The role of cereal soluble fiber in the beneficial modulation of glycometabolic gastrointestinal hormones

Abstract

According to cohort studies, cereal fiber, and whole-grain products might decrease risk for type 2 diabetes (T2DM), inflammatory processes, cancer, and cardiovascular diseases. These associations, mainly affect insoluble, but not soluble cereal fiber. In intervention studies, soluble fiber elicit anti-hyperglycemic and anti-inflammatory short-term effects, partially explained by fermentation to short-chain fatty acids, which acutely counteract insulin resistance and inflammation. ß-glucans lower cholesterol levels and possibly reduce liver fat. Long-term benefits are not yet shown, maybe caused by T2DM heterogeneity, as insulin resistance and fatty liver disease – the glycometabolic points of action of soluble cereal fiber – are not present in every patient. Thus, only some patients might be susceptive to fiber. Also, incretin action in response to fiber could be a relevant factor for variable effects. Thus, this review aims to summarize the current knowledge from human studies on the impact of soluble cereal fiber on glycometabolic gastrointestinal hormones. Effects on GLP-1 appear to be highly contradictory, while these fibers might lower GIP and ghrelin, and increase PYY and CCK. Even though previous results of specific trials support a glycometabolic benefit of soluble fiber, larger acute, and long-term mechanistic studies are needed in order to corroborate the results.

Keywords:

• Cereal fiber
• GLP-1
• GIP
• ß-glucans
• incretins
• soluble dietary fiber

Introduction

Case numbers for type 2 diabetes mellitus (T2DM) are rising steadily all over the world. Even though the apparent modifiable reasons of insufficient physical activity and poor quality of food are well known, recommendations for an improved lifestyle seem to fail in most patients. Given the behavioral inertia to initiate regular exercise, to reduce caloric intake and to increase food variability in our modern world, some impulses for a healthy living could be introduced from the food industry rather than their consumers. An important component of a healthy diet are insoluble and soluble dietary fibers, the latter of which are a heterogeneous, but well investigated group of indigestible carbohydrates found in vegetables, fruits, and also grains and cereals.

In our following review, we first highlight the overall and in particular metabolic health benefits of soluble cereal fiber, both from an epidemiological and interventional point of view. In general, soluble cereal fiber improves insulin resistance and glucose tolerance, accommodated by an amelioration of hypercholesterinemia and systemic inflammation. These effects are at least partially independent of weight loss, indicating specific mechanisms working beyond the gastrointestinal lumen.

We then also summarize the interaction of specific human genes with metabolic outcomes related to the intake of dietary fiber. Surprisingly, several of these nutrigenetic associations are linked to incretins and other gastrointestinal hormones, such as glucagon-like peptide 1 (GLP-1), glucose-dependent insulinotropic peptide (GIP), peptide YY (PYY), ghrelin, and others. This group of endocrine players is well-known for its impact on insulin secretion, eating behavior and inflammation.

Therefore, we conduct a detailed review of randomized controlled trials (RCT) investigating the effect of soluble cereal fiber on glycemia and insulin response, followed by an overview over the reported effects on the release of incretins and similar endocrine signals. This is done for trials on acute effects, in preload studies and in long-term RCTs.

What is the general benefit of cereal fiber and whole-grain in cohort studies?

The type 2 diabetes mellitus epidemic is growing world-wide, with many countries still experiencing an uprise of the unhealthy lifestyle pattern, combining insufficient physical activity and hypercaloric diet. Despite higher standard of living, this leads to a rising prevalence of cardiovascular disease, cancer and other long-term complications, contributing to premature morbidity and death. As T2DM is the common precursor of these outcomes, large prevention trials have investigated, if it can be prevented by lifestyle modification. In randomized controlled trials (RCT) a 40–60% risk reduction by lifestyle treatment was reported (Pan et al. 1997; Ramachandran et al. 2006; Knowler et al. 2002; Tuomilehto et al. 2001). The Da Qing 30-years follow-up indicated a benefit in micro- and macrovascular morbidity and mortality (Gong et al. 2019).

Most patients with T2DM show insulin resistance, which is strongly associated with hypercaloric food intake and obesity as well as lacking physical activity. Besides these major pathomechanistic factors, specific dietary aspects, such as saturated fat, simple carbohydrates and excessive intake of alcohol contribute to the metabolic risk, even in the absence of weight gain. Some other nutritional factors have been identified as protective agents: moderate alcohol intake, coffee consumption and cereal fiber from whole-grain products (Ley et al. 2014; Weickert and Pfeiffer 2008; Aune et al. 2013). Consumption of whole-grain products is also associated with lower all-cause mortality (Ma et al. 2016) and mortality by cardiovascular (Li et al. 2016), neoplastic, infectious or respiratory diseases (Aune et al. 2016), but not by stroke or diabetes (Zhang et al. 2018; Benisi-Kohansal et al. 2016). Similar inverse associations besides mortality are found for morbidity: NAFLD (Xia et al. 2020), coronary heart disease (Tang et al. 2015), ischemic (but not hemorrhagic) stroke (Chen et al. 2016; Deng et al. 2018), colorectal, breast, and pancreatic cancer (Lei et al. 2016; Reynolds et al. 2019). These findings are supported by inverse relations between whole-grain intake and levels of LDL cholesterol (Hollaender, Ross, and Kristensen 2015) and obesity (Schlesinger et al. 2019). Surprisingly, soluble fruit and vegetable fiber has not been linked to a lower risk for T2DM (Aune et al. 2013; Ley et al. 2014; Weickert and Pfeiffer 2018). A discrimination of insoluble and soluble fiber within whole-grain products has not been done in cohort studies.

What is the interventional evidence on metabolic benefits of cereal fiber?

The strong benefit of cereal fiber, which was shown in epidemiological studies, has mainly been supported by interventional trials assessing whole grain of certain types and ß-glucans as a particular soluble fiber, which is found in whole grain. Interventional data for insoluble cereal fiber such as cellulose, hemicellulose are sparse and currently cannot mirror the effects of cohort studies on this topic.

Meta-analyses on RCTs testing whole-grain products are limited by study design: not a single study reported endpoint outcomes such as cardiovascular events or T2DM onset. Also, whole-grain treatment did not significantly affect BMI, blood pressure, triglycerides, cholesterol and fasting glucose (Maki et al. 2019; Kelly et al. 2017), but moderate effects for these surrogate parameters were seen in studies specifically addressing dietary fiber (Reynolds et al. 2019). There is also a trend for lower HbA1c and insulin resistance when investigating dietary fiber in RCTs (Maki et al. 2019). This inconsistency may result from the variety of grains. Barley and oat are rich in soluble fiber such as ß-glucans, which have been shown to be effective in lipid reduction, while other grains – and thus their predominantly insoluble fiber fraction – fail to demonstrate an effect on blood lipids (Hui et al. 2019).

However, whole-grain interventions do show benefits. RCT meta-analyses report significantly lower postprandial glucose and insulin excursions in acute trials (Marventano et al. 2017), and a trend for a similar benefit in medium- or long-term trials (Reynolds et al. 2019). There is also evidence for an anti-inflammatory benefit of whole-grain interventions on CRP and IL-6, but not TNF-alpha (Xu et al. 2018). Once again, one can only speculate, which grains or fibers are driving the effect, as not all means of application are tested consistently in the majority of trials. For example, long-term benefits of whole-grain on postprandial metabolism were seen with rice, but not wheat and rye, once again highlighting the heterogeneity of grains and their bioactive compounds (Musa-Veloso et al. 2018). Most certainly, this heterogeneity is explained by differences in proportions of certain types of dietary fiber, but most fiber trials are small and short, lacking statistical power (Maki et al. 2019).

Long-term trials with considerable statistical power are only available on insoluble fiber. The “Protein and Fiber in Metabolic Syndrome” study (ProFiMet; n = 111) demonstrated improvements of insulin sensitivity by supplementation with an insoluble oat fiber blend after 6, but not after 18 weeks. Insulin resistance, caused by a high-protein diet, was mitigated by fiber additive (Weickert et al. 2011). The Optimal Fiber Trial on Diabetes Prevention (OptiFiT, n = 180) was a 2-year diabetes prevention trial using the same fiber supplement. It showed moderate improvements in 2-h glucose, HbA1c and leucocyte count in the entire cohort, but especially in metabolic subgroups (IFG-IGT and obese subjects, respectively) (Honsek et al. 2018; Kabisch et al. 2019a; Kabisch et al. 2019b). Similar subtype-specific responses to fiber-rich diets had been reported previously, indicating that patients with higher fasting glucose levels are more susceptible to these interventions (Hjorth et al. 2018; Hjorth et al. 2019).

In the past few years, the clinical heterogeneity of T2DM has been investigated more thoroughly, indicating that obese, severely insulin-resistant subjects comprise a distinct phenotype among other groups of patients with a different risk profile (Stidsen et al. 2018; Ahlqvist et al. 2018). Cereal fiber might be beneficial for some of those phenotypes in particular, but not necessarily all of them. Inflammatory processes in obesity, insulin resistance, and NAFLD might be responsive to fiber intakes, while primary insulin deficiency may resist this kind of intervention. The metabolic amelioration in the “classical textbook phenotype” of T2DM could be achieved by changing the secretion pattern of gastrointestinal (GI) peptide hormones. Soluble cereal fiber might modulate incretin action and interact with their nutrigenetic background.

Nutrigenetic associations between cereal fiber and diabetes risk genes

Fiber intake is connected to obesity and T2DM risk genes by a nutrigenetic interaction, involving a variety of loci without clear pathomechanistic linkage. Obesity-related SNPs in FTO, SH2B1, and other genes modulate anthropometric and glycemic traits, parallel to differences in the degree of fiber intake (Steemburgo et al. 2013; Perez-Diaz-del-Campo et al. 2020; Vimaleswaran et al. 2016; Villegas et al. 2014).

Some other nutrigenetic associations are additionally related to the incretin system. The rs7903416 and the rs12255372 polymorphisms (TCF7L2, a strong diabetes risk gene) increase the risk for T2DM, partially by impairing incretin secretion (Lyssenko et al. 2007). rs10923931 and rs4457053 reside in the NOTCH2 and ZBED3 genes. Carriers of the T allele of rs10923931 (NOTCH2) show a higher T2DM risk and lower fasting insulin levels (Chen et al. 2014; Gupta et al. 2012). rs4457053 (ZBED3) has been shown to affect glycemic traits in children (Carayol et al. 2020). Both risk alleles in ZBED3 and NOTCH2 are also linked to higher glucagon levels, which might indicate involvement of GIP or CCK, the two nutrient-dependent hormones that increase postprandial glucagon levels (Jonsson et al. 2013). TCF7L2, ZBED3, and NOTCH2 are also involved in the WNT pathway, which is connected to insulin sensitivity and was shown to be activated by SCFAs (Abiola et al. 2009; Bordonaro, Lazarova, and Sartorelli 2007).

While subjects with the protective genotype of the TCF7L2 SNPs show an additional beneficial association of high fiber intake, carriers of the risk allele have a higher T2DM risk and HbA1c, potentiated by high fiber conditions (Hindy et al. 2012; Hindy et al. 2016). Contrary to that, subjects with the T2DM risk genotypes of NOTCH2 (T+) and ZBED3 (GG) benefit from high fiber intake, while those with generally lower genetic T2DM risk do not (Hindy et al. 2016).

Rs10923931 in NOTCH2 and rs780094 in GCKR are among the very few T2DM risk genes, sharing associations with both glycemic and lipid traits (Stančáková et al. 2011). Apart from that, rs780094 is additionally linked to NAFLD risk (Zain, Mohamed, and Mohamed 2015). Carriers of the TT genotype of rs780094 show lower insulin levels with higher fiber intake, while C allele carriers show higher fasting insulin, independent of fiber intake (Nettleton et al. 2010).

Gastrointestinal peptide hormones in the regulation of glucose metabolism

Apart from collecting and digesting food and absorbing nutrients, a major task of the gastrointestinal tract is the hormonal (self)-regulation of digestion and metabolic fate of nutrients. Specialized cells within the single-cell layer of stomach and intestine produce a variety of peptide hormones, which affect the activity of smooth muscle cells, exocrine glands and other endocrine signaling pathways. Among these hormones, the incretin group – represented by glucagon-like peptide 1 (GLP-1) and glucose-dependent insulinotropic peptide (GIP) – elicits the most relevant metabolic effects.

GLP-1 catalyzes insulin secretion in the presence of elevated or rising glucose levels (Meloni et al., 2013), but also suppresses glucagon release (Ramracheya et al. 2018). Furthermore, GLP-1 promotes satiety signals in the hypothalamus and reduces the rate of ß-cell apoptosis (Turton et al. 1996; Costes, Bertrand, and Ravier 2021; Cornu et al. 2009). In response to all passing digestible macronutrients, GLP-1 is released from L cells, which reside in the lower parts of the intestine (Lindgren et al. 2011). GLP-1 secretion is influenced by a multitude of other nutritional factors, among which fiber has been by far more intensively studied in rodents than humans (Bodnaruc et al. 2016). Other factors affecting GLP-1 levels are sex (Adam and Westerterp-Plantenga 2005a), body weight (Adam and Westerterp-Plantenga 2005b), weight loss (Adam Jocken, and Westerterp-Plantenga 2005), and glycemic state. For a decade now, GLP-1 analogues are an important part of antidiabetic pharmacotherapy.

Similar to GLP-1, GIP stimulates ß-cells to secrete insulin given supraphysiological glucose levels and inhibits ß-cell apoptosis (Widenmaier et al. 2009). Contrary to GLP-1, it does not possess the potential to induce satiety (Edholm et al. 2010; Hayes et al. 2021) and increases glucagon secretion (El and Campbell, 2020). GIP secretion derives from another type of enteroendocrine cells, the K cells, which are located in the upper parts of the gut and are exclusively stimulated during passage of saccharides, lipids and amino acids (Reimann et al. 2020). GIP secretion has an impact on postprandial nutrient distribution by modulating splanchnic blood circulation (Koffert et al., 2017). Especially in scenarios with elevated energy intake this may lead to an increase of inflammatory processes, ectopic lipid storage (liver fat and visceral fat) and insulin resistance (Keyhani-Nejad et al. 2015; Gögebakan et al. 2015; Pfeiffer and Keyhani-Nejad 2018; Keyhani-Nejad et al. 2020). In rodents, inhibition of GIP signaling counteracts obesity and NAFLD (Isken et al. 2008; Pfeiffer et al. 2010). GLP-1-GIP-glucagon co-agonists are tested as antidiabetic drugs (Knerr et al. 2020).

Some other GI peptide hormones are secreted in response to passing nutrients, but lack a direct metabolic effect. However, these endocrine players are strong satiety signals. Peptide YY_1-36 (PYY) is released from the L cells in the ileum and colon (which are also the source of GLP-1), but also in other parts of the GI tract. PYY reduces gastric motility and emptying, which supports nutrient absorption in the gut (Savage et al. 1987). On the other hand, PYY inhibits the exocrine secretion from the gastric wall and the pancreatic acinus. Dietary fat and other chymus components are are the main stimuli (Onaga et al. 2002). Enzymatic cleavage of two amino acids leads to the antagonistic residue PYY_3-36.

There are more hormones, which affect gut function and satiety but do not seem to affect insulin secretion: oxyntomodulin (OXM), glucagon-like peptide 2 (GLP-2), cholecystokinin (CCK), and Pancreatic Polypeptide (PP). OXM and GLP-2 are alternative splicing products of preproglucagon (PPG). They are secreted from the GLP-1 producing L cells, depending on carbohydrate or protein ingestion, and possess structural homologies to GLP-1 (Holst et al. 2018; Tang-Christensen et al. 2000; Baldassano, Amato, and Mulè 2016).

CCK is produced in so-called I cells, which are located in the duodenum and jejunum. This hormone is mainly released after meals containing fat and/or protein (Liou et al. 2011; Wang et al. 2011; Daly et al. 2013). Sex and eating behavior have an impact on circulating CCK levels (Burton-Freeman 2005). PP is a poorly understood hormone. It appears to improve hepatic insulin sensitivity (Śliwińska-Mossoń et al. 2017), and is secreted in response to rising levels of GIP (Veedfald et al. 2020).

Last, but not least, there is one unambiguous opponent of all satiety hormones: ghrelin. The ghrelin-producing X/A cells in the gastric wall are highly active during hunger but become dormant after meal initiation. Dietary triggers for ghrelin suppression are mainly proteins and carbohydrates (Tannous dit El Khoury et al. 2006; Overduin et al. 2005), but also fats (Heath et al. 2004). Ghrelin is only active as the acylated derivative of its precursor protein. Acylated ghrelin increases appetite, food-seeking behavior and craving, even exceeding energy balance. It also shows a glucoregulative effect, which is counteracted by the non-acylated ghrelin precursor (Heppner and Tong 2014; Gauna et al. 2005).

Relations between fiber fermentation, incretins, and glucose metabolism

Most soluble fibers, originating from cereals, fruits and vegetables, share a common pattern of viscosity and fermentability, inulin and fructanes being exceptions as non-viscous, but fermentable fibers. Both viscosity and fermentability contribute to the biological effects of soluble fibers. Viscosity is mainly acting by physics, as soluble fiber binds water, increases chyme and fecal mass and might therefore stimulate pressure-sensitive enteroendocrine cells in the GI tract. Viscous fiber was also proposed to produce a viscous “unstirred layer” which may delay the absorption of cabohydrates and thereby lower the GI (Lembcke et al. 1984).

Microbial fermentation in the lower parts of the gut leads to the production of short-chain fatty acids (SCFA), which are feeding both, other microbiota, and gut cells. After absorption into the blood-stream, acetate is circulating for a long time (Scheppach 1994), while most of the propionate is quickly metabolized in the liver to odd-chain fatty acids (Weitkunat et al. 2017). Butyrate is predominantly remaining in gut cells and serves as their metabolic fuel (Roediger 1980). Within the human body, these SCFA activate SCFA receptors: FFA2 (acetate and propionate), FFA3 (propionate and butyrate) and GPR109A (just butyrate). These receptors modulate immune responses (FFA2 and GPR109A) and incretin secretion (FFA2 and FFA3) (Kaji, Karaki, and Kuwahara 2014). In particular, L cells carry SCFA receptors, which highlights the role of SCFAs in the release of GLP-1 and PYY, the main signals derived from these cells. In-vitro stimulation with SCFA precursors stimulates L cell proliferation, while non-fermentable cereal fiber (cellulose) does not (Kaji et al. 2011; Mitsui et al. 2006). Upon administration of acetate or other SCFAs, FFA2 and to a smaller extent also FFA3 stimulate GLP-1 and PYY release (Freeland and Wolever 2010; Larraufie et al. 2018). After absorption, SCFAs from fermentable soluble fiber (including ß-glucans) stimulate insulin secretion and improve insulin sensitivity (Pingitore et al. 2017; Darzi et al. 2016; Hooda et al. 2010; Adam et al. 2014; Miyamoto et al. 2018). By this partially incretin-driven improvement of the insulin system SCFAs may also counteract the development of NAFLD (Weitkunat et al. 2016; Zhou et al. 2018). Alleviation of T2DM is also supported by SCFA-driven reinforcement of bacterial strains feeding on these fiber fermentation products (Zhao et al. 2018).

As SCFAs as fiber fermentation products contribute to the energy balance, excess intake of fermentable fiber may lead to a smaller metabolic benefit due to a different body weight trajectory. Mice being fed with high amounts of fermentable fiber gain weight, as the SCFAs provide excess energy and are metabolized to storable long-chain fatty acids (Isken et al. 2010).

Soluble cereal fiber – a distinct biochemical entity

As shown in cohort studies, whole-grain products and in particular insoluble total fiber (from all sources) contribute to reduced health risks. It is hard to determine, which fiber from which source is the active player. Whole-grain food, for example, contains both soluble and insoluble fiber, minerals, plant protein, vitamins, triterpenoids, and sterols, all of which might explain health benefits. While insoluble fibers in total (from cereals and other sources) are backed up by better evidence from cohort studies, they lack sufficient fiber-specific intervention studies. On the contrary, ß-glucans and other fermentable grain-based fibers have not been identified as strong predictors of lower health risks in epidemiological studies, but the active role of soluble fiber (in particular of cereal origin rather than fruits and vegetables) is well supported by mechanistic in-vitro and in-vivo studies. Additionally, many acute and short-term studies with clear and targeted intervention using ß-glucans confirm that at least some metabolic benefit is apparent. Soluble cereal fiber comprises two main chemical structures:

Cereal ß-glucans are alternating ß-1,3-/ß-1,4-linked glucose polymers. They are found predominantly in barley and oat, but only in minor fractions in rye, wheat and other grains. ß-glucans from yeasts are ß-1,3-/ß-1,6-linked, often water-insoluble glucose polymers, which possibly do not share the medical benefits of cereal ß-glucans (Kohl et al. 2009; Dalonso, Goldman, and Gern 2015). Other sources are algae, bacteria, lichens, yeasts, and specific edible fungi, which are not reviewed in this article.

Second, some members of the hemicellulose family are found to be water-soluble: pentosane-hemicelluloses from rye and oat, including some, but not all arabinoxylans. The third type – resistant starch type 2 (RS2) – comprises starch, which, despite being extracted from the cellular structure, resists digestion and acts as fermentable, soluble fiber. The main source is high-amylose maize.

Solubility depends on the proportion of arabinose monomers and their linkage to xylose (Kiszonas, Fuerst, and Morris 2013). Natural arabinoxylan occurs together with phenolic compounds (such as ferulic acid), binding to the fiber structure.

ß-glucans, soluble arabinoxylans and RS2 are both viscous and fermentable, sharing these properties with other kinds of soluble fiber: fructanes and inulin (non-viscous, but fermentable), polyuronides (pectin, alginates, agar, carrageen). These compounds are found in fruits, vegetables and legumes. As fruit and vegetable fiber has consistently shown not to be associated with lower risk for T2DM and the long-term risk profile (Aune et al. 2013; Ley et al. 2014; Weickert and Pfeiffer 2018), these types of fiber are not reviewed in this article.

Methodology of the review

The currently available literature on our topic was assessed via PubMed, using a search term, that covered whole grain (barley and oats, only), all types of cereal soluble fiber and all types of incretins or satiety-related gastrointestinal hormones: (((“fiber*” or “fiber*”) and soluble) or ß-glucan* or arabinoxylan* or pentosan* or (“whole grain" and (barley or oat)) OR “resistant starch”) and (glp-1 or “glucagon-like peptide” or gip or “glucose-dependent insulinotropic peptide” or glp-2 or ghrelin or “peptide yy” or pyy or cholecystokinin* OR cck or oxyntomodulin OR OXM OR incretin*). We selected all human intervention trials, using at least one specific soluble cereal high-fiber treatment in comparison to a low-fiber treatment or similar amount of insoluble fiber, assessing at least one of the hormones of interest, and being published until 30th June 2021.

We specifically included studies on clearly defined whole-grain foods, if they were derived from barley or oat. By this, we extend the view on possible effects of fiber-rich cereal-based diets. We are aware, that fiber-rich foods differ in their fiber composition (e.g. by containing both soluble and insoluble fiber) and in their overall food matrix, which might impair the generalizability of our review. For example, whole-grain foods and various kinds of grains can contain peptides inhibiting dipeptidyl peptidase IV, which may increase GLP-1 levels (Wang et al. 2015), while the sole fiber possibly would not alter GLP-1 release. Also, the magnitude of grounding affects dietary properties of whole-grain products, influencing the incretin profile and the glycemic response despite identical contents (Eelderink et al. 2017). However, by reviewing evidence for fiber and fiber-rich foods together, we can highlight the consistency of the results.

We excluded studies comparing complex fiber-rich diets to low-fiber diets, that were not based on a sole contrast between whole grain and refined grain, but rather used additional sources of fiber (e.g. fruits and vegetables) or other kinds of food modification (Zamaratskaia et al. 2017; Zhao et al. 2018; Hu et al. 2016; Silva et al. 2015; Bligh et al. 2015; Gonzalez-Anton et al. 2015; Dandona et al. 2015). We also excluded studies, which compared different kinds of grains or grain products, but not on the basis of fiber content (Bo et al. 2017; Stefoska-Needham et al. 2016; Bakhøj et al. 2003; Mano et al. 2018; Alyami et al. 2019; Stringer et al. 2013). Another study comparing glucose solution and breads with different types of soluble cereal fiber was also excluded (Binou et al. 2021).

The literature provides an enormous variety of treatments (fiber supplements, fiber-enriched foods and natural whole-grain foods), study designs (cross-over trials, parallel interventions, acute vs. preload treatment vs. long-term studies), cohorts (healthy subjects, patients with metabolic impairments), and assessments of outcomes (different ELISA or RIA kits, fasting and postprandial levels based on different oral stimulation tests). Therefore, a meta-analysis is not feasible at the time, even under consideration of subgroup analyses. Hence, we decide to provide a comprehensive, narrative review on the literature.

Included studies

Based on our selection criteria, we identified 38 eligible publications with 38 separate experiments or identical trials on separately reported study cohorts (e.g. a healthy normal weight and an obese sub-cohort). Out of these 38 experiments, 24 were reporting acute treatment effects only. Five papers reported a preload effect, in which the outcome was measured at a subsequent meal in the morning. Nine studies reported long-term outcomes after prolonged exposition to high- or low-fiber treatment. Most experiments on acute effects (n = 15) used an intervention with resistant starch or ß-glucans as food fortification, while nine studies investigated effects of specific types of soluble cereal fiber in other forms of application.

Most trials did not exceed 20 subjects per group and the majority of trials predominantly included female participants. Most trials were focused on young-to-middle-aged healthy and normal-weight subjects. Sex ratio is almost balanced for acute and preload trials, while women are overrepresented in the reviewed long-term studies on soluble cereal fiber. Details on cohort structure, reviewed interventions and assessed hormone parameters of all included trials can be found in Tables 1–4.

Table 1. Reviewed acute studies with use of a specific fiber.

Author	Year	Condition	Subjects	Treatment	Control	Assessed hormones
Raben et al. (1994)	1994	Healthy	10 (0 ♀)	Resistant starch	Resistant starch and sirup	GLP-1, GIP
Juvonen et al. (2009)	2009	Healthy	20 (16 ♀ )	Viscous ß-glucan beverage	Non-viscous ß-glucan beverage	GLP-1, PYY, ghrelin, CCK
Klosterbuer, Thomas, and Slavin (2012)	2012	Healthy	20 (10 ♀)	Resistant starch or soluble corn fiber alone or in combination with pullulan as meal additive	Meal without additive	GLP-1, ghrelin
Barone Lumaga et al. (2012)	2012	Healthy	14 (6 ♀)	ß-glucan-enriched beverage	Beverage with fruit fiber or without additive	PP, ghrelin
Bodinham et al. (2013)	2013	Healthy	30 (0 ♀)	Amylose-based resistant starch mousse at breakfast and lunch	Mousse without additive	GLP-1
Ye et al. (2015)	2015	Healthy	19 (10 ♀)	Fibersol-2-enriched sweet icetea	Icetea without fiber	GLP-1, GIP, PYY, ghrelin, CCK
Doyon et al. (2015)	2015	Healthy	20 (0 ♀)	ß-glucan w./wo. inulin + yoghurt	Yoghurt without additive	Ghrelin
Rahat-Rozenbloom et al. (2017)	2017	Healthy	12 (5 ♀)	Resistant starch added to an oral glucose tolerance test	Plain oral glucose tolerance tests	GLP-1, PYY, ghrelin
		Ov./obese	13 (8 ♀)
Al-Mana and Robertson (2018)	2018	Ov./obese	10 (0 ♀)	Resistant starch mousse	Mousse without additive	GLP-1

Table 2. Reviewed acute studies with use of fiber-enriched grain products.

Author	Year	Condition	Subjects	Treatment	Control	Assessed hormones
Bourdon et al. (1999)	1999	Healthy	11 (0 ♀)	ß-glucan-enriched barley flour pasta	Control wheat flour pasta	CCK
Juntunen et al. (2002)	2002	Healthy	20 (10 ♀)	ß-glucan-enriched rye bread	Whole-kernel rye bread, dark durum wheat pasta and white wheat bread	GLP-1, GIP
Beck, Tapsell, et al. (2009)	2009	healthy	14 (7 ♀)	ß-glucans in different dosage, breakfast cereals	Breakfast cereals without additive	Ghrelin, CCK
Beck, Tapsell, et al. (2009)	2009					PYY
Vitaglione et al. (2009)	2009	healthy	14 (7 ♀)	ß-glucan-enriched wheat bread	White wheat bread	PYY, ghrelin
MacNeil et al. (2013)	2013	T2DM	12 (7 ♀)	Resistant starch bagels	Bagels without additive	GLP-1, GIP
Hartvigsen et al. (2014)	2014	MetSy	15 (8 ♀)	ß-glucan rich bread or white wheat bread	Arabinoxylan-fortified bread	GLP-1, GIP, ghrelin
Ames et al. (2015)	2015	Healthy	12 (5 ♀)	ß-glucan-enriched barley tortillas	Control tortillas (amylose)	GLP-1, PYY
Gentile et al. (2015)	2015	Healthy	24 (24 ♀)	Resistant starch pancakes	Pancakes without additive	GLP-1, GIP, PYY, ghrelin
Ekström, Björck, and Östman (2016)	2016	Healthy	19 (10 ♀)	Resistant starch enriched grain bread	White wheat bread, guar gum rye bread	GIP, PYY, ghrelin
Emilien, Hsu, and Hollis (2017)	2017	Healthy	27 (12 ♀)	Resistant starch muffins	Muffins without additive	GLP-1, PYY, CCK
Emilien et al. (2018)	2018	Healthy	41 (19 ♀)	Resistant dextrin wheat drinks	Drink without additive	GIP, PYY, ghrelin, CCK
Zaremba et al. (2018)	2018	Healthy	33 (22 ♀)	Oat ß-glucan breakfast	Breakfast without additive	GLP-1
Belobrajdic et al. (2019)	2019	Healthy	20 (15 ♀)	Resistant starch breads	Breads without additive	GLP-1, GIP, ghrelin
Wolever et al. (2020)	2020	Healthy/overweight	28 (12 ♀)	Oat bran-enriched (different dosages) breakfast cereals	Conventional breakfast cereals	PYY, ghrelin

Table 3. Reviewed preload studies.

Author	Year	Condition	Subjects	Treatment	Control	Assessed hormones
Nilsson et al. (2008)	2008	Healthy	15 (5 ♀)	Ordinary (cut) barley-kernel bread	White wheat bread	fasting and postprandial GLP-1, GIP
Johansson et al. (2013)	2013	Healthy	19 (13 ♀)	Boiled barley kernels	White wheat bread	Fasting and postprandial GLP-1, GIP and ghrelin
Boll et al. (2016)	2016	Healthy	19 (10 ♀)	Arabinoxylan-fortified white wheat bread	White wheat bread	Fasting and postprandial GLP-1, Fasting, but not postprandial GLP-2
Kim et al. (2016)	2016	T2DM	13 (9 ♀)	Oat-fiber multi-grain cereal	Corn flakes	fasting and postprandial GLP-1, GIP
Sandberg, Björck, and Nilsson (2017)	2017	Healthy	21 (11 ♀)	Whole-grain rye or Whole-grain rye-/rye-kernel bread, with resistant starch or no additives	White wheat bread	Fasting and postprandial PYY

Table 4. Reviewed cross-over and parallel-designed long-term studies.

Author	Year	Condition	Subjects	Duration	Treatment	Control	Assessed hormones
Cross-over studies
Robertson et al. (2005)	2005	Healthy	10 (6 ♀)	2 × 4 weeks	Resistant starch in sachets	Digestable starch in sachets	Fasting and postprandial ghrelin
Stewart et al. (2010)	2010	Healthy	20 (10 ♀)	2 × 2 weeks	Resistant starch type 3, pullulan, soluble fiber dextrin, corn fiber as supplement	Control supplement (maltodextrin)	Postprandial ghrelin
Nilsson, Johansson-Boll, and Björck (2015)	2015	Healthy	20 (17 ♀)	2 × 3 days	Barley-based kernel bread	White wheat bread	Fasting and postprandial GLP-1, PYY, GLP-2, ghrelin and OXM
Nilsson et al. (2016)	2016	Healthy	21 (8 ♀)	2 × 4 days	Barley-based kernel bread	White wheat bread	Fasting and postprandial GLP-1, PYY and GLP-2
Zhang et al. (2019)	2019	healthy	19 (10 ♀)	2 × 4 weeks	Resistant starch type 2 in full diet	Diet without additive	GLP-1, PYY
Johnstone et al. (2020)	2020	Overweight/obese	19 (8 ♀)	2 × 10 days	Resistant starch 3 in low-fat weight maintenance diet	Low-fat weight maintenance diet without resistant starch	AUCs of GLP-1, GIP, PYY and ghrelin
Parallel-designed studies
Beck et al. (2010)	2010	Overweight/Obese	56 (56 ♀)	3 months	ß-glucan-enriched cereal products (2 dosages) + diet	Control products + diet	Fasting GLP-1, PYY, PYY3-36, CCK
Maziarz et al. (2017)	2017	Overweight	18 (15 ♀)	6 weeks	Resistant starch muffins	Control muffins without additive	Fasting and postprandial GLP-1, PYY
White et al. (2020)	2020	Overweight/obese with prediabetes	59 (39 ♀)	12 weeks	Resistant starch type 2 in yoghurts and meal packets	Amylopectin in yoghurts and meal packets	AUCs of GLP-1, PYY, total and active ghrelin

Effects of soluble cereal fiber or whole-grain food on glucose and insulin levels

The reviewed studies on acute effects of soluble cereal fiber or whole-grain often showed lower postprandial glucose and insulin excursions in their fiber-rich interventions. In healthy subjects, an average reduction of 20% in glucose AUC or up to 60% in glucose iAUC could be seen in interventions with purely supplemented ß-glucans (Juvonen et al. 2009), but especially in studies investigating this kind of fiber in a starchy matrix (enriched to grain-based food products) (Vitaglione et al. 2009; Ames et al. 2015; Ekström, Björck, and Östman 2016). Some of those studies on fiber-fortified food reporting an improved glycemia also showed a marked effect on postprandial insulinemia (Ames et al. 2015; Ekström, Björck, and Östman 2016), while one trial failed to show a parallel finding on insulin (Vitaglione et al. 2009). Other studies on ß-glucans or resistant starch as cereal food additive reported 15–20% lower postprandial insulin levels, but no difference in glucose excursions (Juntunen et al. 2002; Beck, Tosh, et al. 2009; Beck, Tapsell, et al. 2009; Emilien, Hsu, and Hollis 2017).

Resistant starch as additive to grain-based foods such as breads or pancakes lowered postprandial glycemia by 20–40% (Gentile et al. 2015; Belobrajdic et al. 2019). Oat-bran fortified breakfast cereal also showed a 20–30% improved glycemia and insulinemia (Wolever et al. 2020). Smaller effects (10–20%) on postprandial glucose and insulin were reported after interventions with resistant starch supplements in comparison to control (Raben et al. 1994; Klosterbuer, Thomas, and Slavin 2012), and in one trial investigating oat ß-glucan-fortified breakfasts (Zaremba et al. 2018).

Unchanged glucose excursions were seen in studies on the acute effects of ß-glucans in non-starchy foods. Herein, insulin levels weren’t altered as well (Barone Lumaga et al. 2012; Doyon et al. 2015). Such non-significant results were also seen in some studies on ß-glucans in starchy foods (Bourdon et al. 1999), on resistant starch supplements (Bodinham et al. 2013) or RS-fortified foods (Emilien et al. 2018). For the specific soluble fiber Fibersol-2 no measurements on glycemia or insulinemia were reported (Ye et al. 2015).

In overweight or obese subjects, test meals with resistant starch (vs. non-fortified control) elicited small (10%) or no significant effects on glycemia and insulinemia (Rahat-Rozenbloom et al. 2017; Al-Mana and Robertson, 2018). Study participants with metabolic syndrome also achieved a glycemic benefit by ß-glucan enriched bread (vs. arabinoxylan-enriched bread), characterized by 20% lower early and late postprandial AUCs for both glucose and insulin (Hartvigsen et al. 2014). In patients with T2DM, resistant starch lowered late, but not early postprandial glycemia and insulinemia (MacNeil et al. 2013).

In preload studies with healthy subjects, administration of barley kernels or barley kernel bread in the evening before a test meal lowered post-peak glucose decrements and insulin iAUCs, while leaving fasting and early (0–15 min) postprandial levels unaffected (Nilsson et al. 2008; Johansson et al. 2013). Contrary to that, an arabinoxylan-fortified white bread evening preload did not change fasting glucose before next day’s breakfast, but lowered its postprandial glucose AUC by about 15%. On the other hand, in this study the arabinoxylan preload reduced fasting insulin by 16-24% and postprandial insulin excursions by only 6–10% (Boll et al. 2016). Whole-grain rye bread had a similar effect, especially when fortified with resistant starch. These preload interventions lowered both postprandial glucose and insulin levels by 15–25% during a standard breakfast in the next morning, but fasting glucose and insulin were not altered (Sandberg, Björck, and Nilsson 2017). In the only preload study in T2DM patients (comparing oat-fiber and conventional cereal flakes), the fiber intervention induced 20% (peak) to 50% (iAUC) lower glycemia in the morning standard test meal (Kim et al. 2016).

The long-term studies in this review also investigated glycemia and insulin response following their fiber and control interventions. Two cross-over trials investigated barley-kernel bread vs. white wheat bread in healthy persons over few days. Despite the short intervention, both studies noticed 20–25% lower postprandial glucose levels and 10–15% lower postprandial insulin concentrations (Nilsson et al. 2015, 2016). A 4-weeks cross-over intervention trial using resistant starch supplements found a 33% improvement in insulin sensitivity, a 44% increased insulin-adjusted glucose clearance, despite overall lower insulin levels (Robertson et al. 2005). In another rather similar study, glucose levels did not differ between intervention and control condition, but early postprandial (30 min) levels of c-peptide and insulin were 15% and 25% higher after 4 weeks (Zhang et al. 2019). A 10-day cross-over intervention with resistant starch 3 paralleling a weight-maintenance diet after weight loss did also not affect fasting or postprandial glucose levels, but significantly lowered pre- and postprandial insulin by about 20% (Johnstone et al. 2020). Prolonged supplementation with pullulan, corn fiber or soluble fiber dextrin over 2 weeks had no effect on fasting glucose or insulin levels (Stewart et al. 2010).

A few long-term RCTs on soluble fiber and GI peptide hormone release were conducted in parallel design, all of them in overweight or obese persons. One study, investigating ß-glucan-fortified food products (vs. non-fortified controls) over a period of 3 months did neither effect glucose nor insulin levels (Beck et al. 2010). Another trial examined the effect of resistant starch over 6 weeks. Similarly, fasting and postprandial glucose and insulin levels were not affected by the intervention (Maziarz et al. 2017). Major cardiometabolic outcomes of the study by White et al. (White et al. 2020) had been published earlier (Peterson et al. 2018). Apart from HbA1c, which increased in the control group and remained stable in the intervention, no other glycometabolic outcome (fasting or postprandial) showed a significant different result between the study arms (Peterson et al. 2018).

Effects of soluble cereal fiber or whole-grain food on GI peptide hormones

Glucagon-like peptide-1

In healthy subjects, several trials on the acute effects of soluble fiber found a weaker postprandial GLP-1 stimulation after ingestion compared to control. Reductions of 20–60% in AUC were reported in studies on resistant starch (Raben et al. 1994; Klosterbuer, Thomas, and Slavin 2012; Bodinham et al. 2013), where the impact of resistant starch was augmented by adding pullulan (Klosterbuer, Thomas, and Slavin 2012). In one study, investigating viscous vs. non-viscous ß-glucans, postprandial GLP-1 was found to be 60% lower in the viscous fiber condition (Juvonen et al. 2009). An acute 40% increase in postprandial GLP-1 has been shown solely for high doses of Fibersol-2 dissolved in icetea and compared to placebo icetea in healthy subjects (Ye et al. 2015). Unchanged postprandial GLP-1 secretion after acute fiber interventions was found in studies in healthy subjects, investigating wheat-based resistant starch or dextrin (Rahat-Rozenbloom et al. 2017; Emilien, Hsu, and Hollis 2017), ß-glucans (Ames et al. 2015; Zaremba et al. 2018) and grain products enriched with ß-glucans or resistant starch (Juntunen et al. 2002; Gentile et al. 2015; Belobrajdic et al. 2019). Non-significant differences between fiber intervention and control were found in two studies in overweight or obese patients testing the effects of resistant starch (Rahat-Rozenbloom et al. 2017; Al-Mana and Robertson 2018). Another study showed indifferent GLP-1 stimulation using bagels with different amounts of resistant starch in T2DM patients (MacNeil et al. 2013).

Evening preloads with barley kernels induced 50% higher fasting and postprandial GLP-1 levels at a following morning test meal in healthy subjects when compared to white wheat bread (Johansson et al. 2013). A similar study found higher postprandial (AUC +25%), but not preprandial GLP-1 levels (Nilsson et al. 2008). Preloads with arabinoxylans and/or resistant starch (in healthy subjects) and fiber-enriched flakes (in T2DM patients) did not significantly change pre- or postprandial GLP-1 compared to control (Boll et al. 2016; Kim et al. 2016).

Similarly, prolonged intake of soluble cereal fiber led to controversial results. In healthy subjects, barley-based bread (compared to white wheat bread) increased pre- and postprandial GLP-1 release by 56% and 13%, respectively (Nilsson et al. 2015). In a similar cross-over experiment of the same research group, the average postprandial, but not the fasting GLP-1 levels were significantly (+14%) higher under barley conditions compared to control (Nilsson et al. 2016). A Chinese study examined the effect of resistant starch supplementation over 4 weeks and found a significant increase of early postprandial GLP-1 levels (+30% after 30 min), while fasting and overall postprandial concentrations did not differ between the interventions (Zhang et al. 2019). In overweight subjects resistant starch compared to control elicited no change in fasting or stimulated GLP-1 levels after 6 weeks (Beck et al. 2010). Fiber-fortified food products (resistant starch type 2 or 3) did not change GLP-1 secretion in overweight or obese subjects after 10 days or 12 weeks (Maziarz et al. 2017; White et al. 2020). Also, ß-glucan supplementation during weight loss diet did not augment effects on fasting GLP-1 after 3 months in obese persons (Johnstone et al. 2020).

Glucose-dependent insulinotropic peptide

The literature on fiber effects on GIP release appears to be slightly more consistent, but most of the studies reporting GIP levels after acute stimulation with soluble cereal fiber used a starchy food matrix to deliver the intervention. Therefore, only sparse information is available about the action of soluble cereal fibers in a non-starchy meal or oral test solution. Overall, an increase in GIP levels has not been reported in any of the reviewed acute, preload, or long-term trials.

In healthy subjects, ingestion of ß-glucan supplemented rye bread, resistant starch and resistant dextrin in different applications, each compared to other cereal foods or suitable control, led to significantly lower postprandial GIP release. Those reductions achieved 20–30% in AUC and up to 50% in iAUC (Raben et al. 1994; Juntunen et al. 2002; Gentile et al. 2015; Ekström, Björck, and Östman 2016; Emilien et al. 2018). Acute reduction of GIP stimulation of similar magnitude was also reported in a trial with T2DM patients investigating the effect of resistant starch (MacNeil et al. 2013). FiberSol-2 and resistant starch did not have an impact on postprandial GIP levels in two other studies with healthy subjects (Ye et al. 2015; Belobrajdic et al. 2019). Also, ß-glucan-enriched bread did not induce significantly different postprandial GIP levels compared to white wheat or arabinoxylan-fortified bread, when tested in patients with metabolic syndrome (Hartvigsen et al. 2014).

Evening preloads with barley kernels or different kinds of breads, including barley bread, did not affect fasting and postprandial GIP levels in healthy subjects (Nilsson et al. 2008; Johansson et al. 2013). Fiber-enriched cereal flakes did not elicit a different GIP profile compared to conventional low-fiber flakes, when administered to T2DM patients (Boll et al. 2016).

Only one study investigated long-term effects of soluble cereal fiber on GIP release. In this trial, a 10-day low-fat maintenance diet, administered to overweight-to-obese patients after weight reduction, was fortified with resistant starch 3 (vs. placebo as cross-over condition). However, both groups showed similar AUCs for GIP after the intervention period (Johnstone et al. 2020).

Peptide YY

Acute fiber effects on PYY secretion have been investigated in more than a dozen cross-over trials, providing quite consistent evidence for some fiber species.

An increased secretion by about 15% was detected in healthy subjects undergoing an interventions with ß-glucan enriched bread (Vitaglione et al. 2009). A complete shift in PYY excursions from U-shaped decrease to U-shaped increase was reported in a study investigating ß-glucan-enriched cereals (Beck, Tapsell, et al. 2009). A massive increase of PYY AUC by 250% was found when examining the metabolic effects of FiberSol-2 as placebo-controlled additive to icetea (Ye et al. 2015). About 30–40% higher levels of late postprandial (120–180 min) PYY were also induced by enrichment of pancakes with resistant starch (Gentile et al. 2015).

An 80% decrease in PYY AUC was seen after ingestion of a viscous ß-glucan drinks (compared to a non-viscous ß-glucan drink) in healthy persons (Juvonen et al. 2009).

No significant differences in comparison to control were reported from studies in healthy subjects using resistant starch as plain substance or food-fortification (Rahat-Rozenbloom et al. 2017; Ekström, Björck, and Östman 2016; Emilien et al. 2018). A single study on ß-glucan barley tortillas (vs. non-fortified amylose tortillas) also found no impact of soluble cereal fiber on PYY excursions (Ames et al. 2015). Additionally, no PYY-modulating effect for sources of soluble cereal fiber – namely oat bran and resistant starch – was seen in overweight subjects (Rahat-Rozenbloom et al. 2017; Wolever et al. 2020).

Among preload studies, only one RCT examined effects of soluble cereal fiber on PYY. In healthy subjects, an evening preload of rye whole-grain bread with resistant starch (but not without it) induced 20% higher fasting and postprandial PYY levels compared to white wheat bread (Sandberg, Björck, and Nilsson 2017).

Long-term effects of fiber-rich food on PYY were studied in several trials. 18% higher postprandial PYY levels were reported after 3 days of intake of barley kernel bread in healthy subjects, while fasting levels were not significantly higher (Nilsson et al. 2015). A very similar study using 4 days of supplementation with barley kernel bread in healthy participants showed no significant change in fasting or postprandial PYY levels (Nilsson et al. 2016). Long-term trials with 2–12 weeks of supplementation with resistant starch also failed to show an impact of this soluble fiber on PYY secretion (Zhang et al. 2019; Johnstone et al. 2020; Maziarz et al. 2017; White et al. 2020). In a 3-months trial investigating the effects of ß-glucans on the metabolic and endocrine profile in obese subjects undergoing a weight-loss diet, only the moderate dose of fiber (5–6 g/d) led to a marked 30% reduction of fasting PYY_1-36 levels, while the high dose (8–9 g/d) did not. The authors could not explain this U-shaped dose-response pattern of effect. Also, controversely, in this study, plasma concentrations of PYY_3-36 did not change at all (Beck et al. 2010).

Other non-incretin satiety hormones

Very few studies investigated the impact of soluble cereal fiber on other satiety-related gastrointestinal hormones.

A three-day intervention comparing barley kernel bread with white wheat bread in healthy subjects did not show an impact on fasting or postprandial OXM secretion (Nilsson et al. 2015). This study is the only one on OXM.

GLP-2 has not been examined in any trial on immediate or long-term effects of soluble cereal fiber. However, one trial in healthy subjects assessed if an overnight preload with arabinoxylan and/or resistant starch modulates pre- or postprandial GLP-2 secretion, but did not find significant changes in comparison to white wheat bread (Kim et al. 2016).

After three days of ingestion of barley kernel bread, a trendwise 9% increase in fasting GLP-2 levels and a 13% significant rise in postprandial GLP-2 was seen in healthy subjects (Nilsson et al. 2015). A similar study from the same research group could not replicate these results in a four-day setting with the same intervention in once again healthy participants (Nilsson et al. 2016).

ß-glucans in starchy foods increased the immediate postprandial CCK release in healthy subjects by about 20–40% (Bourdon et al. 1999; Beck, Tosh, et al. 2009). Another study on non-viscous ß-glucans (vs. viscous ß-glucans) in a non-starchy food matrix found a 40% reduction of CCK AUC in healthy subjects (Juvonen et al. 2009). Other studies on resistant starch in muffins or wheat drinks (Emilien, Hsu, and Hollis 2017; Emilien et al. 2018) and Fibersol-2 as additive to icetea had no such effect on postprandial CCK levels (Ye et al. 2015). The only long-term study reporting CCK levels found no impact of ß-glucans on these hormone levels in obese subjects under caloric restriction (Beck et al. 2010).

Postprandial PP appeared to be strongly stimulated (+35%) by ß-glucan ingestion in a study on acute effects in healthy subjects, which is the only trial on this hormone (Barone Lumaga et al. 2012).

Ghrelin

Most studies about fiber effects on ghrelin levels show non-significant results. Very few cross-over RCTs reported a stronger suppression of postprandial ghrelin by ß-glucans, administered to healthy subjects. Ghrelin levels turned out 10–30% lower compared to control conditions (Juvonen et al. 2009; Barone Lumaga et al. 2012; Vitaglione et al. 2009). In another trial on various isocaloric and isoproteinemic yoghurt snacks, the combination of added inulin and ß-glucans, but neither of the single additives augmented the postprandial ghrelin suppression (Doyon et al. 2015). Lack of changes in postprandial ghrelin levels were described in acute trials investigating resistant starch, pullulan and Fibersol-2 as additives to non-starchy oral stimuli (Klosterbuer, Thomas, and Slavin 2012; Ye et al. 2015; Rahat-Rozenbloom et al. 2017), but also in four studies examining the effect of RS-fortified grain products (Gentile et al. 2015; Ekström, Björck, and Östman 2016; Emilien et al. 2018; Belobrajdic et al. 2019). A single trial on ß-glucans measured postprandial acylated ghrelin levels, but found no differences between the interventions (Beck, Tosh, et al. 2009). No effect on postprandial ghrelin suppression was also seen in an RCT comparing ß-glucan rich vs. arabinoxylan-fortified bread in patients with Metabolic Syndrome and in overweight subjects ingesting oat bran vs. control (Hartvigsen et al. 2014; Wolever et al. 2020).

The only preload study investigating effects of soluble cereal fiber on ghrelin compared barley kernels vs. wheat bread in healthy subjects, but found no effect on fasting ghrelin levels throughout a morning test meal (Johansson et al. 2013).

Several studies assessed the long-term effects of soluble fiber on ghrelin secretion. In a 4-weeks cross-over study on resistant starch in healthy subjects, a 34% increase in fasting ghrelin was reported, while the postprandial suppression was not affected (Robertson et al. 2005). Pre- and postprandial ghrelin levels did not change significantly after a three-day stimulation with barley kernel bread (vs. white wheat bread) in healthy participants (Nilsson et al. 2015). Postprandial ghrelin suppression was also not altered by interventions with pullulan, resistant starch type 3, soluble fiber dextrin or soluble corn fiber in a 2-weeks cross-over trial in healthy subjects (Stewart et al. 2010). Resistant starch type 2 and 3 also turned out to be ineffective on ghrelin levels during 10 days or 12 weeks of supplementation with fiber-fortified food products in overweight and obese patients (Johnstone et al. 2020; White et al. 2020).

Comprehension and outlook

Natural food items are affecting the incretin system in a very elaborate and complicated way (Ríos et al. 2019). The reviewed selection of well-controlled, randomized, but usually rather small studies in mostly healthy subjects shows somewhat conclusive results for some of the investigated parameters, but lacks the statistical power and the generalizability to clearly describe the role of soluble cereal fiber in the modulation of incretins and other GI peptide hormones. The existing literature includes studies using different types and dosages of fiber, most of the RCTs investigated acute, but not prolonged effects. The overall evidence has to be evaluated as low-to-moderate.

Our current work now provides insight for soluble fiber, which has been shown to acutely affect glucose and lipid metabolism. Up-to-date, there is only insufficient literature in order to conduct a systematic review with meta-analysis on incretin effects due to administration of soluble fiber. There are clearly many RCTs on various types of soluble fiber (ß-glucans, inulin, resistant starch), but most of them lack an analysis of gastrointestinal hormones. So far, only two meta-analyses have investigated the effects of prebiotics on ghrelin (da Silva Borges et al. 2020) and resistant starch on several GI peptides (Guo, Tan, and Kong 2020). Effects of all soluble cereal fibers on glucose metabolism and the involved GI peptides in humans can be summarized as follows:

1. Epidemiological evidence points toward a strong antidiabetic association for whole grain and cereal insoluble fiber, but not soluble fiber of any kind. Whether whole grain effects are supported by soluble fiber is doubtful, as some sources of soluble cereal fiber are uncommon in the daily diet of most people. Oats are only consumed by a fraction of the population and barley is mainly used for livestock feeding and beer production, limiting the intake of ß-glucans by natural sources.

RCTs show consistent metabolic benefits for soluble fibers in acute trials, especially for ß-glucans, but there is insufficient data for long-term effects. Once again, studies on “whole-grain” fail to pinpoint their crucial component. Also, it is unclear, if certain effects of cereal fiber depend on a specific food matrix, even though ß-glucans are effectively reducing cholesterol levels as a supplement.

• There are some nutrigenetic studies showing an interaction between (a) genes, which are involved in the regulation of incretin release, (b) whole-grain intake, and (c) diabetes risk or diabetes-associated traits.

• More than 25 cross-over RCTs have assessed the immediate effects of soluble cereal fiber and fiber-enriched cereal products on glucose metabolism and relevant GI peptide hormones.

• A glycometabolic benefit of ß-glucans and resistant starch can be consistently demonstrated, when these types of soluble cereal fiber are tested embedded in a starchy food matrix. Improvements in insulin levels were reported in several trials, but not consistently linked to lower glucose levels. Acute studies on non-starchy fiber interventions and trials in patients with metabolic impairment (obesity and T2DM) showed less impressive effects on glycemia and insulinemia. Both in acute and preload studies, fasting levels of glucose and insulin remained unaltered, while especially late postprandial metabolism was improved. In long-term studies, reduced insulinemia was more common than improved glycemia and generally improvements of metabolism were not consistently found over the spectrum of interventions.

• Results of acute trials are completely inconclusive for GLP-1, as we cannot identify a consistently directed effect of any kind of soluble cereal fiber of fiber-rich food on this endocrine player. This is rather surprising, as this group of fiber is known to be fermented to SCFAs which increase GLP-1 levels. Increased GLP-1 levels could correspond to improved glycemia and insulin resistance, but higher GLP-1 was mostly reported in preload and long-term studies, but not after acute stimulation. Patients with overweight, obesity and/or T2DM did not show a change in GLP-1 levels to either acute, preload or prolonged fiber stimulation.

  Postprandial GIP seems to be either suppressed or unaffected by ß-glucans and resistant starch in acute feeding trials. Lack of effect was predominantly seen in subjects with metabolic syndrome.

  Consistently, ingestion of fiber-enriched foods (ß-glucans, resistant starch and other fiber species) in healthy subjects elicited either an increase or no changes in PYY levels. An increase would be expected, as SCFAs as fermentation product of soluble fibers increase PYY secretion. However, studies in patients with overweight mostly failed to show an effect on PYY.

  In particular ß-glucans seem to increase CCK and to suppress postprandial ghrelin. A reduction of GIP could contribute to the acute antihyperglycemic effects seen in soluble fiber treatment and might also indicate a therapeutic potential in NAFLD. Augmented PYY and CCK and suppression of ghrelin might explain weight loss effects, especially for ß-glucans and resistant starch (Rahmani et al. 2019; Snelson et al. 2019). Surprisingly, studies investigating fiber effects on ghrelin in patients with obesity or T2DM failed to show such a nutrient-driven modulation. For CCK, there is also some evidence for nutrigenetic mechanisms involving fiber intake (Jonsson et al. 2013).

  Overall, studies with insignificant effects on hormone secretion used comparable dosages of dietary fiber compared to RCTs with prominent interventional changes.

• Only few trials investigated the effects of fiber-rich preloads on GLP-1, GIP, PYY, and ghrelin. They do not present any consistent effect for soluble fiber on any of the aforementioned hormones.

• Long-term studies are sparse and most of the data does not show an impact of any kind of soluble cereal fiber on GI hormones.

• Data on PP, CCK, OXM, and GLP-2 are available in only very few trials.

Type-2 resistant starch and ß-glucans are the most promising representatives of soluble cereal fiber and should be investigated in large, high-quality acute and long-term trials, addressing both the endocrine and the resulting metabolic outcome. As these fibers are prone to fermentation, identification of the respective individual SCFA proportions is warranted for each fiber species. Currently, most trials on GI hormones have been conducted in healthy subjects, while patients at metabolic risk (prediabetes, early T2DM) are rarely investigated. Further studies with a broader view on the metabolic potential and mechanisms of soluble cereal fiber (including inflammation and liver fat) are desired.

Authors’ contributions

SK, MOW and AFHP wrote the paper and serve as guarantors of this work.

Abbreviations
CCK	=	cholecystokinin
GIP	=	glucose-dependent insulinotropic peptide
GLP-1	=	glucagon-like peptide 1
GLP-2	=	glucagon-like peptide 2
IFG	=	impaired fasting glucose
IGT	=	impaired glucose tolerance
NAFLD	=	nonalcoholic fatty liver disease
OXM	=	oxyntomodulin
PP	=	pancreatic polypeptide
PPG	=	preproglucagon
PYY	=	peptide YY
RCT	=	randomized-controlled trial
RS	=	resistant starch
T2DM	=	type 2 diabetes mellitus

Disclosure statement

SK received a travel grant from Rettenmaier & Soehne, Holzmuehle, Germany, including conference fees and accomodation. All authors have conducted studies with non-financial support from Rettenmaier & Soehne, Holzmuehle, Germany. The authors declare no further conflicts of interest associated with this manuscript.

Funding

The author(s) reported there is no funding associated with the work featured in this article.