Evaluation of yellow pea fibre supplementation on weight loss and the gut microbiota: a randomized controlled trial

A great deal of focus has been directed towards elucidating and evaluating the beneficial metabolic effects of fibre intake on gut as well as whole-body health in humans [9, 10]. The primary proposed mechanisms for how dietary fibre (specifically soluble and fermentable fibre) contributes to improved weight control and insulin sensitivity include delayed nutrient absorption, stimulation of gut hormones that regulate food intake and modulation of gut bacteria.

A recent meta-analysis of randomized controlled trials showed that increased fibre intake, whether from diets containing foods rich in fibre or soluble fibre supplements, was associated with reduced HbA1c and fasting plasma glucose in patients with type 2 diabetes [11]. The improvement in glycemic responses following a high-fibre meal has also been shown to carry through after a subsequent meal, the so-called “second meal effect” [12]. Part of the beneficial effect of dietary fibre on postprandial glucose control may be related to the release of gut peptides that influence insulin responses and food intake, such as glucagon like peptide 1 (GLP-1), peptide YY (PYY) and ghrelin. For example, feeding the prebiotic fibre oligofructose for 12 weeks reduced levels of the orexigenic hormone ghrelin and increased levels of the anorexigenic hormone PYY in overweight adults which corresponded with lower self-reported energy intake [13]. In the long-term, these changes in postprandial peptide secretion may promote weight loss through reductions in ad libitum food intake.

Members of the Firmicutes and Bacteroidetes phyla represent the majority of species in the human gut, while Proteobacteria, Actinobacteria, Chlamydiae, Spirochetes, and Fusobacteria are present in markedly lower abundance [14]. The gut microbiota phenotype can be both genetically and environmentally determined. There is some degree of genetic influence, because relatives show greater similarities in phylotypes than unrelated individuals even when these family members live in different environments [15]. However, investigations have largely been targeted towards determining how microbiota can be environmentally influenced, specifically by disease state and lifestyle interventions. For example, obese individuals have been observed to have greater proportions of Firmicutes and lesser Bacteroidetes [15, 16]. A lower proportion of Bacteroidetes has also been observed as glucose tolerance worsens [17] and during progression of non-alcoholic steatohepatitis [18].

Both bacterial abundance and composition can be altered with lifestyle changes. Indeed, part of the role of intestinal bacteria in weight control has been elucidated through study of patients receiving bariatric surgery. Kong et al. [19] profiled the gut microbiota of 30 morbidly obese women before and after receiving Roux-en-Y gastric bypass surgery [19]. After the surgery, Bacteroidetes was increased and Firmicutes and bifidobacteria populations were reduced. Further, the “richness” (referring to the number of species present) was greater after surgery, and the gut microbiota composition correlated more strongly with white-adipose tissue gene expression compared to pre-surgery [19]. Weight loss produced via dietary restriction results in an increase in the relative proportion of Bacteroidetes and a reduction in Firmicutes, such that the microbiota profile becomes more similar to lean healthy control subjects [16]. Feeding specific dietary fibres can also beneficially modulate gut bacteria populations. For example, in a double-blind placebo-controlled trial in obese women, feeding 16 g/d inulin and oligofructose (50/50) vs. placebo (maltodextrin) for 3 months decreased Bacteroides, which was associated with a slight reduction in fat mass [20]. In this study, ¹H-NMR was also performed, but there was no clear clustering effect of treatment that could be detected for gut microbial analysis or plasma/urine metabolomic profiles. Indeed, the specific metabolic function of different bacterial species is still being elucidated. However, metagenomic mapping of gut bacteria indicates a highly enriched and functionally diverse genome related to numerous pathways including metabolism of carbohydrates, energy, amino acids, and vitamins [21].

Part of the beneficial metabolic impact of gut bacteria has been attributed to the byproducts produced during fermentation of nondigestible carbohydrates (i.e. dietary fibre) by these organisms. Short-chain fatty acids (SCFA) are the primary end-products of carbohydrate fermentation by gut bacteria, and include acetate, propionate, and butyrate [22, 23]. SCFA are a major energy source for colonocytes [22]. Recently, it has been discovered that SCFA are natural ligands for the G-protein coupled receptor GPR43, which is expressed in numerous tissues including immune cells, adipocytes, and endocrine cells [23, 24]. While the metabolic implications of SCFA are still being elucidated, emerging data suggests that the binding of SCFA to GPR43 may improve glucose tolerance and fat metabolism. For example, in primary culture the binding of SCFA to intestinal GPR43 stimulates GLP-1 secretion by colonic L cells [25]. In animals, mice overexpressing GPR43 in adipose tissue remain lean on a high-fat diet and exhibit lower body weight, white adipose weight, and smaller adipocyte size [26]. In contrast to wild-type mice, these animals did not develop hepatic steatosis in response to a high-fat diet and had normal glucose tolerance. Finally, the GPR43 mice also displayed greater energy expenditure and increased fat oxidation [26]. While these findings are preliminary, the implication is that SCFA may have important effects on the host organism beyond energy production in the gut.

Recently, the interest in pulse crops (e.g. dried beans and peas) as sources of dietary fibre has increased [27]. Yellow peas are a native Albertan pulse crop [28] with the potential to have beneficial effects on plasma lipids, glucose metabolism, and weight control. Pea hulls are comprised of ~82% fibre making them an excellent source of dietary fibre that could be incorporated into food products [29].

To date, few studies have examined the metabolic effects of pea fibre in animal models. Briefly, in rats with diet-induced glucose intolerance, feeding dry pea seed coats improved glucose tolerance as well as fasting and glucose-stimulated insulin levels [30]. Investigation of skeletal muscle expression of proteins involved in insulin signaling suggested that pea seed coat consumption increased glucose transport. Finally, pea seed coat consumption also appeared to reduce markers of oxidative stress [30]. In hypercholesterolemic Golden Syrian hamsters, feeding whole pea flour or fractionated pea flour (hulls only) did not reduce plasma lipid levels, but decreased fasting glucose and insulin concentrations, and showed a statistical trend towards reducing body fat [31]. More recently, we have demonstrated that yellow pea fibre reduces glycemia in diet-induced obese rats while yellow pea flour (rich in pea protein) reduces adiposity (Eslinger et al., Submitted). Both pea fibre and pea flour reduced the percent of fecal Firmicutes.

The state of the literature regarding metabolic effects of pea hull fibre specifically in humans is limited to few studies. In an acute meal setting, adding pea fibre (10 g) to a low-fibre (2.8 g) test meal resulted in lower postprandial plasma cholesterol levels compared to soybean fibre in normolipidemic men, but postprandial glucose, insulin, and plasma triglyceride were not affected by fibre type [32]. Similarly, isoenergetic high-fibre (26 g) and low-fibre (9 g) meals produced similar postprandial glucose and insulin in healthy normal men, however plasma free fatty acids were better suppressed after the high-fibre meal which contained pea fibre baked into wheat bread [33]. Conversely, in one study of healthy adults, pea fibre (15 g) incorporated into a mixed test meal significantly reduced the postprandial glucose incremental area under the curve as compared to sugar beet fibre and wheat bran, while addition of all fibres reduced the postprandial insulin response [34].

As opposed to the effects of acute ingestion of pea fibre, the metabolic effects of consuming diets containing foods fortified with pea fibre are also limited to a few studies. In one study, addition of pea hull fibre to existing foods (resulting in an average of 3 g/d pea fibre) increased stool frequency in elderly residents of a long-term care facility, representing a safe, economical, and non-invasive alternative to relieve constipation in the elderly [35]. Feeding young healthy adults a low-fibre diet or a low-fibre diet supplemented with 33 g/d pea fibre product (corresponding to 20 g dietary fibre) in a randomized cross-over study showed no differences in fasting cholesterol levels however triglyceride levels in the plasma and those arising from the liver were reduced in the pea fibre supplemented group [36]. In addition, incorporation of the fibre into breakfast and lunch meals resulted in significantly lower plasma and chylomicron triglyceride responses, as well as plasma insulin concentrations. Incorporation of whole yellow pea flour into baked products reduced postprandial glycemia in healthy men and women compared to products containing whole wheat flour [37]. Finally, in a population who could potentially benefit the most from dietary fibre provision, providing baked products containing pea fibre (12 g/d of fibre) to overweight hypercholesterolemic adults for 28d reduced fasting insulin concentrations and improved postprandial glucose responses after a standardized breakfast meal [38]. Given the limited state of evidence for the potential benefits of pea fibre in individuals at-risk for obesity and metabolic syndrome, clinical trials specifically examining the clinical outcomes and mechanisms related to pea fibre consumption are warranted.

This study will utilize a double-blind, placebo-controlled, parallel group design in which 60 overweight and obese adults (age 18-70 y; BMI 25-38 kg/m²) will be randomized to either a placebo (n = 30) or pea fibre supplemented (n = 30; 15 g/d yellow pea fibre) group for 12 weeks. Randomization will be carried out using computer generated numbers and stratified according to age, sex, and BMI.

Adults with a stable body weight for at least 3 months prior to the study, defined as weight loss or gain of <3 kg within preceding 3 months to enrollment, will be eligible to participate. Exclusion criteria include: concomitant use of any weight loss medication (e.g. Orlistat) or diet/exercise regime designed for weight loss; use of corticosteroids, anti-depressants, or anti-epileptic, lipid-lowering, and diabetes medications; previous bariatric or other intestinal surgery; pregnancy or lactation; use of bulk laxatives or probiotic/prebiotic supplements; chronic use of antacids; use of antibiotics within the preceding 3 months of enrollment; clinically significant cardiovascular or respiratory disease; liver disease; alcohol or drug dependence; body weight >350 lb; active malignancy; and chronic infections. Eligibility will be assessed by the researchers using a screening questionnaire and phone interview.

The overall study design is described in Figure 1, and the outcome measures are described individually below. Anthropometric data (weight, BMI, waist circumference) and food intake (3-day weighed food record) will be collected at baseline and at 4 week intervals. Subjective ratings of appetite will be measured using validated visual analogue scales (VAS) by the subjects at home on a weekly basis. At baseline and follow-up (12 weeks), body composition will be measured by DXA, and an oral glucose tolerance test will be performed to assess glucose tolerance. Plasma inflammatory markers (including TNF-α and CRP), lipids (total, LDL- and HDL-cholesterol, and triglyceride), and satiety hormones (including GLP-1, PYY, ghrelin and leptin) will be measured using methods already established in the laboratory and described below. Subjects will be provided with an ad libitum lunch buffet to measure objective food intake. Serum samples will undergo metabolomics analysis using proton nuclear magnetic resonance spectroscopy (¹H-NMR) [39]. ¹H-NMR will allow us to measure global metabolic responses which will complement the plasma biomarkers concurrently measured. Stool samples will be collected for analysis of gut microbiota and for fecal water metabolomics using ¹H-NMR.

Sample size was calculated based on anticipated changes in body fat using data from our previous study of prebiotic fibre supplementation in overweight and obese adults [13]. The expected change in body fat in the pulse fibre group is -1.11 kg whereas the expected weight change in the control group is +0.20 kg, with a standard deviation of ±1.6 kg. Assuming an alpha = 0.05 and power = 0.80 (80%) we would need to recruit 24 subjects in each study arm. An additional 6 patients per group will be added to account for the expected 20% loss to follow up for a total of 30 patients per group.

For metabolomics analysis, in addition to univariate tests, multivariate analysis will conducted using SIMCA-P + 12.0.1 software (Umetrics, Sweden). Data will be reprocessed by mean-centering and unit variance scaling. A supervised partial least squares discriminant analysis (PLS-DA) approach will be used to compare the variance of metabolite concentrations between the two sample classes (diet: control and pea fibre). A supervised orthogonal partial least squares analysis will be used (control versus pea fibre) for a direct comparison of the variance between diet type (Y variable) and metabolite concentrations (X variable). The results from the metabolomics analysis will also be combined with the plasma satiety hormones and other biological measurements and regressed to the diet type using O2-PLS-DA (orthogonal PLS-DA).

This proposal has been approved by the Conjoint Health Research Ethics Board of the University of Calgary. Voluntary, written informed consent will be obtained from each participant.