Appetite is regulated by many factors, including oro-sensory and gastric signals. There are many studies on contributions of and possible interaction between sensory and gastric stimulation, but there are few studies in humans using simultaneous oral and gastric stimulation. We investigated the effect of simultaneous, but independently manipulated, oral and gastric stimulation on appetite ratings and energy intake. We hypothesized that compared with no stimulation, oral and gastric stimulation would equally and additively decrease appetite ratings and energy intake. Healthy men (n = 26, 21 ± 2 years, BMI 22 ± 3 kg/m2) participated in a randomized crossover trial with four experimental conditions and a control condition. Experimental conditions consisted of oral stimulation, with either 1 or 8 min modified sham feeding (MSF), and gastric stimulation, with either 100 or 800 ml intragastrically infused liquid (isocaloric, 99 kcal, 100 ml/min). The control condition consisted of no oral or gastric stimulation. Outcome measures were energy intake 30 min after the treatment and appetite ratings. Compared with the control condition, energy intake decreased significantly after the 8 min/100 ml (19% lower, P = 0.001) and 8 min/800 ml conditions (15% lower, P = 0.02), but not after the 1 min/100 ml (14% lower, P = 0.06) and 1 min/800 ml conditions (10% lower, P = 0.39). There was no interaction of oral and gastric stimulation on energy intake. Hunger and fullness differed across all conditions (P ≤ 0.01). In conclusion, duration of oral exposure was at least as important in decreasing energy intake as gastric filling volume. Oral and gastric stimulation did not additively decrease energy intake. Longer oro-sensory stimulation, therefore, may be an important contributor to a lower energy intake.
The food supply in the industrialized world is characterized by a large variety of highly palatable foods, which are often energy dense and easy to consume. Foods that are high in energy density and that are consumed quickly may contribute to a positive energy balance because they have low-satiating efficiency per calorie (1,2,3).
Long-term studies suggest that addition or removal of liquid calories from the diet does not lead to compensatory changes in energy intake (4,5,6). The low-satiating capacity of caloric beverages may for a large part be due to their swift passage through the mouth (3,7,8). In line with this notion, it has been shown in a number of studies that the duration of oro-sensory signals contributes to the satiating efficiency of foods, implying that a longer sensory exposure time leads to a lower food intake (7,8,9,10).
On the other hand, it is also clear that the amount of food in the stomach influences the regulation of food intake. It has been found that factors like volume and weight of the gastric load, and to a lesser extent energy content, influence the short-term regulation of food intake (11,12). In their exhaustive review, Poppit and Prentice suggested that also on the longer term, people eat a constant weight or volume of food, and not a constant amount of energy (13). This idea has been confirmed by a number of studies of Rolls and colleagues (1,2,12,14) and implies that a higher energy density will lead to a higher total energy intake.
A large number of human studies using only oral stimuli (15,16,17,18) or gastric stimuli (19,20,21,22,23) showed that both oral and gastric stimuli contribute to satiety. Several studies investigated this (24,25,26), but not simultaneously, because this is difficult to do in humans.
Therefore, the objective of this study was to investigate the contribution of oral and gastric stimulation to satiety simultaneously, measured as energy intake and appetite ratings. To this end, gastric and oral stimulation were manipulated independently: gastric stimulation was done via a nasogastric tube, and oral stimulation by using modified sham feeding (MSF) (17,18).
We hypothesized that oral and gastric stimulation would equally and additively contribute to a lower subsequent energy intake and that they would decrease appetite ratings. Energy intake and appetite ratings were predicted to be lowest when long oral stimulation is combined with a large gastric filling.
Methods and Procedures
Healthy young men were recruited from Wageningen and surroundings via posters, a mailing list and flyers. Exclusion criteria were BMI <20 and ≥25 kg/m2, change in body weight of ≥2 kg in the last month, restraint eating score ≥2.26 as measured with the Dutch Eating Behavior Questionnaire (27), smoking, any medication or drug use, gastrointestinal diseases, diabetes, thyroid diseases or any other endocrine disorders, lack of appetite and hypersensitivity or food allergy for products used in the study.
In total, 43 men were included for a training session. Subjects had to perform the study procedures successfully (see study procedure section) to be further enrolled in the study. After the training session, four subjects decided to stop and three subjects were excluded because their performance of MSF was insufficient (see study procedure section). The remaining 36 subjects were enrolled in the study, during which four subjects withdrew because of personal reasons, one dropped out because he started on medication and one stopped because of an adverse event (not related to this study).
Thirty subjects completed all experimental sessions. Data of four subjects were excluded: one because of a problem with the tube and pump during one condition, one due to missing data of one of the conditions, and two because the MSF performance was too low for one of their conditions (both 83% recovery of dry mass). Data of the remaining 26 subjects were used in the data analyses.
Subjects were unaware of the exact study aim. They were told that the aim was to examine the effect of oro-sensory stimulation on subsequent digestion of intragastrically infused food. Before the training session, subjects signed an informed consent form. The Medical Ethical Committee of Wageningen University approved the study protocol (NL30728.081.09) and the trial was registered with the Dutch trial register at www.trialregister.nl as NTR2173. All subjects received financial compensation.
The study had a randomized crossover design with five conditions varying in duration of oral stimulation and volume of the intragastric load:
5. No oral or gastric stimulation, but with a nasogastric tube inserted (control condition);
Oral stimulation and intragastric infusion started simultaneously. The rate of infusion was always 100 ml/min, therefore the duration of infusion was 1 min for the 100 ml volume, and 8 min for the 800 ml volume. The five conditions were randomized by means of a Latin Square (Williams design).
Training session. All subjects participated in a training session where MSF was performed (8 min with the test cake) and a 400-ml cake-solution was infused via a nasogastric tube (99 kcal, 100 ml/min). When this was well-tolerated and cake recovery after MSF was ≥85% dry mass, the training was considered successful. To determine the amount of cake accidentally swallowed, the dry mass of the expectorants was compared with the dry mass of the provided amount of cake. The threshold for recovery of cake during MSF was set at ≥85% dry mass. This threshold is a feasible percentage which excludes subjects who encounter difficulties expectorating the cake. Other studies with MSF report a recovery of 89–97% of food in the expectorants (15,28,29).
Furthermore, height and weight of the subjects were measured at the start of the training session.
Experimental sessions. Subjects came to the laboratory on five nonconsecutive mornings with a wash-out period of at least 6 days. Arrival time for a subject was the same for every session and varied between 9.30 and 12.00 h. Subjects were instructed to avoid intensive physical activity from the preceding evening onwards and to eat their breakfast 2.5 h before their arrival. After breakfast subjects were not allowed to eat or to drink energy-containing beverages. Upon arrival any consumption was prohibited. Subjects kept a diary of their breakfast and beverage consumption on each study day to check compliance with the instructions. Within 30 min after arrival a nasogastric tube was inserted (between t = −60 and t = −30 min, Figure 1). One hour after their arrival, baseline feelings of appetite were measured by means of Visual Analogue Scales (t = 0, Figure 1). Subsequently, subjects received the gastric load and performed the MSF, both starting at the same time. Subjects were instructed and supervised individually. All procedures took place in one room which was partitioned into smaller spaces to give the subjects privacy during their visit.
Test foods. Commercially available cake (Grootmoeder's cake; Albert Heijn, Zaandam, the Netherlands) was used for the oral stimulation and the intragastric load. Each study day a dietician prepared the required foods according to the center's hygiene guidelines. Energy content and nutrient composition of the cake and the gastric loads are shown in Table 1.
Table 1. Energy (kcal/kJ) and macronutrient content of the cake and the gastric loads
MSF. Subjects were orally exposed to the cake for 1 or 8 min by means of a MSF paradigm. With this technique subjects chew on the food without swallowing it and then expectorate it. To allow subjects to chew similar to their normal chewing rate, they received a plate with an excess amount of cake established during the training sessions. This corresponded to ∼100 g for the 1-min MSF conditions, and 250 g for the 8-min MSF conditions. Three subjects needed 350 g of cake to have an excess amount during the 8 min MSF. Subjects were instructed to take a bite, chew on it and expectorate it at the moment they would normally swallow it. It was stressed that all cake should be expectorated. This was repeated until the end of the 1 or 8 min. After that, the subjects rinsed their mouth with 25-ml water which was also expectorated in the cup. Similar to the training session, recovery of cake after MSF had to be ≥85% dry mass.
Intragastric load. Gastric stimulation was done by tube feeding via a nasogastric tube (CH8–110 cm, external diameter 2.69 mm; Nutricia, Schiphol, the Netherlands). A nurse inserted the tube in accordance to a clinical intubation-protocol of a nearby hospital (Ziekenhuis Gelderse Vallei, Ede, the Netherlands). The distal end of the tube was positioned ∼75 cm from the nostril, exact distance depended on the subject's length. Correct placement of the tube was confirmed by a litmus test on fluid sucked via the tube by a syringe (pH <5.5 was correct).
Subjects received 100- or 800-ml cake-solution (isocaloric, 99 kcal, 100 ml/min) intragastrically (Table 1). The liquid consisted of cake blended with boiled water, which was cooled down to <7 °C directly after preparation. Shortly before start of the infusion (<15 min) the solution was heated to 37 °C and was kept at 37 °C in a stove. The solution was administered by means of a pump (520 U; Watson-Marlow, Cornwall, UK). As subjects could see whether their tube was filled or not, the pump was also activated during the control condition until the tube was filled with solution. The rate of the pump was calibrated every test day at 100 ml/min. The tube was removed between t = 8 and t = 15 min.
Ad libitum intake. Half an hour after the treatment started (t = 30, Figure 1), subjects were offered a test meal of which they could eat ad libitum for 30 min. In the Netherlands people usually eat bread with different toppings at lunch, therefore the test meal consisted of bread rolls, butter and commonly used sweet and savory toppings. The provided products and their energy content are presented in Table 2. All products were offered in unusual amounts or sizes, e.g., the bread rolls weighed 22 g instead of the usual 50 g. Subjects were instructed to eat until they were comfortably full and had to stay at the dinner table until the 30 min were over. Subjects could ask for more if they wanted.
Table 2. Macronutrient content and offered weights of the ad libitum test meal products
Appetite and wellbeing. To measure appetite subjects scored their hunger, fullness, prospective consumption, desire to eat, desire to eat something sweet, desire to eat something savory, and two wellbeing ratings—wellbeing and nausea—on 100-mm Visual Analogue Scales (30). This was done at five different moments: right before the treatment started (t = 0) and at 8, 15, 30, and 60 min after the treatment started (Figure 1). The scales were anchored “not at all” on the left and “extremely” on the right.
The analyses were performed using the statistical package SAS, version 9.1 (SAS Institute, Cary, NC). Data are presented as means ± SD and P values <0.05 were considered significant.
Sample size was calculated with a SD based on day-to-day variation in energy intake of 27% (31); and an estimated difference in energy intake—control condition vs. 8 min/800 ml condition—of 20%, based on data presented on the effects of increasing volume of foods on satiety sensations (11). With a power of 0.80 and a drop-out rate of ca. 10% it was estimated that enrollment of 35 men—after the training session—would be enough to detect this difference.
Ad libitum energy intakes in the five conditions were compared with an ANOVA (PROC GLM): with subject and condition as fixed factors. To investigate whether there was an interaction effect we compared the four conditions with oral and gastric stimulation with each other. This was done with a two by two ANOVA (PROC GLM): with subject, oral stimulation duration, gastric filling volume, and the interaction term “oral stimulation duration × gastric filling volume” included in the model. Similar analyses were done for weight of the food intake and for energy intake per test meal product.
Appetite and wellbeing ratings were compared across the five conditions with a mixed model ANOVA (PROC MIXED): with condition, time and condition × time in the model. To investigate whether there was an interaction effect we compared the appetite and wellbeing ratings of the four conditions with oral and gastric stimulation with a mixed model ANOVA (PROC MIXED): with oral exposure duration, gastric filling volume, time and their interaction terms in the model.
For significant overall effects, post hoc comparisons with Bonferroni correction for multiple comparisons were performed.
Data of 26 subjects were included in the data analysis. The subjects had a mean ± SD age of 21 ± 2 years, body weight of 75 ± 13 kg, BMI of 22 ± 3 kg/m2, and restraint eating score of 1.5 ± 0.4.
Recovery percentage after MSF procedure
An overall mean recovery percentage of 97.8 ± 2.7% was found, which means that on average 1.5 ± 2.6 g cake was swallowed. Per condition this was: 97.7 ± 2.4% (0.6 ± 0.6 g swallowed) for 1 min/100 ml; 98.5 ± 2.6% (1.8 ± 2.4 g swallowed) for 8 min/100 ml; 96.8 ± 2.9% (0.7 ± 0.6 g swallowed) for 1 min/800 ml; and 98.1 ± 2.8% (2.8 ± 4.3 g swallowed) for 8 min/800 ml.
Ad libitum intake
A significant difference in ad libitum energy intake between the five conditions was found (F4,100 = 4.64; P = 0.002; Figure 2). The post hoc analyses showed that the long oral stimulation conditions resulted in significantly lower energy intakes compared with the control condition: 19% lower energy intake in the 8 min/100 ml condition (T = −4.02; P = 0.001) and a 15% lower intake in the 8 min/800 ml condition (T = −3.20; P = 0.02). The short oral stimulation conditions did not differ significantly from the control condition: 14% lower energy intake in the 1 min/100 ml condition (T = −2.81, P = 0.06) and 10% lower in the 1 min/800 ml condition (T = −2.09, P = 0.39). Comparison of the four conditions with oral and gastric stimulation showed no interaction effect of oral and gastric stimulation on ad libitum energy intake (F1,75 = 0.00, P = 0.95). No interaction effect on energy intake per test meal product was found.
Analysis of the weight of the ad libitum meal gave similar results as the analysis of energy intake. A significant condition effect was found (F4,100 = 6.02; P < 0.001). Post hoc testing showed significant differences between the control condition and the longer oral stimulation conditions (8 min/100 ml vs. control T = −4.43, P <0.001; and 8 min/800 ml vs. control T = −3.89, P = 0.002). No interaction effects were found with the two by two ANOVA.
All appetite ratings were similar across conditions at baseline (t = 0; all P values ≥0.44) indicating that the subjects followed the instructions. Similarly no difference was found after the test meal (t = 60; all P values ≥0.22), indicating that the subjects ate to satisfaction. The ratings for fullness and hunger are presented in Figure 3. The ratings for prospective consumption, desire to eat, desire for something sweet and desire for something savory gave similar figures as the hunger ratings. As expected, the fullness rating showed a mirror image of the hunger rating, with a maximum for fullness directly after the treatment (t = 8) and a minimum right before the ad libitum test meal (t = 30).
Both hunger and fullness ratings differed across conditions (for both: F ≥ 3.17, P ≤ 0.01). Post hoc testing showed that hunger was significantly higher in the control condition than in the other four conditions (Δ5–8 mm, for all comparisons: T ≥ 2.85, P < 0.05). The fullness rating was on average 7 mm lower in the control condition than in the 8 min/800 ml condition (T = −3.08, P = 0.02).
No interaction effects of oral and gastric stimulation were found for hunger or fullness ratings.
Nausea and wellbeing
Wellbeing and nausea ratings were similar across conditions at baseline (t = 0) and after the test meal (t = 60) (all P values ≥0.44). A significant condition effect was found for nausea (F4,600 = 4.64; P = 0.001). The mean nausea rating of the 1 min/800 ml condition was significantly higher than that in the control condition and the other three conditions (Figure 4, mean Δ4–6 mm, for all ratings T ≤ 2.92, P ≤ 0.04).
In the present study, we investigated the contribution of oral and gastric stimulation to appetite and energy intake by simultaneously exposing subjects to independent oral and gastric stimulation. We hypothesized that oral and gastric stimulation would equally and additively contribute to a reduction in appetite and energy intake. Our main finding was that, compared with the control condition, ad libitum energy intake at 30 min was significantly decreased in the 8-min oral stimulation conditions but not in the 1-min conditions. In contrast, the 800-ml gastric volume conditions did not result in a stronger suppression of ad libitum energy intake than the 100-ml conditions. Thus, in our study oro-sensory stimulation was at least as effective in suppressing energy intake as gastric volume.
As far as we know this study is the first in which oral and gastric stimulation were simultaneously manipulated by using MSF and intragastric infusion. Our finding that oro-sensory stimulation was effective in suppressing ad libitum energy intake is in line with studies of Zijlstra et al. The authors showed that the low-satiating efficiency of liquids is due to their high-eating rate, which is associated with a lower degree of oro-sensory stimulation. Zijlstra et al. found that subjects consumed significantly more of a test product when oral processing time was 3 s rather than 9 s (7). This was also observed in a study of Bolhuis et al. where short oro-sensory exposure resulted in 34% higher ad libitum soup intake than long oro-sensory exposure (32). In another study of Zijlstra et al. they found a 30% higher energy intake of a liquid product compared with a semisolid product, while standardizing the eating rate reduced this difference to 12% (8). In a more recent study, Zijlstra et al. showed that ad libitum intake (g) in the highest quartile of eating rate was almost twice as high as in the lowest quartile of eating rate (10).
The present results suggest that there is no optimal satiety without adequate sensory stimulation. These findings are in agreement with the studies of Cecil et al. (25,26), who found that the oral consumption of soup resulted in lower appetite ratings than administration of the same soup directly into the stomach or duodenum. It may be that humans need enough sensory stimulation in congruency with appropriate gastric stimulation for an optimal regulation of energy intake (3). The physiological mechanisms that contribute to the satiating effect of longer oro-sensory stimulation are probably related to the so-called cephalic phase responses (33,34). Cephalic phase responses are physiological responses to sensory signals, conceived to prepare the gastrointestinal tract for the inflow of nutrients, with the aim to maintain homeostasis. In various studies it has been shown that longer MSF leads to higher satiety ratings (15,16,29), whereas cephalic phase responses to (clear) liquids appear to be absent (35,36). This supports the theory that the taste system is a nutrient sensing system that informs the brain and the gastrointestinal tract on the amount and nature of the nutrients in the body (3). A relatively short sensory exposure may bypass this system (37) and explain the positive association which is found in some studies between sugar sweetened beverage consumption and energy intake and body weight (4,38,39,40).
We found that all conditions with oral and gastric stimulation decreased perceived appetite significantly compared with the control condition (Figure 3). Moreover, there seems to be a short lived effect of gastric filling volume on appetite shown by the first measurement after the treatment (t = 8). At that time point, the 800 ml conditions (1 min/800 ml and 8 min/800 ml) resulted in the highest ratings of fullness, and the lowest ratings of hunger (Figure 3). However, these effects disappeared at t = 30. This could explain why the effects on energy intake are not reflected in the appetite ratings. It appears that the gastric filling volume elicited an effect which too short lived to affect ad libitum food intake 30 min later. This is in contrast to other studies which found that volume and weight of the gastric load had more influence on short-term food intake than energy content (11,12). The low caloric content of the gastric filling, i.e., 99 kcal, combined with its relatively low viscosity may have caused a quick passage of the stomach. Furthermore the stomach might sense very quickly (within minutes) that there is little energy in the gastric load (41). Different effects may be found when more energy is included in the gastric load and/or when the load is more viscous. Another factor that might have played a role is that subjects were not aware of the nature and the amount of the gastric filling. Cecil et al. found that overt infusion suppressed appetite more than the same infusion administered covertly, both without oral stimulation (26).
It was found that MSF was performed properly across all conditions with oral and gastric stimulation; i.e., on average 98% of the dry mass of the cake was in the expectorants. For comparison, in a study of Smeets et al. the mean dry mass recovery percentage was 96% (15). At the moment of the ad libitum test-meal, nausea and wellbeing levels were similar across conditions, which indicates that discomfort was not playing a role in the present results.
Nausea ratings at t = 8 only differed significantly between the 1 min/800 ml condition and the other conditions. This effect may be due to the incongruence between the short oro-sensory exposure time (1 min) and the relatively large volume that was administered intragastrically over a period of 8 min. The short duration of this effect may be caused by a high gastric emptying rate as suggested by the quick disappearance of satiety. We would need gastric emptying rate measures to confirm this idea.
The present findings imply that functional foods for satiety in the form of drinks might not be an optimal solution. Such products lack the necessary sensory contribution to satiety. This does not necessarily mean that specially designed liquid products cannot have a good satiating efficiency. Their metabolic effects may contribute to satiety, but products with more oro-sensory stimulation will have a higher satiating capacity.
Overall, we conclude that the duration of oro-sensory stimulation was at least as important in reducing ad libitum energy intake as gastric filling volume. Furthermore, oral and gastric stimulation did not additively decrease energy intake. Based on our results and that of previous studies we conclude that a longer oral sensory stimulation might be an important factor in lowering energy intake.
We thank all participants, the nurses, the research assistants and dieticians (René Bekx, Pauline Claessen, and Corine Peerenboom), Els Siebelink for dietary advice, and Tineke van Roekel for the chemical analyses. The author's responsibilities were as follows: A.G.M.W. collected and analyzed the data, and wrote the manuscript; A.E., E.A., M.M., P.A.M.S., and C.dG. designed the study and provided significant advice and consultation. This research was funded by Nestlé Research Centre. At the time the research was carried out, A.E. and E.A. were employed by the funder.