Gut-microbiota-targeted diets modulate human immune status
TL;DR
In a small 17-week randomized prospective study, a high-fermented-food diet increased fecal bacterial alpha diversity and decreased multiple secondary inflammatory measures, while the prespecified cytokine-response primary outcome did not change. The intervention combined several foods at a high intake, so it cannot identify yogurt, kimchi, kefir, kombucha, or any organism as the causal component.
Design
- 36 generally healthy adults were used for full analysis (18 per arm; mean age 52 ± 11 years; 73% women).
- Participants were assigned to a high-fiber or high-fermented-food diet. Two fiber-arm participants requested nonrandom assignment; the primary-outcome analysis excluded them.
- After a three-week baseline, participants ramped intake for four weeks, maintained it for six weeks, and entered a four-week choice period.
- Fermented-food intake rose from 0.4 ± 0.6 to 6.3 ± 2.9 servings/day. Foods included yogurt, kefir, fermented cottage cheese, kombucha, vegetable-brine drinks, kimchi, and other fermented vegetables.
- Registry: NCT03275662. Participants and dietitians were not blinded.
Key findings
- The prespecified cytokine response score did not change significantly in either arm.
- In the fermented-food arm, observed amplicon-sequence variants, Shannon diversity, and phylogenetic diversity rose over time. The taxa contributing to the increase were largely not taxa detected in the foods, which argues for ecosystem remodeling or increased detectability rather than simple durable engraftment.
- The authors reported that 19 of 93 cytokines, chemokines, and related serum proteins decreased during the fermented-food intervention under their multiple-testing criteria, including interleukin-6. These were secondary/exploratory outcomes. The Methods describe a 92-analyte Olink panel with 67 proteins detected in more than 75% of samples, leaving an internal denominator discrepancy that the article does not resolve.
- The fermented-food arm reported increased abdominal distention at the end of the four-week ramp (0.06 ± 0.06 to 0.4 ± 0.1; p=0.03), but the difference from baseline was no longer significant after maintenance.
- Glucose, insulin, lipids, blood pressure, waist circumference, stress, well-being, fatigue, activity, and cognition did not change significantly.
Limitations
The sample was small, the comparison was against another active diet rather than a usual-diet control, participants and investigators were unblinded, food exposures were heterogeneous, and most immune findings were not the primary endpoint. Although the trial was registered, the statistical-analysis plan was not published in advance. The study cannot separate live organisms from food matrix or fermentation metabolites, cannot establish persistent colonization, and provides no clinical aging endpoint. gap/needs-replication gap/long-term-unknown