Dr. Axe Wire · Study decoded · gut

Gut Bacteria And Fecal Metabolites Flag Mild Cognitive Impairment In 35 Adults

A stool test combining gut bacteria and metabolite markers distinguished people with mild cognitive impairment from healthy controls, most of them spouses, and flagged who declined fastest over a year.

A fork lifting pale shredded fermented vegetables over a plate with green onion pieces.
Photo: Pexels

A fecal microbiome and metabolite signature distinguished people with mild cognitive impairment from cognitively healthy controls, 16 of them spouses, with an accuracy score (AUC) of 0.944, according to a study published October 6, 2026 in MedComm.

35adults, 16 spouse pairs
1-yearcognitive follow-up

Researchers at Cardiff University and Imperial College London studied 35 older adults, 18 with mild cognitive impairment (MCI) and 17 cognitively healthy controls, including 16 couples living in the same household, publishing in MedComm on October 6, 2026. Compared with controls, MCI cases had higher levels of a bacterial group called Izemoplasmatales and the amino acid proline, with lower levels of Ruminococcus, Lachnospira and fumaric acid. A model combining three gut bacteria types and two metabolites separated MCI cases from controls with an accuracy score (AUC) of 0.944, versus 0.899 for bacteria alone and 0.861 for metabolites alone. In an outside group of 30 MCI patients and 30 controls, a two-marker version of the bacterial signature still scored 0.825. Among 17 MCI patients tracked for one year, 6 rapid decliners had higher Bilophila and lower Coprococcus than 6 slow decliners, plus higher fecal arginine, aspartic acid and hypoxanthine, with these three metabolites predicting rapid decline at an accuracy score of 0.956, versus 0.867 for the bacteria-only model.

The authors tie several of the changed microbes to disrupted amino-acid breakdown pathways, including increased arginine and ornithine degradation, which they link to the cell's energy cycle and say has been associated with Alzheimer's disease in earlier research. As general background, the gut and brain stay in constant two-way contact: gut microbiota, the community of microbes in the intestines, can produce byproducts that affect inflammation and nerve signaling, which is the physiological backdrop for why stool chemistry might track with brain health.

Advertisement20% off Ancient Nutrition Multi Collagen Protein, code MCP20

Diagnostic Accuracy: Combined Gut and Metabolite Model vs Metabolites Alone

Diagnostic Accuracy: Combined Gut and Metabolite Model vs Metabolites Alone35 adults (18 MCI, 17 controls), AUC scores. Combined microbiome + metabolite model: 0.944 AUC; Metabolites alone: 0.861 AUC35 adults (18 MCI, 17 controls), AUC scores
0.944 AUC
COMBINED MICROBIOME + METABOLITE MODEL
0.861 AUC
METABOLITES ALONE
MedComm, 2026, Zhang et al.

The cohort was small, 18 cases and 17 controls for diagnosis and just 6 rapid versus 6 slow decliners for progression, which can inflate how accurate these scores look compared with a larger group. This is an observational study, so the bacteria and metabolite changes were linked to MCI and its pace of decline, not shown to cause it. The study was funded by the Rosetrees Trust with infrastructure support from the NIHR Imperial Biomedical Research Centre, and the authors declare no conflicts of interest.

Dr. Axe’s Take

The gut-brain link shown here makes sense given how much gut bacteria influence inflammation and nerve signaling between gut and brain. Ruminococcus and Lachnospira were lower and Izemoplasmatales was higher in the MCI group, a pattern worth watching rather than acting on with a stool test that doesn't exist commercially yet. What I'd do this week is build gut microbial diversity directly: fermented foods like kefir and sauerkraut, varied fiber from vegetables and legumes, and polyphenol-rich foods like berries and olive oil. If a parent or spouse shows early memory changes, that's reason enough to prioritize gut health now.

Dr. Axe— Dr. Axe

Study summary

Who
35 older adults (18 with mild cognitive impairment, 17 cognitively healthy controls), including 16 cohabiting spouse pairs, all White ethnicity, average age 74-76
Design
Cross-sectional fecal microbiome and metabolomics comparison plus 1-year prospective cognitive follow-up in MCI cases, with the bacterial diagnostic signature tested in one independent external cohort
Dose
Not applicable (observational biomarker study, no intervention)
Primary result
Combined gut microbiota and metabolite model distinguished MCI from controls with an accuracy score (AUC) of 0.944 (95% CI 0.861-1.0), versus 0.899 for microbiome alone and 0.861 for metabolites alone
Secondary
Metabolite signature (arginine, aspartic acid, hypoxanthine) predicted rapid cognitive decline over 1 year with an AUC of 0.956 in 6 rapid versus 6 slow progressors; bacterial signature (Bilophila, Coprococcus) scored 0.867 for the same prediction; external validation cohort of 30 MCI and 30 controls gave an AUC of 0.825 for a 2-marker bacterial signature
Funding / conflicts
Funded by the Rosetrees Trust (Project Grant Reference A2226); NIHR Imperial Biomedical Research Centre provided infrastructure support; authors declare no conflicts of interest
Date
Oct 6, 2026
AdvertisementAncient Nutrition Active Peptides
AdvertisementAncient Nutrition Active Peptides