Where the fibre ends

The official definition follows the same line: dietary fibre is carbohydrate that escapes digestion and absorption in the small intestine. EFSA also counts resistant starch and resistant oligosaccharides within it.⁵⁰ Your own enzymes cannot cut the bonds in question; gut bacteria can, because they hold an enzyme repertoire you lack — such as the GH families that break down arabinans and galactans.⁴

To calibrate against practice: the recommendation sits around 25 g/day, while actual intake in Northern Europe stalls at 16–22 g in women and 18–26 g in men.⁵⁰

The pH gradient rises longitudinally from proximal to distal. In the proximal colon pH sits around 5.7, attributable to fermentation and the short-chain fatty acids and lactate it releases; there is a negative correlation between SCFA concentration and pH. Further along, pH rises as SCFAs are absorbed or consumed and the gut wall secretes bicarbonate. Because the carbohydrates are used up proximally, the community switches distally to proteins and amino acids, with ammonia and urea as products — pushing pH up to roughly 6.7.¹

Protein fermentation therefore takes place mainly in the lower part of the colon, yielding potentially toxic metabolites — which has been proposed as an explanation for the fact that colorectal carcinoma and ulcerative colitis tend to occur precisely there.⁴⁶

Nutrition scientists accordingly argue for reclassifying fibres: the soluble/insoluble distinction is too crude to base food choices on.⁵¹ The functional difference is directly measurable — suppression of H₂S production was stronger with readily fermentable fibres such as FOS and resistant starch than with poorly fermentable ones such as psyllium and sterculia.³⁵

That also makes clear straight away what this piece is about. The relevant variable is not "how much fibre", but how much fermentable substrate arrives, and how far it travels.

The difference is structural. Inulin is a fructan with short, linear chains; the galacto-oligosaccharides in legumes are likewise short-chain. Both are directly accessible to bacterial enzymes and ferment fast, with gas as a by-product. In practice, inulin-type fibres above 8–10 g per day therefore often cause gas, bloating and loose stools.²

Acacia gum is a highly branched arabinogalactan, and it is exactly that branching that makes it ferment more slowly and mildly than linear chains. In fermentation models, replacing part of the FOS/inulin with acacia gum shifts fermentation from a peak in the proximal colon to a gradual process across the whole gut, with more even gas production — and with acidification of all colonic segments, plus lower ammonium production. The prebiotic effect remains intact throughout.³

The motif acacia and pectin share

That both fibres are tolerated comparably is no coincidence. Acacia gum is an arabinogalactan. And in its so-called RG-I or "hairy" region, pectin carries highly branched structures made up mainly of arabinose and galactose, with side chains of arabinans, galactans and arabinogalactans.⁴

So it is one molecular family, broken down by one enzyme repertoire: the glycoside hydrolase families that tackle arabinans (GH51, GH43, GH27, GH127) and GH2 for galactans, present in Bacteroides, Bifidobacterium, Ruminococcus, F. prausnitzii, R. intestinalis and Akkermansia, among others.⁴

Practically: that motif is present in ordinary food as well. RG-I is abundant in carrot, okra, tomato and potato.⁴

The effect of a single pH unit is large. In pH-controlled fermentors with gut microbiota, the proportion of Bacteroides at pH 5.5 was lower than at 6.5, with a greater share of butyrate-producing Firmicutes instead and correspondingly higher butyrate levels. At pH 5.5 fermentation turned strongly butyrogenic: butyrate reached 24–28 mM and exceeded both acetate and propionate after 150 hours.⁵

That same drop in pH prevents overgrowth of acid-sensitive species, including Enterobacteriaceae and Clostridia.⁶ And in fermentors with inulin or pectin as substrate, Faecalibacterium prausnitzii replaced the otherwise dominant Bacteroides as soon as pH was brought from 6.9 down to 5.5.⁴

The chemistry behind it: why "acetate" is a misleading name

What the bacteria form is acetic acid (CH₃COOH). It gives up a proton, and that proton lowers the pH. In the gut the equilibrium then lies far towards the conjugate base: with a pK_a of about 4.76 and a colonic pH of 5.5–6.5, roughly 85–98% is dissociated to acetate. Both terms refer to the same pair.

For the antimicrobial effect, however, only the undissociated fraction counts: only the uncharged molecule crosses a cell membrane, to dissociate inside — where pH is higher — and acidify the cytoplasm from within. That fraction runs from about 2% at pH 6.5 to about 15% at pH 5.5. The weapon sharpens itself: more acid lowers the pH, and a lower pH increases the share present in the active form.

That is exactly why it matters where the acid lands. Distally, at pH 6.7, there is almost no active form left.

Fermentation of fibre in the colon yields short-chain fatty acids plus H₂ and CO₂.⁷ Hydrogen has to be removed, because a high partial pressure thermodynamically inhibits fermentation. That is done by three guilds of hydrogenotrophs: sulfate-reducing bacteria, methanogenic archaea and acetogenic bacteria, which convert hydrogen into hydrogen sulfide, methane and acetate respectively.⁸ Most of the hydrogen does not leave the body through breath or flatus, but is cleared inside the gut itself along one of these routes.⁷

Who are they, concretely?

In the research a fixed trio serves as the representatives: Methanobrevibacter smithii for the methanogens, Desulfovibrio piger for the sulfate reducers and Blautia hydrogenotrophica for the reductive acetogens.⁹

The effect of acetogens has been shown directly: dosing rats with Ruminococcus hydrogenotrophicus (now Blautia) lowered peak hydrogen excretion by 40 to 50% after a lactulose load.¹⁰

The pecking order — and why fibre richness is a precondition

The three are not equal competitors. With sulfate as electron acceptor, sulfate reducers are the most efficient hydrogenotrophs; their hydrogen threshold lies significantly below the mean threshold of acetogens and methanogens.⁸ And M. smithii, in turn, has a lower threshold than acetogens.¹¹

Here the two storylines meet. Acidification favours precisely the clean-up crew that leaves no gas behind. A generously fed colon fermenting along its whole length creates both conditions — hydrogen abundance and low pH — under which the acetogenic route can run. Fibre richness is therefore not a vague recommendation, but the precondition for the low-gas scenario.

With one caveat from the coculture literature: in healthy adults and IBS patients, the three groups tend to co-occur rather than exclude one another.⁹ In mixed culture they initially coexisted without competing for hydrogen, after which D. piger inhibited the growth of the other two at around ten hours — for M. smithii in step with a rising sulfide concentration. The researchers stress that these interactions are condition-dependent and not readily translated to the gut.¹³

The key studies come from the Azpiroz group in Barcelona, using galacto-oligosaccharides as the prebiotic. Healthy volunteers took a low dose for three weeks. In the first phase, flatulence and gas volume after a test meal rose significantly; after three weeks both were back at baseline, while the prebiotic was still being taken. At the same time the share of butyrate producers increased, and that increase was inversely related to gas volume.¹⁴

Follow-up work exposed the mechanism: the adaptation consists of a shift in microbial metabolism towards routes that produce less gas.¹⁴ In one trial gas volume initially rose by 37%, falling back to the starting level after two weeks.¹⁵

That also explains why slow fibres land better. With inulin the symptoms are worst in the first one to four weeks¹⁶ — exactly the window in which people drop out. Acacia gum delivers the same substrate pressure without that opening peak, and is well tolerated up to 30 g/day in randomised trials, including in sensitive people.² It also carries a low-FODMAP certification, which makes it usable within a FODMAP-restricted diet.¹⁷

The same adaptation has been reproduced on pectin. On repeated dosing of carrot RG-I into a gut model over three weeks, the substrate was broken down rapidly despite its complex structure, and fermentation ran faster and more completely as dosing was repeated; in some donors the arabinan side chains were used first, correlating with an increase in Bifidobacterium longum.¹⁸

Odour ≠ volume

About 99% of gut gas is odourless: nitrogen, oxygen, hydrogen, carbon dioxide and possibly methane. The offending fraction of less than 1% consists of sulfur compounds — hydrogen sulfide, methanethiol and dimethyl sulfide.^19,20 Methane is therefore completely odourless; natural gas smells only because of an added odorant.

Levitt and Suarez quantified this: hydrogen sulfide was by far the largest sulfur component, followed by methanethiol and dimethyl sulfide, and the malodour intensity scored by raters correlated significantly with the hydrogen sulfide concentration.²¹ Together they averaged around 50 ppm per flatus.²²

Symptoms ≠ volume

Imaging studies compared gas volumes during reported bloating. Patients with a gut motility disorder did indeed have more gas. But in IBS and functional dyspepsia the gas volume was not raised — instead the diaphragm descended and the abdominal wall and gut contents shifted forward.²³

That phenomenon is called abdominophrenic dyssynergia. Normally the diaphragm relaxes and the abdominal wall contracts as gut contents increase, keeping pressure stable; in APD this inverts, producing a visibly distended belly and a feeling of fullness out of proportion to the actual amount of gas.²⁴ On top of that comes visceral hypersensitivity: the perception can be extreme at a low gas volume.²⁵

So there is a second axis of adaptation that has nothing to do with bacteria. Anyone who had painful reactions every week for years also changes at this level once those reactions stop coming.

Protein fermentation — putrefaction — is strongest in the distal colon, precisely because of protein excess and limited or absent carbohydrate availability there.²⁶ The ratio between available carbohydrate and protein is the decisive factor for substrate use: the higher the availability of complex carbohydrates at a given amount of protein, the less the microbiota uses that protein for metabolism. With ample supply, nitrogenous substrates are used only moderately, and then mainly for bacterial anabolism.²⁷

Isotopes have made this directly visible: after inulin, ¹⁵N excretion in faeces rose significantly, with a proportional fall in urine — the nitrogen was fixed into bacterial biomass instead of being absorbed as ammonia and processed by the liver.²⁸

Acid suppresses putrefaction too — by four routes

Beyond substrate availability, pH itself acts on production. Production of SCFAs and branched-chain fatty acids from peptides and free amino acids was markedly reduced at pH 5.5, with correspondingly lower net ammonia production.²⁹

1. Enzymes outside their optimum. The acidic environment lies outside the pH range of most proteases and peptidases, slowing protein and peptide breakdown.³⁰
2. Genes are suppressed. Transcription of genes for amino-acid catabolism, including deaminase genes, is inhibited.³⁰
3. The bacteria themselves disappear. The pH drop from SCFA production creates an environment unfavourable to proteolytic bacteria.³¹ The cysteine degraders — Escherichia, Salmonella, Klebsiella, Enterobacter, Clostridium, Fusobacterium³² — are largely the same acid-sensitive group that fails to overgrow at low pH.⁶ Among amino-acid fermenters generally, clostridia and anaerobic Gram-positive cocci were the predominant isolates.²⁹
4. Ammonia gets trapped. Only uncharged NH₃ crosses the gut wall; acid shifts the equilibrium towards NH₄⁺, which stays in the lumen. This is exactly the principle lactulose rests on clinically in liver failure — the mirror image of the acetic-acid mechanism above.

One honest caveat: where both factors were tested side by side, low pH and high carbohydrate availability each lowered the rate and net ammonia production from peptides — but carbohydrate proved the more important of the two in the amino-acid vessels.³³

Sulfate-reducing bacteria use H₂ as electron donor and inorganic sulfate as terminal acceptor. The main dietary sources of that sulfate are industrially processed products such as bread and sausage, alongside nuts, dried fruit, beer and brassicas.³⁴ The second route runs via desulfhydrase enzymes that break cysteine down into H₂S, pyruvate and ammonia — without hydrogen.³²

Recent data indicate that H₂S arises mainly through cysteine degradation and to a lesser extent through sulfate reducers.³² In vitro the difference is large: cysteine gave a 300-fold stimulation of free H₂S production, against a two-fold stimulation for sodium sulfate.³⁵

Important for the whole picture: fermentation actively suppresses sulfide production. It was inhibited more strongly by readily fermentable fibres such as FOS and resistant starch than by poorly fermentable ones such as psyllium and sterculia.³⁵

What diet demonstrably does

A controlled human feeding experiment exists. Five healthy men stayed in a metabolic suite and received five successive ten-day diets, with meat intake ranging from 0 g/day (vegetarian) to 600 g/day. Faecal sulfide concentration rose from 0.22 ± 0.02 mmol/kg on the meat-free diet to 3.38 ± 0.31 mmol/kg at 600 g/day, significantly related to meat intake (P < 0.001). The authors' conclusion: protein from meat is an important substrate for sulfide formation by bacteria in the human colon.³⁶

To this comes an anatomical detail that is often missed: of all amino acids, cysteine and cystine are among the worst absorbed from the small intestine.³⁷ So it is precisely the sulfur-bearing building blocks that escape digestion relatively more often. Heating makes this worse — heat-damaged protein has reduced ileal digestibility of cystine.³⁷

In legumes, methionine and cysteine are the limiting amino acids: legumes contain more lysine but a lower content of sulfur-containing amino acids, which is the classic reason to combine them with cereals.³⁹ Even in soy — the legume with the highest content — cysteine and methionine are the limiting amino acids.⁴⁰

So the legume protein that escapes digestion is actually low in the building blocks that feed route 2. On top of that it arrives together with a large load of fermentable carbohydrate, which makes the ratio from Part VI favourable.

But that is exactly where the hydrogen load sits. In Levitt's measurements, total gas per subject over four hours ranged from 106 to 1657 ml, with the largest producer making well over half a litre of hydrogen — after a meal of beans and lactulose.²² With a microbiota in which sulfate reducers control the hydrogen market, and given they have the lowest hydrogen threshold of the three,⁸ that hydrogen ends up as H₂S.

That explains why adaptation solves both problems at once. One shift — hydrogen from sulfate reducers to acetogens — lowers both the volume and the odour. The same beans, the same hydrogen, a different destination.

Alongside dietary protein, the body digests 50 to 100 g of endogenous protein each day that is secreted or shed into the lumen: proteins from saliva and gastric juice, pancreatic enzymes, mucoproteins, shed gut cells and proteins leaking in from the blood. In a standard schema, about 100 g/day comes from food and about 85 g/day from endogenous sources.⁴¹

The endogenous protein leaving the ileum consists largely of accrued mucins from the upper gastrointestinal tract that have resisted digestion.⁴² Recovery is incomplete: in pigs it has been estimated that 79% of gross endogenous secretion is reabsorbed, with recovery high for digestive enzymes and notably lower for mucin.⁴³

On the dietary side, structure and inhibitors matter. Plant proteins contain high concentrations of anti-nutritional factors and complex protein structures, so incompletely digested protein reaches the colon;⁴⁴ trypsin inhibitors are one example.⁴⁵ On a Western diet, an estimated 12 g of dietary protein escapes digestion.⁴⁶

And with that the loop closes one level deeper. When no fermentable substrate is left, bacteria turn to the only protein that is always there: the mucus layer itself. Mucin degraders such as Akkermansia normally do this modestly, but under sustained fibre deprivation that protective layer genuinely thins. Do not feed them, and they eat you.

A low-FODMAP diet lowers gas by letting less fermentable carbohydrate reach the colon. Protein intake stays the same, which shifts the ratio from Part VI. That is precisely the described state: diets high in protein and reduced in carbohydrate shift the colonic microbiome towards a potentially pathogenic, pro-inflammatory profile, with reduced SCFA production and raised concentrations of ammonia, phenols and hydrogen sulfide.⁴⁵ Prolonged strict restriction can harm microbial diversity.⁴⁷

In a randomised comparison in people with functional gut disorders, a prebiotic plus Mediterranean diet and a low-FODMAP diet both did significantly better than baseline — but with the prebiotic the effect persisted for weeks after stopping.^14,48

For completeness: psyllium is a usable middle path, because it barely ferments and in one trial even slowed the fermentation of co-ingested inulin.⁴⁹ It therefore delivers the bulk effect, but not the acidification this whole story turns on.³⁵

Bile acids are conjugated in the liver to either glycine or taurine; taurine contains sulfur. Saturated fat shifts that balance. In a study in Nature, consumption of a diet rich in saturated milk fat — but not in polyunsaturated safflower fat — promoted the expansion of the low-abundance, sulfite-reducing pathobiont Bilophila wadsworthia. The effect was mediated by milk-fat-promoted taurine conjugation of hepatic bile acids, which raises the availability of organic sulfur. Mice on a low-fat diet with added taurocholic acid — but not glycocholic acid — likewise showed a bloom of B. wadsworthia.⁵³

The enzymatic mechanism has been resolved: B. wadsworthia produces H₂S during anaerobic respiration of (bi)sulfite released from organosulfonates — including taurine, abundant in both diet and host — via the glycyl radical enzyme isethionate sulfite-lyase.⁵⁴

The dietary pattern itself also steers the conjugation: vegetarian diets favour glycine conjugation, while diets high in animal protein promote taurine conjugation.⁵⁵