourse.

Subsequent human research was inconsistent.

Reviews and pooled analyses have reported the ratio as higher, lower or unchanged in obesity. A detailed review concluded that methodological differences, lifestyle factors and poor reproducibility make the ratio unsuitable as a reliable hallmark of obesity.

The 2025 consensus statement goes further and explicitly discourages reporting the Firmicutes-to-Bacteroidetes ratio in clinical microbiome testing.

This is a useful example of how science evolves.

An interesting early association generated a memorable biomarker.

Replication made the story less simple.

Consumer health communication often preserves the early simple story long after the literature has become more cautious.

“Good bacteria” and “bad bacteria” are often context-dependent categories

Microbiome reports frequently classify organisms as beneficial or harmful.

Some organisms do have well-established pathogenic roles. That should not be confused with broad categorisation of commensal taxa.

The effect of a microbial species can depend on strain, abundance, host genetics, diet, coexisting organisms and location.

An organism associated with a favourable phenotype in one study may not be protective in another population. The same taxon can perform different metabolic functions depending on strain-level genes and environmental substrate availability.

Taxonomy is therefore not equivalent to function.

This becomes increasingly important as research moves from asking “who is there?” toward “what are they doing?”

Metabolites, pathways and host response may ultimately be more clinically informative than the abundance of one taxonomic label.

A report that treats every organism as intrinsically good or bad can give an ecological system a simplicity it does not possess.

More diversity is not always automatically better

Microbial diversity is another attractive summary metric.

In some disease states, reduced gut microbial diversity is associated with poorer health. That has encouraged the idea that diversity itself should always be maximised.

The relationship is not universal.

Different body sites have different ecological structures, and high diversity is not synonymous with health in every microbial ecosystem. Even within the gut, diversity is only one property among many.

A high-diversity community can still contain undesirable functional characteristics, while a lower-diversity community can be stable and compatible with health.

The clinical value of diversity therefore depends on context.

Treating one summary statistic as a universal health score repeats the same error seen with inflammation, glucose and cortisol: the visible metric becomes the outcome.

Personalised nutrition is promising, but microbiome testing is only one component

Microbiome research may eventually contribute substantially to personalised nutrition.

Several studies have shown that postprandial metabolic responses differ between individuals and that models incorporating microbiome features can improve prediction.

A 2024 randomized trial in Nature Medicine reported improvements in several cardiometabolic outcomes from a personalised nutrition programme. The programme used microbiome information as one part of a much larger prediction system that also included dietary, physiological and behavioural data.

That trial is important.

It does not establish that a standalone consumer stool test can generate clinically validated dietary advice.

The intervention was the entire personalised programme, not the microbiome assay in isolation.

This distinction mirrors the bundled-intervention problem seen elsewhere in health research. If multiple inputs and behavioural components change together, the result cannot automatically be assigned to the microbiome component.

Personalised nutrition may be effective while the incremental value of one stool test remains uncertain.

Test-guided probiotics remain largely unproven

The most commercially intuitive recommendation is to test the microbiome and then prescribe a probiotic containing organisms that appear deficient.

The logic sounds compelling.

It assumes that the measured abundance is valid, that the organism should be increased, that oral supplementation changes the relevant ecological niche and that the change improves an outcome.

Each step requires evidence.

Current consensus guidance does not support routine post-test therapeutic recommendations from microbiome testing providers.

There is also no established general rule that a bacterium appearing at low relative abundance in stool should be replaced with a probiotic strain bearing a similar species name.

Probiotics are strain-specific interventions. Colonisation can be transient, and clinical benefits depend on disease, strain and endpoint.

“Low on report” is not equivalent to “needs supplementation”.

Antibiotics, diet and stool characteristics can dominate the result

One reason microbiome interpretation is difficult is that many exposures affect the measured community.

Antibiotics can produce large shifts.

Diet can influence composition and microbial metabolism.

Stool consistency and transit time can alter measured relative abundance.

Medication, age, geography and disease status also matter.

This creates a problem for reference ranges.

A healthy individual eating one diet may differ markedly from another healthy individual eating another. That difference may reflect adaptation rather than pathology.

Without a stable definition of health, a deviation from the database mean should not automatically be labelled abnormal.

Reference populations are choices.

The choice should be scientifically justified.

Clinical microbiome testing does have legitimate uses

Criticising consumer interpretation should not be confused with claiming that the microbiome has no clinical value.

Microbiome-based diagnostics and therapeutics are active areas of legitimate research. Faecal microbiota-based therapies are already clinically important in recurrent Clostridioides difficile infection. Microbiome signatures are being studied in inflammatory bowel disease, colorectal cancer, liver disease and response to immunotherapy.

The international consensus statement was not an argument against the field.

It was an argument for standards.

The authors explicitly described microbiome testing as promising while concluding that routine clinical use remains ahead of the evidence.

That is a normal stage in translational science.

Promising is not the same as ready.

The main risk is medicalising normal variation

Consumer testing creates a specific psychological problem.

If a report measures hundreds of organisms, some values will inevitably look unusual.

The person may be told that diversity is too low, one beneficial species is deficient and several potentially harmful species are elevated.

The result can transform ordinary variation into a list of problems requiring correction.

This is especially powerful because the findings are invisible. The consumer cannot independently observe whether their microbiome is “imbalanced”. The report defines both the problem and the solution.

If recommendations then include proprietary supplements, repeated testing or subscription services, the commercial loop becomes obvious.

Again, this does not mean the company is necessarily dishonest.

It means the evidential burden should be high because the test can create its own market by converting uncertain biological variation into actionable abnormalities.

A sophisticated report can still be clinically weak

Microbiome reports often look more advanced than conventional laboratory tests.

They contain sequencing data, machine-learning scores, multidimensional visualisations and taxonomic hierarchies.

That complexity can create an impression of precision medicine.

But clinical usefulness does not increase automatically with the number of features measured.

A simple validated test can be more useful than a thousand poorly interpreted variables.

The crucial question is whether the information changes a decision in a way that improves health.

Until that link is demonstrated, analytical sophistication should not be confused with clinical maturity.

What a defensible report would say

A careful microbiome report would make uncertainty prominent.

It would state which sequencing method was used, what part of the microbiome the stool sample can and cannot represent, how technical variation is controlled, whether results are relative or absolute, and whether repeated sampling is needed for the question being asked.

It would avoid unsupported universal healthy ranges.

It would avoid presenting the Firmicutes/Bacteroidetes ratio as a diagnostic biomarker.

It would explain that dysbiosis lacks a universally accepted clinical definition.

And it would distinguish research associations from validated treatment recommendations.

That would make the report less impressive.

It would make it more scientific.

Conclusion

The microbiome matters.

That is not the same as saying every microbiome test matters.

Modern sequencing can describe microbial communities in remarkable detail, but the translation from sequence to diagnosis remains difficult. Different providers can generate materially different profiles from the same reference material. Stool is only one sampling compartment. Taxa vary over time. Relative abundance is not absolute quantity. Healthy reference ranges remain poorly defined. Dysbiosis indices and the Firmicutes/Bacteroidetes ratio do not currently function as universal clinical biomarkers.

Personalised microbiome medicine may become genuinely useful.

Some forms already are.

But the correct sequence is measurement, validation, prediction and then intervention.

Consumer testing often reverses that order: the report is generated first, and clinical meaning is added afterwards.

A stool sample can provide data.

It does not automatically provide a diagnosis.

---

References

1. Porcari S, Mullish BH, Asnicar F, et al. International consensus statement on microbiome testing in clinical practice. The Lancet Gastroenterology & Hepatology. 2025;10(2):154–167. https://doi.org/10.1016/S2468-1253(24)00311-X

2. Servetas SL, Gierz KS, Hoffmann D, Ravel J, Jackson SA. Evaluating the analytical performance of direct-to-consumer gut microbiome testing services. Communications Biology. 2026;9:269. https://www.nature.com/articles/s42003-025-09301-3

3. Vandeputte D, De Commer L, Tito RY, et al. Temporal variability in quantitative human gut microbiome profiles and implications for clinical research. Nature Communications. 2021;12:6740. https://doi.org/10.1038/s41467-021-27098-7

4. Allaband C, McDonald D, Vázquez-Baeza Y, et al. Microbiome 101: Studying, Analyzing, and Interpreting Gut Microbiome Data for Clinicians. Clinical Gastroenterology and Hepatology. 2019;17:218–230. https://pubmed.ncbi.nlm.nih.gov/30240894/

5. Wei S, Bahl MI, Baunwall SMD, Hvas CL, Licht TR. Determining Gut Microbial Dysbiosis: a Review of Applied Indexes for Assessment of Intestinal Microbiota Imbalances. Applied and Environmental Microbiology. 2021;87:e00395-21. https://doi.org/10.1128/AEM.00395-21

6. Magne F, Gotteland M, Gauthier L, et al. The Firmicutes/Bacteroidetes Ratio: A Relevant Marker of Gut Dysbiosis in Obese Patients? Nutrients. 2020;12:1474. https://pubmed.ncbi.nlm.nih.gov/32438689/

7. Sze MA, Schloss PD. Looking for a Signal in the Noise: Revisiting Obesity and the Microbiome. mBio. 2016;7:e01018-16. https://pubmed.ncbi.nlm.nih.gov/27555308/

8. Kedia S, et al. Human gut microbiome: A primer for the clinician. JGH Open. 2023. https://doi.org/10.1002/jgh3.12902

9. Bermingham KM, Linenberg I, Hall WL, et al. Effects of a personalized nutrition program on cardiometabolic health: a randomized controlled trial. Nature Medicine. 2024. https://www.nature.com/articles/s41591-024-02951-6

10. Hoffmann DE, Langel FD, von Rosenvinge EC, et al. Is the current regulatory framework for direct-to-consumer microbiome-based tests sufficient to protect consumers from medical, economic, and dignitary harms? Journal of Law and the Biosciences. 2025;12:lsaf024. https://pubmed.ncbi.nlm.nih.gov/41450770/

---

This article discusses the clinical interpretation of consumer microbiome testing. Persistent gastrointestinal symptoms or suspected disease require validated clinical assessment rather than interpretation of a commercial microbiome score alone.