THE ARISING · Longevity & Healthspan
At-home gut tests gave different answers from the same stool
Seven services analyzed the same reference sample. The findings challenge repeatability, not the value of studying the microbiome.
The same reference material sent to seven commercial gut-testing services produced reports that differed, sometimes even across kits from the same provider.
That is the central result of a laboratory study in Communications Biology, led by Stephanie Servetas and colleagues at the US National Institute of Standards and Technology (NIST). It is an analytical test of the products, not a study of health outcomes in people. The team ordered three kits from each of seven direct-to-consumer services, 21 kits in all, and loaded every kit with a homogenized pool of human fecal material that NIST developed as a reference standard. Because the material was uniform, any differences in the reports came from the testing, not the sample.
The companies were not told about the assessment until all the final reports had come back, between 2 and 8 weeks after shipping. The authors report major discrepancies both across providers and within them. Their key finding: for many of the shared groups of microbes they examined, differences among methods were comparable to differences between donors. To make that comparison, NIST also analysed material from seven other donors with its own method. In effect, the choice of company changed the picture about as much as swapping one person's gut for another's.
The reports differed in basic ways. Some companies sequenced a single marker gene, the 16S rRNA gene. Others read all the DNA in the sample, a method called whole metagenome shotgun sequencing. One also looked for fungi. The number of sequence reads ranged from under 20,000 to more than 10 million. More reads did not mean more species found. Company C reported over 600 species from between 20,000 and 30,000 reads. Company D reported 95, 98 and 99 species across its three kits, each with more than 10 million reads.
The cutoffs for reporting a microbe also varied widely, from 0.1% of the sample down to 8 × 10⁻⁶ %. Collection instructions differed too: whole bowel movement or used toilet paper, different sampling tools, and a buffer liquid in some kits but not others. No kit included ice, so every sample travelled at ambient temperature. The authors attribute the differences to methodological variability and a lack of sufficient quality control. They add that reliable measurement is a prerequisite for sound clinical recommendations, and they call for standards.
The study tested the consistency of reports for one standardized specimen. It did not test whether following a company’s diet or supplement advice improves health. Improving measurement would be a prerequisite for answering that separate question.