Low glycemic index common bean snack increased satiety without modifying energy intake in adults with normal weight: randomized crossover trials

Abstract This study aimed to evaluate the short-term effect of a common bean baked snack (CBBS) and cooked bean consumption on energy intake, satiety, glycemic response, and palatability in subjects with normal weight (Study 1) and overweight (Study 2) and to determine the glycemic index of CBBS (Study 3). For studies 1 and 2, satiety and glycemic response were measured over 45 min after consuming CBBS, cooked beans or white bread preload, and energy intake at an ad libitum test meal was calculated. Energy intake remained similar after consuming the three preloads in both studies. Compared to white bread, CBBS consumption increased fullness by 52% in subjects with normal weight but not in those with overweight. The CBBS calculated glycemic index was considered low (42). Consumption of low glycemic index CBBS increased satiety in adults with a normal weight. Long-term trials assessing the effects on body weight management are necessary.


Introduction
Overweight and obesity increase the risk of cardiovascular diseases, diabetes mellitus, dyslipidemia and certain cancers (Lega and Lipscombe 2020). These disorders are associated with an unbalanced energy intake caused by unhealthy dietary habits, including a constant consumption of energy-dense foods, saturated and trans fats, refined sugars, and deficient protein and dietary fibre intake (Lane et al. 2021). Food consumption responds to numerous sensory, endocrine and nervous signals that control appetite and satiety (Mont egut et al. 2021). Regular high protein and dietary fibre intake is a good strategy to enhance satiety (Munekata et al. 2021). Among foods, pulses have been demonstrated to improve long-term body weight management by increasing satiety and decreasing food intake (Clark and Duncan 2017).
Common bean (Phaseolus vulgaris L.) is one of the most significant pulses cultivated and consumed in the Americas and Africa (Rawal and Navarro 2019). The inclusion of common beans in a daily diet may also reduce the risk of certain non-communicable diseases due to their bioactive components such as proteins and peptides, dietary fibre, polyphenols, saponins and phytosterols (Ferreira et al. 2021;Mullins and Arjmandi 2021). Despite these benefits, pulse consumption does not meet recommendations in several regions. For instance, only 17% of the United States Population meets the recommended cooked pulses intake of 1.5 cups per week (Dietary Guidelines Advisory Committee and Data Analysis Team 2020; USDA and HHS 2020). Consequently, research on pulses has extensively increased to provide reformulated food alternatives with pulses to promote consumption (Monnet et al. 2019;Tas and Shah 2021).
As consumers seek healthier food options to satisfy their appetite between meals, common bean-based snacks could meet this growing demand. Most commercially available snacks are not nutritionally balanced because they are produced from cereals and added with excessive fat, sugar and salt (Cuj-Laines et al. 2018). Furthermore, savoury snack foods, including extruded snacks, chips and popcorn, have an average glycemic index of 60, whereas boiled beans only have 42 (Atkinson et al. 2021;Singh et al. 2021). Glycemic index values are relevant due to their influence on the prevention and control of diabetes mellitus and body weight management (Livesey et al. 2019;Zhu et al. 2021).
Regardless of nutritional properties, common beanbased snacks must overcome obstacles such as reducing antinutritional factors, achieving consumer sensory acceptance and scientifically demonstrating their health benefits. Short-term clinical trials remain to be performed in individuals with overweight to evaluate whether common bean consumption has similar effects on satiety and food intake as people with normal weight. This information is important because individuals with excess body weight have a dysregulation of orexigenic and anorexigenic hormones in fasting and postprandial states (Li et al. 2014;Farhadipour and Depoortere 2021;Smith and Moran 2021).
In a previous study, a common bean-baked snack (CBBS) with a high protein and dietary fibre content and low content of a-galactooligosaccharides was developed and demonstrated to be sensory accepted and improved cardiovascular health . However, the effects related to satiety, food intake and glycemic response remain unknown. Therefore, the purpose of this study was to evaluate the acute effect of CBBS and cooked bean consumption on energy intake in individuals with overweight and normal weight. In addition, subjective satiety, glycemic response, and palatability were assessed. Finally, the glycemic index of CBBS was also determined.

Materials
Raw Peruano bean cultivar (Phaseolus vulgaris L.) and white bread (Grupo Bimbo, Mexico City, Mexico) were purchased at a local market and stored at room temperature. Common beans were cleaned to remove debris and damaged seeds. The common bean baked snack (CBBS) (Figure 1) was made as described by  and was individually packaged in resealable bags and stored at room temperature. Meanwhile, cooked beans were prepared the day before each test day. Beans were soaked in purified water (1:3 w/w) for 24 h at 20-25 C. Then, the soaking water was removed, and the beans were cooked in boiling water (1:5 w/w) for 1.5 h. Lastly, cooked beans were individually packaged in resealable bags at 4 C.
Studies 1 and 2: energy intake, satiety, glycemic response, and palatability Trial design Two similar acute randomised 3 Â 3 crossover studies were conducted with normal-weight and overweight subjects. Participants were assigned to one of the three study sequences of a 3 Â 3 Latin square design with computer-generated randomisation. They underwent three separate study meals in three different sessions ( Figure 2). The Ethics Committee of the University of Guadalajara (CI-09620, 18 December 2020) approved the study protocol, and all participants provided written informed consent. This trial was registered at ClinicalTrials.gov as NCT05230979.

Participants
For Study 1, eligible participants were men and women with a BMI between 18.5 and 24.9 kg/m 2 . Meanwhile, men and women with a BMI from 25.0 to 29.9 kg/m 2 were recruited for Study 2. Both studies included adults aged 18-50 years. Subjects were not included if they had one of the following conditions: pregnancy, lactation, established plans to lose or gain weight in the next two weeks, surgical event or modification of diet or physical activity in the last three months, diagnosis of diabetes, cancer, cardiovascular disease, gastrointestinal disorder, pancreatitis, kidney, liver or thyroid disease, eating disorder, smoking or drug use, sensitivity for common bean consumption and pharmacological treatment or consumption of non-prescription drugs, herbal or nutritional supplements known to modify appetite or satiety.
Participants were recruited by open invitation in Guadalajara, Mexico. Interested individuals were screened using a self-administered online form and body weight and height measurements to establish eligibility. Before the beginning of the study, blood samples were collected after ten-hour fasting to determine the baseline biochemical characteristics of participants. The study was carried out at two centres, the University Centre for Health Science of the University of Guadalajara and the Centre for Research and Assistance in Technology and Design of the State of Jalisco, from November 2021 to April 2022.

Interventions
Participants were asked to consume a similar dinner meal the night before each study day, restrict the intake of alcohol and other caffein beverages, and avoid intense physical activity for 24 h prior to a testing day. Participants attended the institute after 10-14 h fasting. Initially, they completed a 24-h diet record, a physical activity questionnaire, and the first visual analog scale (VAS) of subjective satiety questionnaire, followed by a finger-prick blood sample (Flint et al. 2000). Immediately, they were instructed to consume 40 g of the preload (CBBS, cooked beans or white bread) and 100 mL of water within 10 min ( Table 1). The serving size was established based on the local regulation for food and non-alcoholic beverages with modifications in their composition (Secretar ıa de Salud 1996). Common bean baked snacks and white bread preloads had the same amount of available carbohydrates (19.5%) except for cooked beans (6.4%). Cooked beans were microwaved for 30 s before serving, and white bread was served as slices at room temperature. The VAS and blood glucose were measured every 15 min after starting to eat the test food. Another VAS was used to assess the palatability as a hedonic evaluation of the preload products. An ad libitum meal consisting of freshly prepared sandwiches and 250 mL of water was served after 45 min because, at this time, there is a maximum sensitivity to preload energy differences (Forde 2018). Sandwiches were prepared with white bread, turkey breast ham, lettuce, and mayonnaise or sour cream (Table S1). The meal was weighed before serving, and leftovers were weighed again to determine the amount of food consumed. Finally, the VAS of subjective  satiety was assessed again. We decided not to use a placebo because no placebo for this common bean baked snack or cooked beans could achieve successful blinding of participants.

Primary and secondary outcomes
The primary outcome was a change in energy intake after the CBBS and cooked beans consumption periods. Meanwhile, the secondary outcomes were blood glucose changes and the VAS of satiety, hunger, prospective consumption, and palatability. Palatability consists of the hedonic evaluation of the integrated sensory food cues such as smell, taste (sweet, salty, bitter, sour, and umami), texture, and overall acceptance, as it influences how much a person eats (McCrickerd 2016).

Outcome measurements
Participants completed questionnaires to rate satiety, hunger, and prospective consumption using a 100 mm visual analog scale (VAS) and AUC were calculated. The amount of food eaten was converted into calories to estimate the energy intake using the data from the nutrition facts label of each ingredient. Capillary blood glucose was determined using an Accu-Chek Guide glucose metre (F. Hoffmann-La Roche Ltd, Basel, Switzerland). Besides, VAS was also used to assess the palatability as a hedonic evaluation of the preload products, including taste, smell, texture and overall acceptability. Regarding the baseline biochemical characteristics of participants, blood samples were centrifugated, and serum was stored at À80 C until analysis. The total cholesterol, triglycerides, high-density lipoproteincholesterol and glucose levels were measured by multilayered colorimetric slides on a Vitros 350 Analyser (Ortho Clinical Diagnostics, Inc., Raritan, NJ, USA). The low-density lipoprotein-cholesterol was calculated using the Friedewald formula (Friedewald et al. 1972). Insulin was determined with an ARCHITECT i2000SR immunoassay analyser (Abbott, Chicago, IL, USA), and the homeostatic model assessment of insulin resistance (HOMA-IR) was estimated according to the Matthews formula (Matthews et al. 1985).
Bodyweight, body fat percentage and muscle mass were obtained using a bioelectrical impedance scale (BC-585F, Tanita Corporation, Tokyo, Japan), and height was taken using a Seca 213 stadiometer (Medical Measuring Systems and Scales Seca, Hamburg, Germany). Participants completed the Short Form of the International Physical Activity referring to the day before each session.

Study 3: determination of glycemic index (GI)
The study was an acute trial under the ISO 26642:2010 method for determining the GI (International Organization for Standardization 2010). Ten participants received a reference glucose solution containing 25 g of anhydrous glucose per 250 mL of water in two separate sessions and 54 g of CBBS, the equivalent of 25 g of available carbohydrates. The randomisation sequences for participants were computergenerated. Participants attended the institute after 10-14 h fasting. At first, two baseline finger-prick blood samples were collected with a 5 min interval between the samples using an Accu-Chek Guide glucose metre (F. Hoffmann-La Roche Ltd, Basel, Switzerland). Then, they consumed the reference glucose solution or the test food with 250 mL of water within 12-15 min. Later, finger-prick blood samples were taken at 15, 30, 45, 60, 90 and 120 min after starting to eat the test food. Incremental areas under the glucose response curves (iAUC) were calculated using the trapezoidal rule, ignoring the areas below the fasting value.

Statistical analysis
A total of 18 participants were estimated to detect a 15% change in AUC of satiety ratings, assuming a within-subject SD of 25% at a 5% significance level with 80% power. The AUC and iAUC were calculated using Prism GraphPad version 9.3.1 (GraphPad Software, San Diego, CA, USA). Statistical analyses were performed using NCSS Statistical Analysis and Graphics v22.0.4 (NCSS, LLC., Kaysville, UT, USA). All the variables were assessed for normality, and the values were presented as means ± SEM or medians [IQRs]. Data were compared using the repeated-measures ANOVA and Tukey's post hoc test.

Participants
Eighteen individuals with normal weight and eighteen with overweight were recruited and completed the sessions (Figures 3 and 4). No subject dropped out of the studies. The baseline characteristics are shown in Table 2.

Energy intake
In both studies, energy intake was similar between the three preloads (p > .143), ranging from 318 to 365 kcal for individuals with normal weight and from 357 to 375 for participants with overweight ( Figures 5 and 6).
Subjective satiety, glycemic response, and palatability For subjects with normal weight, the snack consumption increased fullness by 52% (p < .001) and reduced hungry and prospective eating by 24 (p < .001) and 19% (p < .001), respectively, compared to white bread ( Figure  5). On the other hand, for participants with overweight, subjective satiety did not significantly differ between preloads ( Figure 6). As expected, the glycemic response was lower for common bean baked snack (CBBS) and cooked beans compared to white bread in both studies (p < .001). No significant differences were found for palatability attributes in adults with normal weight (Table  3). In contrast, subjects with overweight did perceive sensorial discrepancies. Their most pleasant preload was cooked beans, and white bread was the lowest.

Study 3: glycemic index
Ten subjects, five females and five males were recruited and completed the sessions (Table S2). The  CBBS blood glucose concentrations peaked at 45 min with a mean value of 109 mg/dL ( Figure S1). The mean within-subject CV of the iAUC (0-120 min) for the glucose test was 24%. The glycemic index (GI) of the CBBS was 42 ± 6, and according to the ISO 26642-2010 recommendation for food classification, the snack can be considered a low-GI product.

Discussion
Few studies have assessed the consumption influence of pulse snacks on satiety and energy intake. Similar results were reported by Reister and Leidy (2020), who examined the effect of 240 kcal of hummus with a serving of pretzels, a granola bar, or no snack in 39 healthy adults. A reduction in hunger and desire to eat after consumption of hummus was found, but not in energy intake. Diversely, Johnston et al. (2021) evaluated in 26 healthy adults the acute effects of 50 g of corn extruded snacks and extruded snacks in which 40% of the corn was replaced by whole yellow peas, split yellow peas, green lentils, chickpeas, or pinto bean flour. There were no differences in food intake or appetite, probably because the number of pulses was insufficient or because the texture may have been indistinguishable from each other. Food texture could affect satiety and energy intake by determining the eating rate. Hard foods are eaten more slowly than soft foods. Evidence suggests that hard textures lead to higher satiety and lower energy intake (Mccrickerd and Forde 2016;de Graaf and Boesveldt 2017).
Another attribute of the CBBS responsible for the satiety increment is its nutritional profile. The snack contains high protein and dietary fibre levels, key  nutrients that promote satiety after consumption. The protein fraction is considered the macronutrient with the highest satiating effect. Particularly proteins from pulses have been shown to stimulate the secretion of anorexigenic hormones such as cholecystokinin (CCK) (Overduin et al. 2015;Santos-Hern andez et al. 2018). Meanwhile, dietary fibre directly affects satiety and energy intake by decreasing the eating rate, delaying gastric emptying, slowing intestinal mobility of food and stimulating satiety signals. Additionally, dietary fibre indirectly promotes the production of glucagon-like peptide 1 (GLP-1), peptide tyrosine-tyrosine (PYY) and CCK through their fermentation products, the short-chain fatty acids, which act as ligands to receptors on cells of the intestinal endocrine system (Munekata et al. 2021). However, the last mechanism could not be responsible for the outcomes in this study due to the short testing period. To our knowledge, no studies have assessed the satiety and food intake effect of common beans in subjects with normal and overweight to observe differences. As mentioned above, the consumption of CBBS did not affect the satiety in participants with overweight. Individuals with overweight and obesity seem to lose their capacity to respond to satiety signals due to alterations in the gastrointestinal and adipose tissue hormones, including the blunted postprandial response of GLP-1, PYY and CCK and reduced  postprandial suppression of ghrelin and leptin resistance. A more significant quantity of CBBS may be needed to increase satiety in individuals with overweight (Lean and Malkova 2016). Previous evidence supports the pulses for improving satiety and appetite. A systematic review and meta-analysis that evaluated nine trials reported a 31% increment in subjective satiety after a meal of pulses compared with a control meal, where the median dose was 160 g [IQR 12 -276 g] (Li et al. 2014). Since cooked bean consumption did not affect satiety or energy intake in both studies, it can be considered that the 40 g portion of cooked beans was insufficient. Furthermore, regardless of the energy intake outcome, a decreased appetite by pulse intake may facilitate long-term body weight management (Kim et al. 2016;Clark and Duncan 2017). Finally, physical activity and thirst were similar across sessions in each study; therefore, they did not influence the outcomes.
Despite differences in palatability between cooked beans and white bread for participants with overweight, a consensus states that palatability impacts satiation (meal completion process) more than subsequent satiety (appetite suppression after ingestion). Consequently, these variabilities could not have altered satiety responses (McCrickerd 2016).
The low GI of both CBBS and cooked beans was responsible for the low glycemic response. Besides, although the concentration of available carbohydrates was higher in CBBS, the glycemic response was as low as cooked beans. It is well-documented that pulse consumption produces a low glycemic response and improves long-term glycemic control in adults with and without type 2 diabetes (Mollard et al. 2014;Hafiz et al. 2022). The main mechanisms for these glycemic responses are based on the complex structure of starch, the dietary fibre content and the inhibition of the glycosidases activity, which delay carbohydrate digestion and thus glucose absorption (Barrett and Udani 2011;Binou et al. 2022).
The GI of the CBBS is lower than those of conventional snacks. For instance, potato chips, cheese puffs and popcorn are medium-GI foods with mean values of 56, 63 and 62, respectively (Atkinson et al. 2021). Moreover, pulse enrichment of conventional snacks has reduced the GI successfully since pulses are low-GI with a mean value of 34 (Atkinson et al. 2021;Binou et al. 2022). As an example, Zambrano et al. (2013) changed the GI category of a cereal bar from 66 (medium-GI) to 49 (low-GI) by replacing 30% of cereals with common beans. Furthermore, Flores-Silva et al. (2015) formulated a chickpea-based gluten-free snack composed of chickpea flour (60%), maize flour (30%) and unripe plantain flour (10%). The snack showed a predicted GI of 28 by an in vitro starch hydrolysis index assay.
Although the study has some strengths, such as an appropriate design, an independent evaluation of subjects with normal weight and overweight, and a short period following preload consumption, some factors limited it. This trial could be underpowered to observe differences in food intake. Besides, caloric and volume differences could have contributed to the effects of cooked beans and the bean snack on subjective satiety, energy intake and glucose response. Including a control dinner before the session and scheduling the session according to the variation of hormones in participants with a premenopausal state could improve the study design. Lastly, quantifying satiety and appetite hormone responses would have provided more detailed outcomes.

Conclusion
In conclusion, the consumption of low glycemic index common beans as a baked snack increased subjective satiety in adults with a normal weight. Long-term trials assessing the effects on body weight management are necessary. Furthermore, the common bean baked snack is an appropriate low glycemic index option to avoid blood glucose spikes due to its dietary fibre and complex carbohydrate content. In addition, similar studies are required in populations with obesity or type 2 diabetes. Finally, common bean-based snacks have been demonstrated to be well-accepted, highly nutritious, and could be considered as functional foods. This snack alternative could increase pulse intake and satisfy the consumer demand for healthier snacks.

Disclosure statement
The authors report there are no competing interests to declare.