Energy-yielding liquids elicit weak suppressive appetite responses and weak compensatory responses, suggesting that liquid calories might lead to a positive energy balance. However, data is often derived from foods differing in many characteristics other than viscosity.
To investigate the effect of viscosity on ad libitum food intake in real-life setting and to investigate whether a difference in ad libitum intake is related to eating rate and/or eating effort.
In real-life setting 108 nonrestrained subjects (26±7 years, BMI 22.7±2.4 kg m−2) received a chocolate flavored liquid, semi-liquid and semi-solid milk-based product, similar in palatability, macronutrient composition and energy density. In laboratory setting 49 nonrestrained subjects (24±6 years, BMI 22.2±2.3 kg m−2) received the liquid or semi-solid product. Effort and eating rate were controlled by means of a peristaltic pump.
In real-life setting the intake of the liquid (809±396 g) was respectively 14 and 30% higher compared to the semi-liquid (699±391 g) and semi-solid product (566±311 g; P<0.0001). In laboratory setting, removing eating effort, resulted in a 29% (P<0.0001) intake difference between liquid (319±176 g) and semi-solid (226±122 g). Standardizing eating rate resulted in 12% difference between liquid (200±106 g) and semi-solid (176±88 g; P=0.24). If not controlled, the difference in intake between liquid (419±216 g) and semi-solid (277±130 g) was comparable to the real-life setting (34%; P<0.0001).
Products different in viscosity but similar in palatability, macronutrient composition and energy density lead to significant differences in intake. This difference is partially explained by the higher eating rate of liquids.
Several studies have shown that energy-yielding liquids seem to elicit weaker suppressive appetitive responses compared to solids. Postprandial hunger scores are significantly lower and/or satiety scores significantly higher after ingestion of more viscous/solid foods compared to liquid foods.1, 2, 3, 4 Furthermore, compared to solids, liquids seem to elicit a weak compensatory response in balancing energy intake throughout the day.5, 6 This effect is not only apparent in short-term preloads studies but is also observed in longer term intervention studies. In a 4-week crossover study7 subjects consumed daily a liquid or a solid carbohydrate load for 4 weeks. Free energy intake during the period that the solid was consumed was significantly lower than prior to this period. No decrease in energy intake occurred during the period that the liquid load was consumed. Total 24 h energy intake even increased by an amount equal to the energy content of the liquid food load. Also other long-term studies have shown that energy ingested from energy yielding liquids add to the total energy intake during the day.8, 9, 10, 11, 12 These data suggest that liquid calories may lead to a positive energy balance and subsequent weight gain.
The underlying mechanism responsible for the difference in satiety responses between liquids and solids is still unclear. A possible mechanism that has been suggested in earlier studies is that the act of chewing the more solid foods may give a satiety signal which is not induced by swallowing a liquid.4, 5 This may be related to the notion that the faster transit of fluids compared to solids leads to smaller sensory exposure time in the mouth.
The objective of our study is first to investigate whether products differing in viscosity lead to differences in ad libitum intake and furthermore to investigate the hypothesis that the difference in ad libitum intake between products with different viscosities might be due to differences in oral exposure time. This implies that this effect operates through sensory mechanisms. A thicker product is probably eaten more slowly than a more liquid product. We assume that eating slower results in a longer stay of the product in the oral cavity and therefore in a longer oral exposure time.
We hypothesize that the effect of viscosity on food intake runs via sensory mechanisms. As explained in the satiety cascade of Blundell,13 sensory processes have an early onset during the course of eating and primarily affect satiation. Satiation is the process that eventually brings the period of eating to an end and thus is responsible for meal termination. The importance of sensory aspects in meal termination has also been shown in the study of Hetherington.14 Therefore in the design of our studies we chose to focus on satiation/meal termination, which we defined as ad libitum food intake.
To achieve our objective we performed two studies. The aim of study 1 was to investigate the effect of viscosity on ad libitum food intake in a real-life setting. In this study it was found that subjects consumed more from more liquid products. Two possible explanations were differences in eating rate and differences in the effort to get the product in the mouth. Therefore the aim of study 2 was to investigate whether the observed difference in ad libitum intake could be related to eating rate and/or eating effort. Since interpretation of studies investigating viscosity may be hampered by the use of foods that differ along other dimensions than viscosity, we used foods with different viscosities but with the same energy content, energy density, nutrient composition and similar palatability.
Subjects and methods
Subjects were recruited in the surroundings of Ede and Wageningen (the Netherlands) via flyers, advertisement in local papers and email. Subjects had to be healthy, aged 18–50 years, normal weight (BMI 18.5–30.0 kg m−2) and had to like chocolate flavored dairy products. Exclusion criteria were restrained eating (Dutch eating behavior questionnaire (DEBQ) men: score>2.37, women: score>3.2415), lack of appetite for any (unknown) reason, following an energy restricted diet during the last 2 months, change in body weight of >5 kg during the last 2 months, stomach or bowel diseases, diabetes, thyroid disease or any other endocrine disorder or hypersensitivity for milk components.
Subjects were not aware that the primary outcome of the studies was ad libitum food intake, as this could affect the outcome of the study. Subjects of study 1 were informed that the aim of the study was to test the palatability of chocolate flavored milk products while watching a movie. Subjects of study 2 were informed that the aim of the study was to investigate the effect of certain factors, such as eating rate and effort to consume a product, on taste experience. The studies were approved by the Medical Ethical Committee of the Wageningen University and all subjects gave their informed consent. We certify that all applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during this research.
A total of 132 subjects participated in study 1, of which 108 subjects were included in the data analysis (Table 1). Subjects were excluded from data analysis because of incomplete data with respect to ad libitum intake (n=7) and mistakes in weighing the ad libitum intake (n=13). Furthermore, 4 subjects were accidentally included who were restrained eaters (n=2) and did not fulfill the criteria of a normal BMI (n=2). Fifty subjects participated in study 2, of which 1 subject was excluded from data analysis because of not adhering to the study protocol. Thirty-four subjects participated in both studies.
The test products were milk-based products with chocolate flavor specially developed for this study (NIZO food research, Ede, the Netherlands). The basis of all chocolate products was whole fat milk (68%), water (18%), sugar (6.5%), modified starch (3.5%), cream (2%), cacao (1.5%) and carrageenan (0.05%). The type starch was varied in the products to obtain three identical products differing solely in physical state, a liquid (comparable to commercially available chocolate milk), a semi-liquid and a semi-solid product (comparable to commercially available chocolate custard).
The products were made in the following way: After analysis of the protein, fat and lactose content, the fresh milk was standardized at 2.2% protein and 3.7% fat, by adding cream. To the standardized milk a dry mix of sugar, starch mixture (Perfectamyl A3108 and/or Farinex VA85T and/or Farinex VA40), cacao and carrageenan was added. After mixing of the ingredients, the product was pasteurized (30 s, 85 °C) and cooled down (<7 °C). Pasteurization was performed to prolong shelf life, however products still had to be produced in several batches to maintain palatability and microbiological safety throughout the studies. Products are commercially available at NIZO food research. Table 2 gives an overview of the mean liking scores and the sensory profile of the test products.
Figure 1 shows the viscosity measurements of the experimental products. Measurements were made with a AR 2000 small strain rheometer (TA Instruments, Etten-Leur, the Netherlands) at 5 °C with shear rates increasing from 0 to 1000 per second in 5 min. Viscosity at a shear rate of 50 per second at 5 °C was on average 85 mPa s for the liquid product, 233 mPa s for the semi-liquid product and 788 mPa s for the semi-solid product. For comparison, a commercially available chocolate milk and chocolate custard were also measured at 42 per second and 57 per second; for the chocolate milk this was 60 and 56 mPa s, and for the chocolate custard 2360 and 1870 mPa s.
Test products were equal in energy content, energy density, volume and macronutrient composition. Macronutrient content was determined by chemical analysis; protein by the Kjeldahl method, fat by the Rose Gottlieb method, subsequently carbohydrate was calculated by subtracting moisture, ash, protein and fat from total weight. Energy content was calculated from the macronutrient composition by means of the Atwater factors. For both studies, products were produced in several batches. For the chemical analyses, food samples were taken from a homogenous mixture of all batches within a study. Data from the different studies were averaged and are shown in Table 3. In study 1, the liquid, semi-liquid and semi-solid products were used and in study 2, the liquid and semi-solid products.
The studies were randomized crossover experiments. The order of test products and conditions were randomized within subjects. The time between sessions was at least one day. Ad libitum intake was the primary outcome of both studies and was calculated from the weight of the products before and after consumption. Products were weighed on a digital scale with a precision of 0.1 g (model XP-3000, Denver Instruments, Göttingen, Germany and model 1203-MP, Sartorius, Gottingen, Germany). Subjects were not aware of the weighing. Products were served chilled (temperature varying from 5–13 °C in study 1 and 3–5 °C in study 2).
Study 1: Ad libitum intake in a real-life setting
To create a real-life setting and to distract subjects from visual and weight cues, study 1 was performed in a cinema. Each subject participated in three sessions, which were chosen out of total of six available test days. Each test session started at 1800 hours. On each test day a different movie was shown, however the type of movie was similar (romantic comedy). During the movie, subjects were instructed to eat as little or as much of the test product as they wanted. In order to keep any visual or weight cues from the subjects and to stimulate ad libitum intake, a surplus of the test product was served in white carton ‘bag-in-boxes’ (max content of 2 l, height 23 cm, width 12 cm, depth 8.5 cm) and consumed with a thick straw (length 26 cm, diameter 0.9 cm). The total duration of each movie was set at 90 min. The movie was divided into three parts of 30 min with breaks of 15 min in between. Table 4 gives an overview of the time schedule of a test session. During the break, subjects left the theater and handed in the box with test product. Before the restart of the film, they received a new box with test product. On average, three times 1501±15 g of product was offered (total session 4504±34 g). At the beginning and end of the movie, satiety parameters and sensory attributes were rated.
Study 2: Ad libitum intake during controlled eating rate and/or effort
In study 2, subjects returned for six sessions. All test session were held during dinnertime. Subjects could choose out of three starting times (16.15, 17.15 or 18.15 h), which had to be the same for all six sessions. After arrival (t=0 min) subjects received a preload (see ‘Standardization of satiety state before ad libitum intake’). After 45 min (t=45 min) subjects received the test product.
This study took place in the sensory cabins of the Department of Human Nutrition, Wageningen University. There were three experimental conditions: (1) free eating rate, different effort; (2) free eating rate, no effort and (3) fixed eating rate, no effort. During all conditions subjects were instructed ‘to eat from the test product until pleasantly satiated’. It was not required to remain in the laboratory for a set period of time.
During condition 1 (control), subjects consumed from the blinded box by means of a thick straw, identical to the manner of consumption in study 1. In this condition, on average 1498±23 g of product was offered per session. During condition 2, an electric peristaltic pump (Watson-Marlow, type 323Dz and 505s, Watson-Marlow pumps Bredel, USA/Canada) with a silicon tube (length 2 m, diameter 4.8 mm; Rubber BV, Hilversum, the Netherlands) was used to deliver the product in the subject's mouth, so that no effort was required to consume the product. Subjects still needed to orally process the product and swallow. Subjects could control the pump themselves (minimum speed that was used: 12 g min−1, maximum speed: 224 g min−1). During condition 3, the peristaltic pump was set at a standardized rate to control eating rate. The pump was adjusted to pump for 10 s, hereafter to stop for 10 s and to repeat this cycle, to allow subjects to swallow and to imitate a natural drinking situation. The pump was set at a fixed rate of 100 g min−1 for men and 80 g min−1 for women. Since the pump was set to pump the product every other 10 s the total intake of the subjects was fixed to 50 g min−1 for men and 40 g min−1 for women. Flow rates were higher for men since they have a higher energy need compared to women. During pre-testing, the chosen rates were experienced as convenient. In this condition the pump was not visible to the subject.
In all conditions total time of consumption was measured. In condition 1, ad libitum intake was calculated by weighing the boxes before and after consumption and in conditions 2 and 3, ad libitum intake in grams was registered every 30 s. This was performed by placing the box with test product on the digital balance and recording the number on the display every 30 s. Subjects were not aware of the weighing.
At the beginning and end of the test session subjects rated satiety parameters, and before ad libitum intake sensory attributes were rated.
Standardization of satiety state before ad libitum intake
To standardize the individual state of satiety, subjects in study 1 and 2 were instructed to eat the same breakfast and lunch during all test days and record their food and drink consumption in a diary. Furthermore, the individual state of satiety was standardized by means of a preload, which was a piece of commercially available mini pizza (1130 kJ per mini pizza, brand Chaupain, Kreko, Dordrecht, the Netherlands). The amount of pizza was based on individual energy needs, estimated by means of the Schofield I equation,16 taking into account age, weight and a physical activity level (PAL) of 1.6. About one-sixth of energy of the daily estimated energy needs was provided by the preload, this is about half the amount of energy provided by the evening meal in the Netherlands.17 For logistic reasons three energy groups were formed; group 1 (estimated energy needs ⩽8.5 MJ): 1 mini pizza, group 2 (estimated energy needs between 8.5 and 11.9 MJ): 1.5 mini pizzas, and group 3 (estimated energy needs ⩾11.9 MJ): 2 mini pizzas. In study 1, 7 subjects received 1 mini pizza, 78 received 1.5 mini pizzas and 23 received 2 mini pizzas. In study 2, 4 subjects received 1 mini pizza, 37 received 1.5 mini pizzas and 8 received 2 mini pizzas.
During the test session, subjects were not allowed to eat or drink anything other than the preload and the test products.
In both studies, before and after ad libitum intake, subjects rated hunger, fullness, desire to eat, appetite for something sweet, appetite for something savory, prospective consumption and thirst on 10-point category scales. Scales were anchored from not at all/very little until very much. Changes in scores were calculated for the satiety parameters by subtracting the ratings before ad libitum consumption from after ad libitum consumption; thus negative scores indicate a decrease in rating.
Data are presented as means±s.d. Statistical analyses were performed by means of SAS (version 9.1.3; SAS Institute Inc., Cary, NC, USA). Significance was set at P<0.05.
In study 1, analysis on ad libitum intake and on the satiety parameters were conducted by one-way ANOVA (Proc GLM, SAS) with participant and product as independent variables. If type of product had a significant effect, least squares means were used for post hoc comparisons. The liking scores before ad libitum intake were significantly different between products (see Table 3), therefore the analysis on the ad libitum intake was also performed with a correction for liking by adding the liking scores as independent variable to the model. To test whether the difference in intake between products was dependent on BMI, linear regression analysis was performed.
In study 2, the ad libitum intake data, the duration time of ad libitum eating, the rate of consumption and the satiety parameters were analyzed by two-way ANOVA (proc GLM, SAS), with participant, product and condition as independent variables including condition × product interaction.
The analysis on ad libitum intake was also performed with correction for liking by adding the liking scores as independent variable. Least squares means were used for post hoc comparisons.
To test whether the difference between the liquid and the semi-solid was altered as an effect of condition, the difference between the intake of the liquid and the semi-solid was calculated for each condition (intake liquid minus intake semi-solid). These intake differences were compared among the three conditions by one-way ANOVA (proc GLM, SAS) with condition and participant as independent variable. If condition had a significant effect, least squares means were used for post hoc comparisons. To test whether the difference in intake between the liquid and the semi-solid was dependent on BMI, linear regression analysis was performed.
Furthermore, linear regression analysis was performed within the ‘free eating, no effort’ condition on the ad libitum intake data versus consumption rate.
Ad libitum intake test products in real-life setting
The intake of the three test products was significantly different from each other (P<0.0001; Figure 2). Ad libitum intake was 809±396 g for the liquid, 699±391 g for the semi-liquid and 566±311 g for the semi-solid products. The difference between the liquid and the semi-solid products was 243±304 g (30%; P<0.0001). The difference in intake between the liquid and the semi-liquid products was 110±299 g (14%; P<0.0001). The difference between the semi-liquid and the semi-solid products was 133±233 g (19%; P<0.0001).
The intake of the products in the first half-hour of the test session were liquid 443±267 g, semi-liquid 384±234 g and semi-solid 316±209 g. In the second half-hour intakes were liquid 187±133 g, semi-liquid 161±130 g and semi-solid 124±88 g. In the last half-hour of the test session, mean intakes were 179±115 g for the liquid, 155±126 g for the semi-liquid and 126±97 g for the semi-solid.
Linear regression analysis showed that the differences in ad libitum intake were not statistically significantly dependent on BMI.
The correction for liking did not influence the mean ad libitum intake data or the P-values. Mean ad libitum intakes after correction were 802 g for the liquid, 699 g for the semi-liquid and 576 g for the semi-solid.
Ad libitum intake test products in laboratory setting
Subjects consumed significantly more of the liquid product than of the semi-solid product (Figure 3) in the ‘free eating rate, different effort’ and in the ‘free eating rate, no effort’ conditions. However, the ad libitum intake of the liquid and the semi-solid product was not statistically significantly different in the condition where eating rate was standardized. These effects are reflected in a significant condition × product interaction (F=9.24, P=0.0001).
In the ‘free eating rate, different effort condition’ (condition 1), the absolute intake of the liquid and the semi-solid was the highest (419±216 and 277±130 g, respectively) and the difference in intake between the liquid and the semi-solid was the largest, 143±200 g (34%; P<0.0001).
In the ‘free eating rate, no effort condition’ (condition 2) the absolute intake of the liquid and the semi-solid were smaller (319±176 and 226±122 g, respectively) and the difference between the products was also smaller, 93±137 g (29%; P<0.0001). However, this difference of 93 g was not significantly different from the difference of 143 g in the ‘free eating rate, different effort’. So, controlling for effort had no statistically significant effect on the ad libitum intakes.
In the ‘fixed eating rate, no effort condition’ (condition 3) the absolute intake of the liquid and the semi-solid were the smallest and not significantly different from each other (200±106 and 176±88 g, respectively; P=0.24). The difference between the intake of the liquid and the semi-solid was also the smallest, 23±78 g (12%). This difference was significantly different from the difference in intake between the liquid and the semi-solid in conditions 1 and 2 (P<0.0001 and P=0.01, respectively). So controlling for effort and eating rate has a significant effect on ad libitum intakes.
Linear regression analysis showed that the differences in ad libitum intake between the liquid and the semi-solid product were not statistically significantly dependent on BMI.
After correction for liking, the means of the ad libitum intake were not changed and significant results remained the same. Mean ad libitum intakes after correction were 419 g for the liquid and 278 g for the semi-solid in the ‘free eating rate, different effort’ condition, 316 g for the liquid and 229 g for the semi-solid in the ‘free eating rate, no effort’ condition and 197 g for the liquid and 177 g for the semi-solid in the ‘fixed eating rate, no effort’ condition.
Eating time and eating rate in laboratory setting
In study 2, the total time the subjects took for ad libitum intake (without the time of the sensory rating of the product) was not significantly different between the liquid and semi-solid in the three testing conditions. The overall median time was 2.6 min for the liquid product and 2.8 min for the semi-solid product.
Figure 4 shows the cumulative consumption over time for test conditions 2 and 3. This figure stops at 4 min, which is the 75 percentile value for both the liquid and semi-solid products. In condition 1, the ‘free eating rate, different effort’ condition cumulative consumption could not be measured. In the ‘fixed eating rate, no effort’ condition the mean consumption rate overall was 43.7±5.8 g min−1 for the liquid product and 41.5±6.5 g min−1 for the semi-solid product. The mean consumption rates in the ‘free eating rate, no effort’ condition were significantly higher, 89.5±50.1 g min−1 for the liquid product and 56.7±20.2 g min−1 for the semi-solid product (P<0.001). Furthermore, regression analysis within the ‘free eating rate, no effort’ condition showed that there is a significant positive relation between ad libitum intake and consumption rate; for the liquid product R-square=0.43 and P<0.0001, and for the semi-solid product R-square=0.44 and P< 0.0001.
Table 5 gives an overview of the satiety ratings before and after ad libitum intake in the real-life setting (study 1). Before ad libitum intake there were no significant differences in ratings between the products. After ad libitum intake, change in thirst scores was significantly different between the products (P=0.007). The liquid product decreased thirst more than the other products. Similar ratings were obtained in the experimental study (study 2; data not shown).
The results of our studies show that food differing in viscosity lead to clear differences in ad libitum intake, both in the realistic setting as well as in the laboratory setting. As far as we know these are the first studies that demonstrated such a clear effect of viscosity on ad libitum intake/meal termination. This effect was not due to differences in energy content, energy density or macronutrient content as they were all identical. Thus the potential confounders in the interpretation of liquid–solid differences in satiety were controlled for. The effect was, furthermore, not due to the effort to get the product in the mouth. In the present studies the effect was explained by the faster eating rate of the liquid compared to the semi-solid product.
In our study, we found that the eating rate of the liquid product was significantly higher than the eating rate of the semi-solid product. Other studies which measured eating rate also showed that the consumption rate of liquids is much higher than that of solids, which is in agreement with our results. In the study of Haber et al.4 subjects had to consume a fixed amount of apples, apple purée or apple juice. The apple juice was eaten more than 10 times faster than the apple and nearly 3 times faster than the puree. Also in the study of Kissileff et al.,18 it was found that the liquefied version of the food was eaten faster (108 g min−1) than the solid version (71 g min−1) when both foods had to be eaten with a spoon.
The finding that the differences in intake between the liquid and the semi-solid are explained by eating rate, suggests that the transit time in the oral cavity plays an important role in this respect. With eating rate we mean the consumption volume per minute rather than the total time available for an eating episode. A liquid is eaten at a much higher rate and does not stay in the mouth for a long time, while a thick product is eaten more slowly and stays in the mouth much longer. This could be an important factor in the explanation of differences in satiety responses between liquids and solids. When a product stays in the mouth for a longer time, the exposure time to sensory receptors in the oral cavity is longer and there is more opportunity for more exposure to taste, smell, texture and so on.
In our second study in the laboratory setting we found that the ad libitum intake in the conditions with a pump was lower compared to the intake by means of a straw. Additionally, we found that the total eating time in this study was low. This could have been a result of the fact that subjects were not instructed to remain in the laboratory for a fixed period of time. Perhaps in future studies it would be better to set a fixed time of presence in the laboratory, to make sure subjects are eating normally. Furthermore in this second study we were able to remove the effort to get a product into the mouth by using a pump. We did not remove the effort to orally process the product. However, we assume that these factors did not have a large effect on our data since the products used in our studies required no chewing. Nevertheless, further studies in this area are needed. Eating rate and oral physiology are also subjects of our future studies.
Other factors that could have influenced the ad libitum intake in study 2 is that drinking by means of a peristaltic pump was unnatural or that the fixed eating rate of 50 g min−1 (men) or 40 g min−1 (women) was low and the cycle of the pump to stop every 10 s was unpleasant or became boring. However, the satiety ratings were not significantly different between the different test conditions. Apparently subjects did consume the products until satiated.
The finding that there were no differences in satiety parameters after ad libitum intake in our studies, despite the considerable differences in intake, is interesting. Apparently subjects did not feel less hungry or fuller after the larger consumption of the liquid product. This was also seen in the study of Wansink et al.19 in which subjects ate significantly more soup from self-refilling bowls than subjects eating from normal bowls. Although they consumed 73% more compared to the subjects eating from normal bowls, they did not feel more satiated. Some other studies investigating satiety responses to liquids and solids did find differences in satiety ratings. Mattes and Rothacker2 performed a study with a thin and thick version of a shake with equal volume, energy content and macronutrient composition. They found that hunger ratings were significantly lower after ingestion of the more viscous shake. Also Hulshof et al.1 found clear differences in appetite ratings between three preloads differing in physical states, with the solid preloads suppressing appetite ratings the most. In the study of Tsuchiya et al.,3 it was found that satiety ratings following a yogurt preload were significantly higher than following a preload of fruit drink or dairy fruit drink. No difference however was found between the two yogurts, which were a semi-solid yogurt and the same yogurt homogenized to liquid form. However, the yogurts contained more protein than the beverages and it has been shown that of the macronutrients, protein is more satiating.20, 21
All of the above mentioned studies that did find differences in satiety ratings were preload studies. In these types of studies, subjects consume a fixed amount of the several test products, usually equal in energy content. In our study subjects could eat ad libitum until satiated. Apparently energy intake and feelings of satiety do not correspond with each other between liquids and solids. When the same amounts of calories are consumed, the subjective feelings of satiety are different and when the subjective feelings of satiety are the same, the amount of calories consumed is different. That hunger ratings are not accurate predictors of energy intake has already been shown in earlier studies, for example, by Mattes.22 Perhaps the human appetite system is not well equipped to sense liquid calories. In nature, calories in liquid form do not occur, except for milk during infancy, which is a period of rapid growth. During this period there is a largely direct relationship between food viscosity and caloric content since the viscosity and the caloric density of human breast milk appear to vary together.23, 24 Breast feeding may provide an important initial exposure to a general rule that thicker substances contain more calories than thinner substances.24, 25 This could mean that the difference in satiety responses between liquids and solids is based on learned behavior. Furthermore that solely the mouth feel of a product already could have an effect on the relationship between viscosity and intake.
So far, based on the studies described in this article, we conclude that food intake increases with decreasing viscosity and that the mechanisms through which viscosity affects food intake work (partly) through a higher/longer sensory exposure time and/or a longer transit time of the product in the oral cavity.
Hulshof T, De Graaf C, Weststrate JA . The effects of preloads varying in physical state and fat content on satiety and energy intake. Appetite 1993; 21: 273–286.
Mattes RD, Rothacker D . Beverage viscosity is inversely related to postprandial hunger in humans. Physiol Behav 2001; 74: 551–557.
Tsuchiya A, Almiron-Roig E, Lluch A, Guyonnet D, Drewnowski A . Higher satiety ratings following yogurt consumption relative to fruit drink or dairy fruit drink. J Am Diet Assoc 2006; 106: 550–557.
Haber GB, Heaton KW, Murphy D, Burroughs LF . Depletion and disruption of dietary fibre. Effects on satiety, plasma-glucose, and serum-insulin. Lancet 1977; 2: 679–682.
Mattes RD . Dietary compensation by humans for supplemental energy provided as ethanol or carbohydrate in fluids. Physiol Behav 1996; 59: 179–187.
Tournier A, Louis-Sylvestre J . Effect of the physical state of a food on subsequent intake in human subjects. Appetite 1991; 16: 17–24.
DiMeglio DP, Mattes RD . Liquid versus solid carbohydrate: effects on food intake and body weight. Int J Obes Relat Metab Disord 2000; 24: 794–800.
De Castro JM . The effects of the spontaneous ingestion of particular foods or beverages on the meal pattern and overall nutrient intake of humans. Physiol Behav 1993; 53: 1133–1144.
Tordoff MG, Alleva AM . Effect of drinking soda sweetened with aspartame or high-fructose corn syrup on food intake and body weight. Am J Clin Nutr 1990; 51: 963–969.
Malik VS, Schulze MB, Hu FB . Intake of sugar-sweetened beverages and weight gain: a systematic review. Am J Clin Nutr 2006; 84: 274–288.
Schulze MB, Manson JE, Ludwig DS, Colditz GA, Stampfer MJ, Willett WC et al. Sugar-sweetened beverages, weight gain, and incidence of type 2 diabetes in young and middle-aged women. JAMA 2004; 292: 927–934.
Raben A, Vasilaras TH, Moller AC, Astrup A . Sucrose compared with artificial sweeteners: different effects on ad libitum food intake and body weight after 10 wk of supplementation in overweight subjects. Am J Clin Nutr 2002; 76: 721–729.
Blundell JE, Green S, Burley V . Carbohydrates and human appetite. Am J Clin Nutr 1994; 59: 728S–734S.
Hetherington MM . Sensory-specific satiety and its importance in meal termination. Neurosci Biobehav Rev 1996; 20: 113–117.
Van Strien T . Eating behaviour personality traits and body mass. PhD thesis, Wageningen, Landbouwhogeschool, 1986.
WHO. Principles for the estimation of energy requirements. Energy and Protein Requirements. Report of a Joint FAO/WHO/UNU Expert Consultation, 2nd edn. World Health Organization: Geneva, 1985, pp 34–52.
Hulshof K, Ocké M, Van Rossum C, Buurma-Rethans E, Brants H, Drijvers J et al. Resultaten van de Voedselconsumptionpeiling 2003. Report 350030002/2004 RIVM, National Institute for Public Health and the environment, 2003.
Kissileff HR, Klingsberg G, Van Itallie TB . Universal eating monitor for continuous recording of solid or liquid consumption in man. Am J Physiol 1980; 238: R14–R22.
Wansink B, Painter JE, North J . Bottomless bowls: why visual cues of portion size may influence intake. Obes Res 2005; 13: 93–100.
Anderson GH, Moore SE . Dietary proteins in the regulation of food intake and body weight in humans. J Nutr 2004; 134: 974S–979S.
Halton TL, Hu FB . The effects of high protein diets on thermogenesis, satiety and weight loss: a critical review. J Am Coll Nutr 2004; 23: 373–385.
Mattes R . Hunger ratings are not a valid proxy measure of reported food intake in humans. Appetite 1990; 15: 103–113.
Picciano MF . Human milk: nutritional aspects of a dynamic food. Biol Neonate 1998; 74: 84–93.
McDaniel MR, Barker E, Lederer CL . Sensory characterization of human milk. J Dairy Sci 1989; 72: 1149–1158.
Davidson TL, Swithers SE . A Pavlovian approach to the problem of obesity. Int J Obes Relat Metab Disord 2004; 28: 933–935.
We thank all participants in these studies and all practical assistants, especially Themistoklis Altinzoglou, Els Siebelink and Miranda Janssen; CineMec cinema in Ede for their hospitality and especially Vico Duisings for his assistance during the real-life setting experiment; NIZO food research, Ede, the Netherlands especially Marja Kanning; Truus Kosmeyer for the chemical analyses and Harry Baptist for the rheological measurements. The studies were funded by Top Institute Food and Nutrition. TI Food and Nutrition, formerly known as WCFS, is a unique public/private partnership that generates vision on scientific breakthroughs in food and nutrition, resulting in the development of innovative products and technologies that respond to consumer demands for safe, tasty and healthy foods. Partners are major Dutch food companies and research organizations.
About this article
Cite this article
Zijlstra, N., Mars, M., de Wijk, R. et al. The effect of viscosity on ad libitum food intake. Int J Obes 32, 676–683 (2008). https://doi.org/10.1038/sj.ijo.0803776
- food intake regulation
- sensory satiety
- eating rate
American Journal of Physiology-Regulatory, Integrative and Comparative Physiology (2020)
Characterization of individual particle movement during in vitro gastric digestion in the Human Gastric Simulator (HGS)
Journal of Food Engineering (2020)
The effects of psyllium supplementation on body weight, body mass index and waist circumference in adults: A systematic review and dose-response meta-analysis of randomized controlled trials
Critical Reviews in Food Science and Nutrition (2020)
Sweet satiation: Acute effects of consumption of sweet drinks on appetite for and intake of sweet and non-sweet foods