Why do microbes live in the gut, and does more fiber add more kinds?

Microbes, mostly bacteria, live in the gut. Fiber people cannot digest reaches the colon, and the microbes ferment it for energy. The number of kinds is not itself a health score. Surveys have still found metabolic and inflammatory markers sitting with a lower count, so diet trials treat diversity as an outcome. One trial asked that question by separating a higher-fiber diet from a higher intake of fermented foods.

Why microbes live there

Dietary fiber is poorly cut by human digestive enzymes, so it reaches the colon. Gut microbes ferment that fiber and make short-chain fatty acids. One of them, butyrate, is called a main fuel for the cells of the colon wall. When microbes already hold the space and the food, a newcomer has a harder time settling. Immune cells trade signals with the microbes that already live there and are trained not to overreact to food and to those residents.

What matters is less the head count of one microbe than whether fermentation, the gut barrier, and that resistance to newcomers keep working. When they do, bowel habit stays relatively steady, and gas or diarrhea after the same meal is less likely to repeat. When they weaken, gas, pain, diarrhea, or constipation can show up. After antibiotics thin the kinds, a toxin-making bacterium can sit in the gap. If the barrier loosens, bacterial components cross more easily, and that has been seen alongside low-grade inflammation. The association alone does not say whether the microbial mix broke first or inflammation changed the setting first.

Imbalance does not mean a ratio of good and bad microbes has slipped. Healthy people already differ with diet and age. In this note, imbalance means the number of kinds has fallen, jobs such as fiber fermentation have weakened, or microbes are growing in the wrong place. Antibiotics, a diet with almost no fiber, a gut infection, and inflammation in the gut can make that state. The survey below puts a low gene count next to body fat and an inflammatory marker at one moment in time. It is not evidence that imbalance caused those markers.

Good and bad lactic acid bacteria

Lactic acid bacteria are bacteria that ferment sugars and mainly make lactic acid. The familiar ones in yogurt and kimchi are in this group. Good and bad are not species names. Strains of the same species differ in how well they survive stomach acid, how they attach in the gut, and the signals they send to the immune system. Effects repeated in human studies belong to a specific strain, not to lactic acid bacteria as a whole.

In this note, a good lactic acid bacterium is a strain that, for that person and that purpose, has been shown to compete with a pathogen for space or to change the gut setting. There is no separate class of bad lactic acid bacteria. A strain with no evidence still does little if stomach acid kills it, and a strain that makes histamine can bother someone sensitive to that substance. When fermentation in the small intestine is already high, a lactic acid product can add gas. In critically ill people, or when immunity is severely reduced, even a usually safe live microbe has rarely been reported in the bloodstream.

Saccharomyces boulardii is often pulled into this story, and it is not a lactic acid bacterium. The papers call this microbe Saccharomyces boulardii. It is a yeast, not a bacterium. Antibacterial antibiotics do not kill it, and after people stop taking it, it leaves in the stool instead of settling. In 2015, Hanna Szajewska and colleagues pooled 21 randomized trials and 4,780 people. Among people taking antibiotics, antibiotic-associated diarrhea fell from 18.7% to 8.5% with this yeast (risk ratio 0.47, 95% confidence interval 0.38–0.57). The estimate is that 10 people need treatment to prevent one case of diarrhea (95% confidence interval 9–13). In children the rates were 20.9% to 8.8% (6 trials, 1,653 people, risk ratio 0.43, 95% confidence interval 0.3–0.6). In adults they were 17.4% to 8.2% (15 trials, 3,114 people, risk ratio 0.49, 95% confidence interval 0.38–0.63). Diarrhea linked to Clostridium difficile had a risk ratio of 0.25 in 2 trials of children (579 people, 95% confidence interval 0.08–0.73). In 9 trials of adults (1,441 people) the risk ratio was 0.8 (95% confidence interval 0.47–1.34), which is hard to call a reduction. Clostridium difficile is a bacterium that sits in the gap after antibiotics and causes diarrhea with a toxin. These numbers are not a rise in the number of microbial kinds in healthy people. The 10-week trial below did not use this yeast.

Why diversity is measured

In this note, diversity means how many kinds of gut microbes are detected and how evenly they are spread. The count of observed sequence variants is the number of kinds. The Shannon index combines that count with evenness.

Emmanuelle Le Chatelier and colleagues, in 2013, counted microbial genes in stool from 292 Danish adults, 169 with obesity and 123 without. Fewer than 480,000 genes was the low group, and the rest was the high group. The low group was 68 people, 23% of the sample. Average gene counts were 380,000 and 640,000, a difference of about 40%.

Median body fat was 37% in the low group (interquartile range 29–42) and 31% in the high group (25–39), after adjustment for age and sex (p = 0.0069, false-discovery q = 0.022). Median fasting insulin was 50 versus 44 pmol/L, median HOMA-IR was 1.9 versus 1.6, and median hsCRP was 2.3 versus 1.4 mg/L. Among people with obesity, the low-gene group had gained more weight over the previous 9 years.

The survey compares one stool sample with blood drawn at the same time. Treated as a continuous number, gene count was not significantly tied to weight or BMI (p = 0.12 and p = 0.11). Gene count is not the same measure as the sequence-variant count in the trial below. It also does not show that raising diversity improves these markers. It is one of the studies the Wastyk paper cites for why a lower count has been found alongside metabolic and inflammatory markers.

What the study looked at

Hannah Wastyk and colleagues, in 2021, split healthy adults into two groups. Eighteen people remained in each group for analysis. The results report a mean age of 51 years (SD 12), a mean BMI of 25 (SD 4), and 73% women.

One group raised fiber for 4 weeks and held that intake for 6 weeks. The other group followed the same schedule with fermented foods. The target was at least 20 g more fiber a day than usual, or at least 6 more servings of fermented foods a day, with room to eat more if it was tolerated. The fiber group was pointed to legumes, seeds, whole grains, nuts, vegetables, and fruit. The fermented-food group was pointed to yogurt, kefir, fermented cottage cheese, fermented vegetables such as kimchi, vegetable brine drinks, and kombucha. The study did not provide the food. Participants logged 3 days of intake every 2 weeks. Stool was used for the microbes and blood for immune measures. The registration number is NCT03275662.

In this trial, one serving was 6 oz of kombucha, yogurt, or kefir, 1/4 cup of fermented vegetables such as kimchi or sauerkraut, or 2 oz of vegetable brine.

Results

Fiber intake in the fiber group rose from 21.5 g/day at baseline (SD 8.0) to 45.1 g/day at the end of maintenance (SD 10.7). Fermented-food intake in that group stayed where it was. Microbial diversity, measured as observed sequence variants, the Shannon index, and phylogenetic diversity, did not change for the group as a whole. No single taxon changed in relative abundance across that whole group over time.

What changed was function. The relative abundance of 11 enzyme genes that break down plant cell-wall carbohydrates rose. The share of stool protein that came from microbes also rose (p = 0.003 versus week −2). People who ate more fiber had more carbohydrate left in stool (p = 0.0008). Butyrate did not rise. Isobutyric acid, isovaleric acid, and valeric acid fell.

The primary outcome, a cytokine response score, did not differ in either group. Inside the fiber group, the immune response did not move one way. The group whose inflammatory signaling rose and a group with more microbial types at the start differed in that starting count (p = 0.037). Total fiber eaten did not differ across those groups.

Stools became softer. The Bristol score moved from 3.3±0.3 at baseline to 4.1±0.3 at the end of maintenance (p = 0.004).

The fermented-food group went from 0.4 servings a day (SD 0.6) to 6.3 at the end of maintenance (SD 2.9). Fiber intake in that group stayed where it was. Microbial diversity rose over time (observed sequence variants p = 0.0023, Shannon index p = 0.0014). Nineteen of 93 inflammation-related proteins in serum fell, including IL-6 (false discovery rate at or below 0.05). Those 19 did not fall in the fiber group.

Among newly detected sequence variants, the share that was also found in the fermented foods peaked early at 5.4%. The rise in diversity is hard to read as the microbes in the food simply taking up residence.

Limits

The participants were healthy adults, 18 in each group. There was no control group. The comparisons were within the same person over time, and between the two diets. People chose their own foods and wrote their own logs. The intervention lasted 10 weeks. These numbers do not transfer as-is to someone with gut disease.

The primary immune result is empty. Diversity and the 19 proteins were examined after that. The hypothesis that more fiber would raise diversity does not fit these data. Carbohydrate left in stool fits the reading that the microbiota did not ferment all of the extra fiber.

The gene-count gap in the Le Chatelier survey does not mean that a change in sequence variants over these 10 weeks would move body fat or hsCRP with it. The two papers use different measures and different designs. The diarrhea rates in the Szajewska review are also not a reading of whether microbial kinds rose or fell in this diet trial. One is diarrhea in people taking antibiotics. The other is a diet in healthy adults.

What this means for a log

The part of this paper that meets a food log is whether fiber grams and fermented foods stay visible as separate lines. A yeast taken as a capsule and the fermentation already in yogurt or kimchi are not the same line. Diversity in this trial did not follow the fiber total from 21.5 g to 45.1 g. If a photo misses yogurt, kimchi, beans, or mixed grains, one of those two lines disappears from the record. Kimchi carries vegetable fiber and fermentation together, so a gram total with the fermentation missing is a different condition from the one this study split apart.