Viscous or gel-forming dietary fibers can increase satiety by a more firm texture and increased eating time. Effects of viscous or gel-forming fibers on satiety by post-ingestive mechanisms such as gastric emptying, hormonal signals, nutrient absorption or fermentation are unclear. Moreover, it is unclear whether the effects persist after repeated exposure.
To investigate satiety and energy intake after single and repeated exposure to gelled fiber by post-ingestive mechanisms.
In a two-arm crossover design, 32 subjects (24 female subjects, 21±2 y, BMI 21.8±1.9 kg m−2) consumed test foods once daily for 15 consecutive days, with 2 weeks of washout. Test foods were isocaloric (0.5 MJ, 200 g) with either 10 g gel-forming pectin or 3 g gelatin and 2 g starch, matched for texture and eating time. Hourly satiety ratings, ad libitum energy intake and body weight were measured on days 1 (single exposure) and 15 (repeated exposure). In addition, hourly breath hydrogen, fasting glucose, insulin, leptin and short-chain fatty acids were measured.
Subjects rated hunger, desire to eat and prospective intake about 2% lower (P<0.015) and fullness higher (+1.4%; P=0.041) when they received pectin compared with control. This difference was similar after single and repeated exposure (P>0.64). After receiving pectin, energy intake was lower (−5.6%, P=0.012) and breath hydrogen was elevated (+12.6%, P=0.008) after single exposure, but not after repeated exposure. Fasting glucose concentrations were higher both after single and repeated exposure to pectin (+2.1%, P=0.019). Body weight and concentrations of insulin, leptin and short-chain fatty acids did not change during the study.
Gelled pectin can increase satiety and reduce energy intake by post-ingestive mechanisms. Although the effects were small, the effects on satiety were consistent over time, whereas the effects on energy intake reduction were not.
Observational studies have shown that dietary fiber intake is inversely associated with body weight.1, 2 A reduction in body weight is likely mediated by a reduction in appetite and energy intake. Dietary fiber, however, is a term that reflects a heterogeneous group of compounds, which differ in their chemical structure and physicochemical properties.3 As a result, different types of fiber may have different effects on appetite and energy intake.4, 5, 6
Earlier research showed that viscous or gel-forming fibers, such as pectin and guar gum, can increase satiety and reduce subsequent energy intake.7, 8, 9 Viscous or gel-forming fibers may affect satiety by various mechanisms, which include increased oral exposure (i.e., presence of food in the oral cavity), delayed gastric emptying, modified neural and hormonal signals in the gut, slowed down or diminished absorption of nutrients, or an altered fermentation pattern in the large intestine.4, 5, 6
In many studies on viscous or gel-forming fiber, the control treatment was a liquid.7 Subsequent effects on satiety may therefore be explained by an increased viscosity or thickness of the test food10 (Wanders et al., submitted). Increased eating time and a subsequent increase in oral exposure have been repeatedly shown to increase satiety11, 12 and can also occur after consuming non-fiber thickeners. In contrast to non-fiber thickeners, such as starch and gelatin, which are broken down by digestive enzymes, viscous or gel-forming fibers stay intact and may affect additional satiety-related mechanisms in the gastro-intestinal tract. Yet, it is unclear what the effects of viscous or gel-forming fibers are beyond oral exposure. To our knowledge, one earlier study compared the effect on satiety between a gelled fiber and an equal gelled non-fiber control.13 This study showed that over a 3.5-h period appetite for a meal reduced by approximately 5% after consumption of the gelled fiber, without a change in subsequent energy intake.
The first aim of the present study was to assess satiety and energy intake after consuming gelled fiber via post-ingestive mechanisms. Matching fiber and control test foods for texture and eating time enabled studying satiety mechanisms in the gastro-intestinal tract. The second aim of the study was to explore whether the effects after single exposure would persist after repeated exposure. There is an ongoing debate whether foods increasing satiety and lowering subsequent energy intake after single exposure result in sustained effects after repeated exposure.14, 15 At present, only a limited number of studies on repeated exposure are available on viscous or gel-forming fibers;16, 17, 18 however, none of these studies were designed to compare single with repeated exposure. The present study comprises a two-arm crossover design, in which subjects consumed each test food for 15 consecutive days.
Subjects and methods
Thirty-five healthy, normal weight (BMI 18.5–25.0 kg m−2) subjects aged between 18 and 30 y were locally recruited. Subjects were excluded if they scored high on restrained eating (DEBQ: score men>2.89, women>3.39).19 They were also excluded if they used an energy-restricted diet, lost or gained more than 5 kg body weight in the last 2 months, had a lack of appetite, had stomach or bowel diseases or disorders, used antibiotics or dietary fiber supplements in the last 2 months, were hypersensitive for any ingredient in the test foods or were vegetarian. They were further excluded if they had diabetes or any other endocrine disorder, had fasting glucose concentrations >5.8 mmol l−1 or if they were smokers or heavy alcohol users (>5 drinks a day). Thirty-two subjects, of which 24 were women, completed the study. They had a mean age of 21.1±2.4 y, a mean BMI of 21.8±1.9 kg m−2 and a mean DEBQ score of 1.7±0.5 for men and that of 2.2±0.5 for women. One subject dropped out during the run-in because of disliking the test food, one dropped out at day 13 (control) because of problems with the ad libitum diet and one dropped out at day 13 (control) because of scheduling difficulties.
The study was approved by the Medical Ethics Committee of Wageningen University and was registered at ClinicalTrials.gov under number NCT01526759. Written informed consent was obtained from all subjects.
In a double-blind, two-arm randomized crossover design, subjects consumed test foods for 15 consecutive days, with a 2-week washout (Supplementary Material Table 1). Measurements were recorded on day 1 (single exposure) and on day 15 (repeated exposure). Test foods were consumed daily between 10.30 and 11.00 hours and contained either gelled pectin or a gelatin-starch blend as a control with similar consistency. During the run-in day, all subjects consumed the control, and during the intervention period subjects were randomly assigned to receive either the pectin or the control. An independent researcher randomized subjects by means of computer-generated random numbers in permuted blocks of size 2. To assess the total energy intake during the first day and the last 3 days of both the intervention periods, subjects consumed ad libitum diets. During the other days, subjects were instructed to consume their habitual diet.
Test foods were isocaloric dietary supplements (58 kcal 100 g−1) equal in consistency and eating time (Table 1) and were especially developed for this study (NIZO food research BV, Ede, the Netherlands). The test foods were served in portions of 200 g in non-transparent plastic cups with spoons. The pectin test food contained (g 100 g−1) 5g pectin (Classic CU901, LM-pectin, DE=10%, molecular weight=max 15kDa; Herbstreith & Fox, Neuenbürg/Württ, Germany), 1.5 g starch, 10.4 g sugar, 0.12 g calcium, 0.4 g citric acid and 0.1 g vanilla aroma. The control test food contained (g per 100 g) 1.5 g gelatin, 3.5 g starch, 10.4 g sugar, 0.12 g calcium, 0.5 g citric acid and 0.1 g vanilla aroma. To calculate available energy, Atwater factors were used for protein and carbohydrate (17 kJ g−1), and fiber was estimated as 8.5 kJ g−1.20 The pectin dose was based on our previous study (Wanders et al., submitted). Table 2 provides the sensory properties of the test foods obtained from an independent panel.
Stiffness of the test foods as consumed and viscosity and water holding capacity in simulated mouth and stomach conditions were measured (Table 1). Stiffness was measured as the slope of the penetration curve at a speed of 0.3 mm s−1 7 °C and at 18 °C in g mm−1 (Texture Analyzer, type TA-XT2, Stable Micro Systems. Godalming, UK). Viscosity and water holding capacity in simulated mouth and stomach conditions were measured according to methods described earlier,21, 22 with adaptations for the high water content of the products.
To assess the total energy intake, all meals were served ad libitum at the research center at the run-in day, at day 1 and at days 13 to 15 of both intervention periods, including breakfast at days 2 and 16. The morning and evening meals were served in buffet style. Hot meals at lunch were individually served in portions corresponding to 200% of the energy content of a normal Dutch hot meal.23 A package with snacks could be taken out for the afternoon and evening. Subjects were not allowed to eat other foods than what was provided. The variety of foods served was limited to everyday foods to prevent subjects from overeating. The buffet at breakfast and evening consisted of wholemeal mini buns, crackers, low-fat margarine, two types of cheese, two types of meat slices, four types of sweet fillings and three types of fruit. The hot lunch was a mixed meal comprising pasta or rice, meat and cooked vegetables, served with a raw vegetable salad and a fruit salad as dessert. As afternoon and evening snacks, we provided apples, gingerbread and currant buns. Subjects were only allowed to drink water, tea and coffee. Other liquid and semi-solid foods were not allowed, as liquid calories have low satiating value.12, 24 Food intake over the day was measured by weighing all foods before serving and weighing any leftovers. Energy and macronutrient intake was estimated using the Dutch Food Composition Database.25 Intakes were calculated for day 1 and day 15 (mean day 13 to 15).
On day 1 and on day 15, each subject completed hourly satiety ratings over a 14-hour period using a Personal Digital Assistant (PDA, HP IPAQ) with software (EyeQuestion version 3.8.3, Logic8 BV, Elst, the Netherlands). In the morning, subjects received the PDA, and they rated hunger, fullness, desire to eat, prospective food consumption and thirst on the basis of an 11-point Likert scale anchored from ‘not at all’ to ‘very much’. Satiety ratings were completed before and after the three meals and further every hour until 22:15 hour. An hourly alarm was programmed to remind the subjects. Satiety ratings were analyzed for the total day and separately for morning, afternoon and evening. Moreover, ratings over the 2 hours after test food intake were also analyzed.
Parallel to the satiety ratings on day 1 and on day 15, the subjects collected hourly alveolar breath samples over a 14-hour period using reusable 250-ml sample holding bags (Quintron, Milwaukee, USA). Samples were collected before breakfast, after lunch, after dinner and further every hour until 21:15 hours. Breath samples were analyzed for hydrogen content, as an indicator for colonic fermentation, using a Quintron Microanalyzer (Quintron Instruments, Milwaukee, USA). Measurements of day 15 in intervention period 1 and that of day 1 in intervention period 2 were discarded from the data set, because of an incorrect way of storing the sample holding bags.
To assess fasting glucose, insulin, leptin and short-chain fatty acid concentrations, after day 1 and day 15, blood samples were collected into 6 ml EDTA-containing tubes. Tubes were centrifuged at 1000 g/2216 rpm for 10 min at 4 °C and stored at −80 °C until analysis. Glucose was measured by the hexokinase method (Modular P800 analyzer, Roche, Basel, Switzerland). Detection limit was 0.11 mmol l−1, and intra-assay and inter-assay coefficients of variation (CVs) at 3.66 mmol l−1 were 1.1 and 1.9%, respectively. Insulin and leptin were measured using commercially available human ELISA kits (Mercodia, Uppsala, Sweden). For insulin, the assay had a detection limit of 0.1 mU l−1 and intra-assay and inter-assay CVs at 11 mU l−1 were 3.4 and 3.6%, respectively. The detection limit for leptin was 0.05 μgl−1 and the intra-assay and inter-assay CVs at 1.88 μg l−1 were 2.3 and 5.2%, respectively. Concentrations of the short-chain fatty acids acetate, butyrate and proprionate were assessed by 1H-NMR spectroscopy according to methods described earlier.26
Body weight was assessed each morning before ad libitum breakfast, without wearing shoes and heavy clothing. Subjects were instructed not to change their physical activity, which was monitored by step counts (Yamax Digi-walker, SW-200, Tokyo, Japan). Baseline body weight was calculated as the mean of the measurements at run-in and day 1. End body weight (day 15) was calculated as the mean of day 14 to 16. The similar procedure was followed for step counts.
During both intervention periods, subjects were asked to keep a diary to register the time of consumption of the test foods, adverse events (bloating, belching, flatulence, nausea, diarrhea or other adverse events), illness and the use of medication. Directly after both intervention periods, evaluation questionnaires were given to rate the test foods for palatability on an 11-point Likert scale anchored from ‘not at all’ to ‘very much’.
Data are presented as means with standard deviations unless reported otherwise. Statistical analyses were performed with SAS (version 9.2; SAS institute Inc., Cary, NC, USA). Significance was set at P<0.05. Test food effects per study day were analyzed by means of ANOVA (proc mixed, SAS). For the analysis of total daily energy intake, fasting glucose, insulin, leptin and short-chain fatty acid concentrations, body weight and step count, variables for test food, intervention period, day, test food*day and test food*intervention period (=order) were included as fixed factors, and a variable for subject was included as a random factor. After confirming that on run-in days the energy intake, fasting glucose, insulin and leptin concentrations were similar, the run-in data were removed from the dataset. Repeated measurements for satiety ratings were analyzed according to a similar procedure, with the addition of time and test food*day*time as a fixed factor in the model. The Likert scales were treated as continuous variables. Sensory and palatability ratings were analyzed according to a similar procedure. Hydrogen excretion data were not normally distributed and were therefore log transformed for analysis and presented as back-transformed geometric means. As a result of discarding hydrogen data from day 15 in period 1 and day 1 in period 2, the variable subject was included as a covariate rather than a random factor. For adverse events and illness, the number of subjects reporting an event at least once was counted and analyzed by the χ2 test.
Subjects complied well with the instructions to consume the test foods each day between 10:30 and 11:00 hours. From the daily records, it was shown that 86% was consumed in time, 11% within 30 min of the instructed time and 3% more than 30 min late or early. Reasons to deviate from the instructed time were mostly events not allowing food intake.
The test foods were comparable with respect to energy content, stiffness and eating rate (Table 1) and for the sensory ratings thickness and melting behavior (Table 2). There were differences in some of the sensory properties, for example, the pectin test food was rated as more sticky and more mouth filling compared with control. Moreover, palatability of the pectin test food was significantly lower compared with control (P<0.001). As hypothesized, physicochemical properties of the two test foods in simulated upper gastrointestinal conditions were different (Table 1). The pectin test food had a fourfold higher viscosity in mouth conditions, an eightfold higher water binding capacity in mouth conditions and a 10-fold higher water binding capacity in stomach conditions.
After single exposure (day 1) to the pectin test food, total daily energy intake was significantly lower than that after the control test food (−5.6%, P=0.013) (Table 3). Both fat (−9.3%, P=0.014) and carbohydrate (−5.4%, P=0.030) intakes were lower compared with the control. After repeated exposure (day 15) to the pectin test food, energy intake was similar to when they received the control (P=0.62). At the different meal times, there were no differences between the two test foods in energy intake (test food*day*meal time interaction: P=0.59). Post hoc tests, however, revealed a borderline significant difference for the evening meal on day 1: energy intake was 3.01±0.86 MJ after the pectin test food and 3.31±1.01 MJ after control (−9.1%, P=0.065).
Satiety ratings reflected the cycle of three meals a day clearly (time-effect: all P<0.0001) (Figure 1). Overall, subjects rated hunger lower (−1.7%; test food-effect: P=0.007), fullness higher (+1.4%; P=0.041), desire to eat lower (−1.7%; P=0.014) and prospective intake lower (−1.8%; P=0.003) when they received the pectin test food compared with when they received control. This difference in satiety rating was comparable after single and repeated exposure to the test foods (test food*day interaction: all P>0.65). Over the total mornings (from breakfast to lunch time), satiety ratings were not different between the test foods, whereas right after consumption of the test foods, (from 11:15 hours until lunchtime), there were differences. In this period, subjects rated hunger lower (−3.3%; test food-effect: P=0.029), desire to eat lower (−3.5%; P=0.020) and prospective intake lower (−3.5%; P=0.010) after the pectin test food compared with control. This difference was comparable after single and repeated exposure (test food*day interaction: all P>0.54). In the evening, subjects rated fullness higher (+2.9%; P=0.017), desire to eat lower (−3.7%; P=0.006) and prospective intake lower (−2.2%; P=0.029) after the pectin test food compared with the control, which was comparable after single and repeated exposure (test food*day interaction: all P>0.31). The two test foods had no different effects on thirst ratings.
Baseline body weight was similar for the two groups (P=0.99) and did not change over the 2-week period when subjects received the pectin test food (day 15: 66.5±9.7 kg) compared with the control (day 15: 66.5±9.7 kg) (test food*day interaction: P=0.60). Step counts also did not differ between the two groups (data not shown).
After single exposure to the pectin test food, breath hydrogen was elevated compared with the control (P=0.0075) (Figure 2). On day 15, breath hydrogen excretion was not different between the pectin test food and the control (P=0.13). When the subjects received the pectin test food, fasting glucose concentrations were higher (test food-effect: 5.1±0.4 mmol l−1, P=0.019), compared with when they received control (5.0±0.4 mmol l−1). This did not change after single or repeated exposure to the test foods (test food*day interaction: P=0.66). No differences were seen in fasting insulin and leptin concentrations (Table 4) and in fasting short-chain fatty acids (data not shown).
Adverse events and other illnesses possibly related to the test food intake were registered in daily records. During the period in which the control and pectin test foods were consumed, bloating was reported by 8 vs. 13 subjects (P=0.18); belching, by 3 vs. 2 subjects (P=0.64); flatulence, by 9 vs. 12 subjects (P=0.42); nausea, by 5 vs. 3 subjects (P=0.45); and diarrhea, by 4 vs. 5 subjects (P=0.72), respectively.
The present study investigated satiety and energy intake after single and repeated exposure to a gelled fiber. To gain insight into the post-ingestive satiety mechanisms, the test foods were matched for texture and eating time. We found that after single exposure to a 10 g dose of gelled pectin, satiety ratings were higher and subsequent energy intake was lower compared with the control. Over the 15-day study period, the increased satiety ratings persisted, whereas the energy intake reduction did not. The results of the present study show that gelled pectin can increase satiety by other means than a change in texture and eating time alone. Moreover, the study shows that, even though the effects are small, the increased satiety persists over time.
State-of-the-art methods were used to measure satiety and energy intake under free living conditions.27 Satiety ratings were measured with PDAs hourly over a full day, and energy intake was measured with an ad libitum diet over multiple days according to standardized methods. Moreover, subjects were not allowed to consume liquids containing calories, as these have a low satiating value.12, 24 The development of our test foods was restrained by structural differences between gels formed by pectin and gels formed by non-fibers. Therefore, not all sensory properties could be made identical and some differences were observed in palatability. Palatability and sensory properties such as taste intensity have shown to affect within-meal satiation.28, 29 However, studies on the effects of palatability and taste intensity on between-meal satiety have shown inconsistent findings.28, 30 On the basis of these earlier studies, it cannot be established whether test food properties that affect oral exposure, other than texture and eating time, have affected the findings in the present study.
Our findings show that both after single and repeated exposure the pectin test food increased satiety ratings with about 3.5% over the 2 hours after consumption, and with about 2% over the whole day, compared with the control that was equal in texture and eating time. In an earlier study in which the same gelled pectin was compared with a liquid control, the 3-hour satiety ratings increased with about 7% (Wanders et al, submitted). Although the effects were small, the present findings suggest that the satiety effects of gelled pectin may also be attributed to effects other than an increased oral exposure. Suggestions for other mechanisms are post-ingestive effects such as delayed gastric emptying, changed appetite regulating neural and hormonal signals in the gut, lower or slowed down absorption of nutrients and altered fermentation products in the large intestine.4, 5, 6
The primary difference between the test foods was the consistency right after ingestion. In simulated mouth and stomach conditions, the control had a lower viscosity and water holding capacity than the pectin test food. This difference can be explained by both the presence of salivary amylase and a difference in sensitivity to temperature. In contrast to starch that is broken down by enzymes, and gelatin that starts to melt in the mouth, pectin is unaffected by enzymes and melts in the mouth to a lesser extent. This difference in texture after ingestion may result in a change in gastric emptying rate. In an earlier study, we showed that a single dose of the same gelling pectin resulted in a reduced gastric emptying rate (Wanders et al, submitted). Moreover, Schwartz et al, showed that repeated pectin consumption persistently delayed gastric emptying, which may be due to adaptive changes.31 A delay in gastric emptying may enhance gastric mechanoreceptor stimulation and as a result prolong fullness feelings.32
Both a delayed gastric emptying and an increased water holding capacity of intestinal content may slow down nutrient absorption in the small intestine. As a result, release of neural and hormonal signals in the gut may be prolonged and lead to longer satiety.33 In the present study, we did not measure post-prandial release of signals in the gut, but we did measure fasting glucose, insulin and leptin. Insulin and leptin concentrations did not change during the study, but fasting glucose concentrations showed a small but significant increase of 0.1 mmol l−1 over the study period. This is an unexpected finding as viscous fiber is generally found to lower post-prandial glucose concentrations,33 and earlier studies showed no changes in fasting glucose levels.34, 35 However, as we studied young healthy adults with normal baseline glucose concentrations, this might be a chance finding. For future studies, we recommend to measure the post-prandial release of appetite signals, such as glucose and insulin, in the gut.
A third process that may explain the satiety-increasing effect of the pectin is fermentation in the large intestine. Three hours after consuming a single dose of pectin, breath hydrogen excretion showed a clear peak, which suggests an increase in colonic fermentation.36 Whereas hydrogen on itself is probably not directly associated with appetite,36 altered activity of gut microbiota may promote the production of short-chain fatty acids,37, 38 which has been related to an increased satiety and a reduced body weight.39, 40 In the present study, the increase in hydrogen production did not persist over time, which was confirmed by the absence of differences in fasting concentrations of short-chain fatty acids. In contrast to findings in the present study, in other studies, fermentable fibers have resulted in sustained changes in hydrogen excretion.39, 41 The inconsistency may be explained by dissimilar fermentation patterns in the large intestine after different types of fermentable fiber.42, 43, 44 After a sustained pectin consumption, adaptation of bacterial metabolism toward efficient fermentation pathways not producing hydrogen or short-chain fatty acids may have taken place.45
After single exposure to the pectin test food, energy intake reduced by about 6%, which was not observed at the end of the study period, nor was this reflected in a reduction of body weight. This finding did not match with the persistent increase in satiety that was observed throughout the 15 days of intervention. We hypothesize that, to affect actual eating behavior, larger changes in satiety ratings may be needed. It was suggested that an 8–10% change in satiety ratings is a relevant effect size,27, 46 which may also be a relevant effect size to influence energy intake. A recent study with a fermentable, but not viscous or gel-forming wheat dextrin supports this suggestion.47 Over 2 weeks of supplementation with dosages of 8, 14, 18 and 24g per day, all dosages increased satiety ratings compared with 0g per day. Interestingly, only the dosages of 14g and higher significantly lowered the body weight.48 These results suggest that to achieve a persistent reduction in energy intake, larger increases in satiety should be aimed for, which may be achieved by a higher dosage of dietary fiber.
In a previous paper, we suggested that, after repeated exposure, other metabolic processes may add toward sustained satiety or reduced longer-term energy intake (Wanders et al, submitted). In the present study, however, no further changes were found after repeated exposure. On the basis of an array of studies (e.g., as reviewed in7), we postulate that dietary fibers can have both ‘mechanical’ effects and ‘colonic’ effects on appetite. The mechanical effects can, for example, be oral exposure, gastric emptying rate and nutrient absorption. These effects are mediated by food matrix or intestinal content characteristics, such as viscosity and accessibility of nutrients in the lumen, and can generally be provoked by relatively small dosages of fiber. In contrast, colonic effects may depend highly on fiber-specific properties such as prebiotic activity, composition of gut microbiota and fermentation end products. The effective dosage of fiber that may induce a change in colonic effects depends largely on the specific fiber type characteristics. This may explain why a gelled and highly fermentable pectin does not change energy intake after repeated exposure, whereas wheat dextrin, a fiber that is not viscous and does not form gels, but is fermentable, has large effects on energy intake.47 This suggestion, however, is preliminary and should be verified with additional research.
We conclude that a relatively small dosage of gelled pectin can increase satiety and reduce energy intake by other means than texture and eating time. Although the effects were small, the effects on satiety were consistent over time, whereas the effects on energy intake reduction were not. Further research with larger dosages and a longer intervention period is needed to strengthen the evidence for the post-prandial effects of gelled fiber.
We would like to thank Lydia van Es, Matty Karsten, Annemieke Janszen and Roxanne van Oeveren for their assistance; Melliana Jonathan and Henk Schols for their advice; and Jacques Vervoort, Anita Bruggink-Hoopman, Nhien Ly, Marja Kanning and Peter de Gijsel for their technical support. This work was funded by the Dutch Ministry of Economic Affairs, Agriculture and Innovation (project KB-05-009-003).
About this article
Supplementary Information accompanies this paper on International Journal of Obesity website (http://www.nature.com/ijo)