Structure of dietary fiber and gut microbiota: what Feng's review shows

The main finding is that structural properties help explain why two fibers with the same general label can follow different paths. A less accessible structure can slow down degradation; a more exposed or less compact one can facilitate contact with microbial enzymes and, therefore, fermentation in the colon.

The authors emphasize that solubility, viscosity, and fermentability are related but not equivalent properties. Viscosity indicates how much a fiber thickens a liquid; fermentability, on the other hand, indicates how well the gut microbiota can utilize it. Crystallinity and chemical cross-linking can limit access to the fiber, while pores, branching, and size reduction can modify it. The practical significance is mechanistic: these elements can influence the metabolites produced by bacteria, but they do not automatically demonstrate a clinical benefit.

This review is relevant because it avoids a common oversimplification: treating all fibers as if they produce the same intestinal response. Instead, it proposes considering composition, physical structure, processing, and the baseline gut microbiota together.

This framework also helps in interpreting experimental results. If the fiber structure changes, the bacteria capable of utilizing it and the short-chain fatty acids—molecules produced during fermentation—may also change. However, a microbial change is not, in itself, a prognostic biomarker (i.e., an indicator of the future course of a disease) nor a predictive biomarker (i.e., an indicator of response to a treatment). Targeted clinical studies are needed to link these steps to health outcomes.

Feng and colleagues (2025) did not assign fibers to participants nor measure specific clinical outcomes. The publication is a narrative review and brings together evidence with different designs, doses, foods, animal species, and laboratory conditions. Therefore, it does not estimate an average effect or a single causal relationship.

It does not demonstrate that modifying porosity, crystallinity, or branching improves health in all people. It does not establish which fiber is best in absolute terms and does not allow for the direct transfer of data obtained in test tubes, in animals, or with isolated fibers to a whole plant-based food. The initial gut microbiota, the overall diet, and the dose can also change the response. Independent evidence clarifies how much the narrative is supported and how much it depends on the context.

A randomized clinical trial by Deehan and colleagues compared three type 4 resistant starches, which are starches modified to resist digestion. In healthy adults, small structural differences produced distinct changes in the gut microbiota and a relative prevalence of either propionate or butyrate; the dominant effects increased with the dose up to a plateau at 35 g per day [1]. The confirmed pathway is fiber structure → selective utilization by bacteria → fermentation profile. The confirmation is direct and relatively strong for those three ingredients, but it does not equate to proof for all fibers or clinical outcomes.

In a fermentation model using fecal microbiota from different subjects, Cantu-Jungles and colleagues observed that more complex fibers utilized by fewer bacteria generated responses that were more similar across donors compared to low-specificity fibers [2]. With insoluble beta-glucan, Anaerostipes increased on average from 0.3% to 16.5% and Bacteroides uniformis from 2.5% to 17.9%; the variations were associated with butyrate and propionate ratios. The confirmation is direct for the principle of selectivity, but the model does not replicate the human organism.

Jensen and colleagues fermented 55 fibers for 72 hours with pooled feline fecal inoculum. The monosaccharide composition was associated with final pH, short-chain fatty acids, and microbial community [3]. This supports the chemical composition → fermentation pathway, but it is indirect confirmation because it concerns an in vitro system and non-human microbiota.

Evidence does not indicate that fiber structure alone determines a predictable response. In a crossover study of 76 healthy adults, with 59 participants completing the protocol, Devarakonda and colleagues compared type 2 resistant starch, type 4 resistant starch, and digestible starch for ten days each. The baseline gut microbiota, habitual dietary fiber intake, and the order of interventions helped predict some responses; furthermore, short-chain fatty acid concentrations increased more with digestible starch [4]. This reduces the generalizability of a linear relationship between structure and metabolites, without denying it.

Moncada and colleagues also compared 16 fibers in vitro using healthy gut microbiota and four fibers using microbiota associated with ulcerative colitis. In the latter case, for all four selected fibers, fermentation, gas, and short-chain fatty acids were lower [5]. This finding indicates that the initial microbial community can limit the effect expected from the structure. This limitation is significant: the comparison with ulcerative colitis is derived from a single donor. No sufficiently relevant directly conflicting evidence beyond these contextual conditions has been identified; this does not prove the absence of contrary results.