Host-diet-gut microbiome interactions influence human energy balance: a randomized clinical trial
Corbin et al. fed 17 young healthy adults (8 women) two diets in random order, each designed by the classic digestibility equations to deliver the same metabolizable energy and macronutrients: a highly digestible Western Diet (WD) and a Microbiome Enhancer Diet (MBD) built to send more food to the colon (more fibre and resistant starch, larger food particles such as whole nuts instead of nut butter, few processed foods).
Energy intake was matched in real time to whole-room calorimetry, so both diets kept participants in balance within about 5 kcal/day over the 6-day measurement. On the MBD an additional 116 ± 56 kcal/day was lost in faeces, so the body received 89.5% of the energy eaten (range 84.2–96.1%) against 95.4% (range 94.1–97.0%) on the WD.
Energy expenditure, hunger and ad libitum intake did not change; weight fell more on the MBD (−625.6 vs −134.4 g). People differed little on the WD and a lot on the MBD.
Protocol
Randomised crossover, block randomisation stratified by sex; 17 participants (8 women, 9 men; 30.8 ± 1.9 y; BMI 25.1 ± 0.52; mean ± s.e.m.); 8 had the WD first, 9 the MBD first.
Each diet: 11 days outpatient feeding then 12 days inpatient on the metabolic ward; at least 14 days washout. 100% of provided food required; adherence 99.6% vs 99.9%.
Diets designed equal in metabolizable energy and macronutrient proportions by standard digestibility equations; MBD drivers: dietary fibre, resistant starch, large food particle size, limited processed foods.
Energy intake set to whole-room calorimetry; energy balance held within the ±50 kcal target over a 6-day chamber stay.
Faecal energy as chemical oxygen demand, normalised to 24 h by a non-absorbable marker (polyethylene glycol); host metabolizable energy = intake minus faecal energy, as a share of intake. Microbiome by 16S and whole-genome sequencing; short-chain fatty acids in faeces and serum; gut transit by a pH capsule; appetite hormones over 12 h.
Key Results
| Outcome | MBD vs WD (full text) |
|---|---|
| Faecal energy | 73.0 ± 6.1 vs 32.1 ± 2.5 gCOD/day (P = 2.96 × 10⁻⁷) |
| Extra energy lost in faeces | 116 ± 56 kcal/day on the MBD |
| Host metabolizable energy (primary endpoint) | 89.5 ± 0.73% vs 95.4 ± 0.21% of intake |
| Range between individuals | 84.2–96.1% on the MBD vs 94.1–97.0% on the WD; 73–390 vs 59–185 non-metabolized kcal/day |
| Consistency | lower energy absorbed on the MBD in 16 of 17 participants |
| Energy expenditure (sleep metabolic rate) | no difference (P = 0.16) |
| Hunger, satiety, ad libitum intake | no difference |
| Body weight on the ward | −625.6 ± 196.5 g vs −134.4 ± 156.1 g (P = 0.04); fat mass −289.9 vs −64.7 g (P = 0.06) |
| What explained the spread on the MBD | faecal propionate and bacterial biomass jointly explained 58% of the variance in host metabolizable energy |
| Microbes and hormones | more bacterial biomass and short-chain fatty acids; leptin lower; pancreatic polypeptide 1.4-fold higher |
Mechanisms
Measured: more energy left in faeces on the MBD, along with more bacterial biomass (16S copy number) and more short-chain fatty acids in faeces and serum.
Modelled (authors' in silico model): microbial biomass made up more than 25% of faecal energy on both diets; the host absorbed about twice as much energy as microbial short-chain fatty acids on the MBD, but this was offset by the extra biomass lost.
Authors' interpretation: the gut microbiome shapes how much energy the host gets from food; they state they cannot establish whether the change is caused by the diet, the microbes, or both.
Inference, not tested: the MBD bundled four changes (fibre, resistant starch, particle size, less processing), so the study does not say which one matters for an athlete's food choices.
Practical Application
The same calories from whole, high-fibre food give the body less energy: a diet built to feed the gut microbes lost an extra 116 ± 56 kcal/day in faeces, so 89.5% of intake was metabolizable against 95.4% on a Western diet, in 17 adults held in energy balance (B+).The diets were designed equal by classic digestibility equations, yet the whole-food diet delivered about 6 percentage points of intake less than the Western diet (95.4 − 89.5). The equations' absolute error for each diet was not reported.
The same food gives different people more different energy on a whole-food diet than on a processed one: 84.2–96.1% metabolizable on the MBD vs 94.1–97.0% on the WD. The widest individual gap, 73 to 390 kcal/day, sustained over weeks, is not trivial next to the 4.3–13.3 kg spread from one identical surplus in REF-1127 (cross-study comparison, not tested).
For an athlete in a hard block: a fibre-rich whole-food diet is likely over-counted by labels and apps too (inference from the designed-diet result), so a log that "should" cover the training can leave a real shortfall; watch body mass, performance and recovery, and lean on carbohydrate targets (SYNTH-69) rather than adding calories by formula (SYNTH-87).
Not a reason for microbiome tests or for dieting with fibre in a build block: the effect came from a whole designed diet, and hunger and ad libitum intake did not rise to compensate — which in a heavy training block is the risk, not the benefit (inference).
Limitations
n = 17 young healthy adults, mostly Black (11 of 17); no athletes; habitual fibre was low (7.6 g/1000 kcal/day), so the MBD was a large change for them.
23 days per diet; long-term adaptation of the microbiome and appetite not tested.
Four dietary drivers combined; the effect of each is unknown, and the exact fibre doses are in the supplement, not the text.
Faecal energy measured as chemical oxygen demand rather than bomb calorimetry (the authors report R² = 0.97 between the two for foods).
The microbial species linked to the spread were not reproduced by a second statistical method and are hypothesis-generating.
How this entry was checked and changed
Why this level
controlled crossover, n = 17 (rubric: crossover ≥15 = B+), primary endpoint pre-registered (NCT02939703), whole energy balance measured; read from the full text
Full text (checked 2026-10-13)
MBD vs WD: an additional 116 ± 56 kcals lost in feces daily; host metabolizable energy 89.5 ± 0.73% (range 84.2-96.1%) vs 95.4 ± 0.21% (range 94.1-97.0%) (full text).
Faecal energy 73.0 ± 6.1 vs 32.1 ± 2.5 gCOD/day (full text).
The range translates to 73–390 non-metabolized kcals/day on the MBD vs 59–185 kcals/day on the WD (full text).
Lower energy absorbed on the MBD in 16 out of 17 participants; no diet difference in sleep metabolic rate (P = 0.16) (full text).
Weight −625.6 ± 196.5 g on the MBD vs −134.4 ± 156.1 g on the WD (P = 0.04); fat mass −289.9 ± 97.30 g vs −64.7 ± 84.6 g (P = 0.06) (full text).
Faecal propionate and biomass jointly explain 58% of the variance in host metabolizable energy on the MBD (full text).
Energy balance within ±50 kcals per 6-day calorimeter stay (WD 4.1 ± 5.1 kcal/day; MBD 5.4 ± 2.8 kcal/day) (full text).