Dietary Index for Gut Microbiota: Evidence from 197,000 Participants

What is DI-GM and how it works

The Dietary Index for Gut Microbiota (DI-GM) is a composite score developed through a systematic review of 121 interventional and observational studies on the effects of specific foods on the composition of the gut microbiota in adults. The index is not tied to specific bacterial species — it reflects the extent to which a diet as a whole supports the diversity and functionality of the gut microbiome.

The final index comprises 14 components: 10 beneficial and 4 harmful. Each beneficial component earns 1 point when consumed at or above the sex-age median intake. Each harmful pattern earns 1 point when avoided (either consumption below the median for harmful foods, or fat share below 40% of energy). Total range: 0–14 points.

Beneficial components (1 point each):

• Chickpeas — fermentable prebiotic, source of inulin and resistant starch
• Soy (tofu, edamame, soy milk) — isoflavones and PUFAs with probiotic properties
• Avocado — monounsaturated fats and fibre; lowers the Firmicutes/Bacteroidetes ratio
• Broccoli and other cruciferous vegetables — glucosinolates, indole-3-carbinol
• Fermented dairy products (yoghurt, kefir) — live Lactobacillus and Bifidobacterium
• Coffee — polyphenols (chlorogenic acid), lowers pH in the large intestine
• Dietary fibre (soluble and insoluble) — primary substrate for butyrate-producing bacteria
• Cranberries — proanthocyanidins; inhibit pathogen adhesion to the epithelium
• Whole grains (oats, barley, brown rice) — beta-glucan, arabinoxylan
• Tea (green, black) — catechins; stimulate growth of Bifidobacterium

Harmful patterns (1 point for avoiding each):

• Refined grains — absorbed rapidly in the small intestine, barely reaching the microbiota
• Red meat — high saturated fat content; fermented to produce TMAO and hydrogen sulphide
• Processed meat (sausages, bacon, hot dogs) — nitrites, preservatives and high fat content disrupt intestinal pH
• Diet with fat share >=40% of daily energy — reduces microbiota diversity regardless of fat source

Two 2026 cohort studies: what the data show

The first study was published in Endocrinology and Metabolism in July 2026. The authors analysed two cohorts simultaneously: UK Biobank (170 575 participants, mean age 56.05 years, 53.9% women, median follow-up 12.44 years; 8 082 all-cause deaths, 875 CVD deaths) and NHANES (26 651 participants, mean age 48.04 years, median follow-up 9.25 years; 2 863 all-cause deaths, 459 CVD deaths).

Results (top DI-GM quartile vs. lowest, multivariable-adjusted models):

• All-cause mortality — UK Biobank: HR 0.86 (95% CI 0.81–0.92), p < 0.001
• All-cause mortality — NHANES: HR 0.79 (95% CI 0.70–0.89), p < 0.001
• CVD mortality — UK Biobank: HR 0.79 (95% CI 0.66–0.95), p = 0.013
• CVD mortality — NHANES: HR 0.68 (95% CI 0.51–0.90), p = 0.007

The authors also conducted mediation analysis using plasma metabolomics: 52.6% of the association between DI-GM and all-cause mortality was explained by a metabolomic profile of 124 plasma metabolites. This points to a measurable biological pathway from diet to outcome via the gut microbiota and its metabolic products.

Second cohort: dose-response per point

The second study was published in the Journal of Nutrition, Health & Aging in 2026. The authors used NHANES 2007–2018 data with additional mediation analysis through biological age (15 810 participants, mean age 47.0 ± 15.7 years, follow-up to 2019).

Key results:

• Continuous variable: each 1-point increase in DI-GM — HR 0.90 (95% CI 0.86–0.93) for all-cause mortality. This corresponds to a 10% risk reduction per additional point out of 14.
• Categorical comparison: DI-GM >=6 vs. 0–3 points — HR 0.64 (95% CI 0.53–0.77), a 36% risk reduction.

Mediation analysis showed that part of the association between DI-GM and mortality operates through biological age (slowing of cellular ageing). This is a distinct mechanism from the metabolomic one identified in the first study — both indicate that the influence of DI-GM is not limited to one or two pathways.

Mechanism: why a microbiota-supporting diet is associated with mortality

Observational cohort data do not establish causality, but the biochemical mechanisms are well characterised in interventional studies. The beneficial components of DI-GM act through several pathways.

Short-chain fatty acids (SCFAs). Fermentation of fibre, inulin and resistant starch by bacteria of the genera Faecalibacterium prausnitzii, Roseburia and Bifidobacterium produces butyrate, propionate and acetate. Butyrate is the primary energy source for colonocytes, and also a signalling molecule that reduces systemic inflammation via NF-kB inhibition. SCFA levels directly correlate with inflammation markers (CRP, IL-6) in a number of interventional studies.

Gut barrier. Impairment of the epithelial barrier function (intestinal permeability) allows bacterial endotoxins (lipopolysaccharides, LPS) to enter the systemic circulation and trigger chronic inflammation. Avocado, whole grains and fermented products reduced plasma LPS-binding protein — an indirect marker of intestinal permeability — in several RCTs.

Microbiota diversity and trimethylamine N-oxide (TMAO). Red and processed meat are rich in choline and carnitine — substrates for bacterial synthesis of trimethylamine, which is then oxidised in the liver to TMAO. Elevated plasma TMAO is consistently associated with atherosclerosis and CVD risk in several prospective cohorts. A high DI-GM diet reduces the substrate load on this pathway.

None of these mechanisms alone fully explains the findings, and their combined contribution to mortality outcomes cannot be precisely established from observational data alone.