- Email: [email protected]
Repeated consumption of a large volume of liquid and semi-solid foods increases ad libitum intake, but does not change expected satiety
Appetite 59 (2012) 419–424
Contents lists available at SciVerse ScienceDirect
Appetite journal homepage: www.elsevier.com/locate/appet
Research report
Repeated consumption of a large volume of liquid and semi-solid foods increases ad libitum intake, but does not change expected satiety q P.S. Hogenkamp a,b, M. Mars a,b, A. Stafleu a,c, C. de Graaf a,b,⇑ a
Top Institute Food and Nutrition, P.O. Box 557, NL-6700 AN, Wageningen, The Netherlands Division of Human Nutrition, Wageningen University, P.O. Box 8129, NL-6700 EV Wageningen, The Netherlands c TNO, P.O. Box 360, NL-3700 AJ, Zeist, The Netherlands b
a r t i c l e
i n f o
Article history: Received 25 November 2011 Received in revised form 23 May 2012 Accepted 11 June 2012 Available online 18 June 2012 Keywords: Food intake Texture Satiation Satiety Energy density Expectations Dietary learning
a b s t r a c t Food intake and a food’s expected satiating effect initially rely on sensory attributes. People will learn about the food’s satiating capacity by exposure. We investigated whether repeated consumption changed the expected satiety effects and intake of iso-energetic liquid and semi-solid foods. In a randomised cross-over study, participants (n = 53; age: 21 ± 2.9 y; BMI: 21.8 ± 2.0 kg/m2) consumed one of two isoenergetic dairy foods (liquid or semi-solid) for breakfast in each 5-day test condition. Expectations regarding satiety were measured on days 1, 2, and 5. Foods were offered ad libitum on days 1 and 5 and in a fixed volume on days 2–4. Appetite sensations were rated up to 180 min after the start of the session on fixed time points. Expected satiety effects of the semi-solid food were higher than of the liquid food on all days (p < 0.0001). Ad libitum intake of the liquid food was higher than of the semi-solid food on day 1 (liquid: 391 ± 177 g, semi-solid: 277 ± 98 g; p < 0.0001) and day 5 (liquid: 477 ± 161 g, semi-solid: 375 ± 148 g; p < 0.0001). On day 2, hunger was rated lower and fullness rated higher after the semi-solid compared with the liquid food; on day 4, no differences were observed (significant product⁄ exposure interaction AUC). Changes in hunger and fullness indicated that the fixed volumes of liquid and solid food were perceived to be equally satiating after repeated consumption, but this did not result in the anticipated changes: expected satiety effects remained lower, and ad libitum intake higher for the liquid compared with the semi-solid food. The effect of texture on a food’s expected satiety effects and its ad libitum intake appears to be large, also after repeated consumption. Expectations based on sensory cues are not easily changed. Ó 2012 Elsevier Ltd. All rights reserved.
Introduction The intake of a particular food is partly based on previous experience with the food. People may learn about the energy content of a food by exposure and link post-ingestive appetite sensations to the food’s sensory attributes (Yeomans, Weinberg, & James, 2005; Zandstra, Stubenitsky, De Graaf, & Mela, 2002). More precise, meal size will be largely controlled by a learned decrease in appetite towards the end of a meal, which has been defined as learned satiety (Booth & Davis, 1973; Booth & Grinker, 1993; Brunstrom, 2007). Learned q Acknowledgements: We kindly thank Jeffrey Brunstrom for sharing with us his computer task, Natasja Hück for research assistance, and Hans Meijer for technical support. The studies were supported by the Top Institute Food and Nutrition. The authors’ responsibilities were as follows—PSH: designed the study, collected and analysed the data, and wrote the manuscript; MM, AS, and CdG designed the study and provided significant advice and consultation. None of the authors had a personal or financial conflict of interest. ⇑ Corresponding author. E-mail address: [email protected] (C. de Graaf).
0195-6663/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.appet.2012.06.008
satiety will result in expectations about the food’s satiating capacity. These expected effects of a food on satiety may play a role in decisions on portion size (Brunstrom & Rogers, 2009), and consequently in energy intake. Learning about the energy content of a food may change these expected effects on satiety (Wilkinson & Brunstrom, 2009), with changes in portion size as a possible consequence. Expectations about the satiety effects of a specific food are among others based on its sensory attributes (Sorensen, Moller, Flint, Martens, & Raben, 2003). One of the sensory attributes that is important in the food’s expected satiety is texture. In a preceding experiment, we observed that an increase in perceived thickness of several dairy products resulted in a consistent increase in the expected satiating capacity of the foods (Hogenkamp, Stafleu, Mars, Brunstrom, & de Graaf, 2011). This is in line with the results of several studies that showed that solid foods produce higher satiety ratings than iso-energetic liquid foods (e.g. (Hulshof, De Graaf, & Weststrate, 1993; Mattes & Rothacker, 2001)), although the effect of texture on satiety and subsequent intake may critically depend on the volume of the liquid and solid preloads and on the time
420
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
Table 1 Overview of the study design. In this cross-over design, participants were served dairy foods for breakfast on five consecutive days in each condition.a Conditiona Liquid Semi-solid
Food intake Measurementsb:
Day 1
Day 2
Day 3
Day 4
Day 5
Ad libitum Expected satiety Sensory attributes Appetite sensations
Fixed Expected satiety Sensory attributes Appetite sensations
Fixed – Sensory attributes Appetite sensations
Fixed – Sensory attributes Appetite sensations
Ad libitum Expected satiety Sensory attributes Appetite sensations
a
The order of conditions was randomised and balanced within subjects. Conditions were separated by a period of at least one week. Expected satiety was measured using the ‘method of adjustment’ as described in the methods. Sensory attributes included pleasantness and familiarity, all rated on a visual analogue scale. Appetite sensations were rated on a visual analogue scale up to 180 min after the start of the session. b
interval to the next meal (Almiron-Roig, Chen, & Drewnowski, 2003; Hogenkamp, Stafleu, Mars, & de Graaf, 2011). In line with the (expected) satiety effects of liquid and solid foods, previous studies consistently showed that ad libitum intake of solid or semi-solid foods was lower than of liquid foods with similar macronutrient composition (Hogenkamp, Mars, Stafleu, & de Graaf, 2010; Mars, Hogenkamp, Gosses, Stafleu, & De Graaf, 2009; Zijlstra, Mars, de Wijk, Westerterp-Plantenga, & de Graaf, 2008). This suggests that learned satiety and expectations are indeed associated with satiation (portion size). Expected satiety of a specific food also depends on the energy content of the selected food item. Previous studies, however, showed that expectations did not change when repeatedly consuming low or high-energy-dense foods that were similar in appearance (Hogenkamp, Brunstrom, Stafleu, Mars, & de Graaf, 2011; O’Sullivan, Alexander, Ferriday, & Brunstrom, 2010). When satiety expectations based on the sensory attributes of a food are not congruent with the food’s energy content and its actual satiety sensations, this may disrupt energy intake regulation (Tepper, Mattes, & Farkas, 1991; Warwick & Schiffman, 1991). On the other hand, a large incongruence (between expected and experienced satiety) may promote changes in a food’s expected satiety effect, as is observed for hedonic expectations. Studies investigating hedonic expectations show that evaluation after consumption is generally adjusted towards expectations (Cardello & Sawyer, 1992; Wansink, Ittersum, & Painter, 2004). Only when there is a large contrast between expectations and evaluation (Yeomans, Chambers, Blumenthal, & Blake, 2008), e.g. when foods taste much better than expected (Wansink et al., 2004), future expectations will be adjusted. We hypothesised that a large contrast between expectations about a food’s effect on satiety and the experienced satiety value would result in adjustments in these expectations, and in changes in intake accordingly. Evaluation of the satiety process may, however, be more subtle than the (explicit) hedonic evaluation, and therefore requires repeated exposures to the food. In the current experiment, we investigated changes in expected satiety after repeated consumption of iso-energetic foods with different sensory attributes, and with different expected satiety at baseline (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011): the expected satiety of the test foods was either low (a liquid dairy food) or high (a semi-solid dairy food). We hypothesised that changes would be most pronounced in the condition with the liquid food, which was assumed to provide a higher perceived satiety than expected. Methods Design In a randomised cross-over design with two test conditions, participants were served dairy foods for breakfast on five consecutive days (Monday to Friday), see Table 1. The order of conditions was randomised and balanced within subjects. Conditions were separated by a period of at least one week. In each test condition,
participants consumed one of the two iso-energetic test foods: food with a low expected satiety at baseline (liquid ‘custard’) or food with a high expected satiety at baseline (semi-solid ‘pudding’). On days 1 and 5, test foods were offered ad libitum and intake was measured; on days 2, 3, and 4, a fixed volume of test food was consumed entirely. Expected satiety of the products was measured on days 1, 2 and 5. Preceding the study, the participants were trained in the procedures. Participants Healthy, normal weight participants, aged 18–40 y, were recruited from Wageningen and the surroundings. Exclusion criteria were: restrained eating [Dutch eating behaviour questionnaire: men score >2.25; women score >2.80 (26)], lack of appetite, following an energy-restricted diet or change in body weight >5 kg during the last 2 months, stomach or bowel diseases, diabetes, thyroid disease or any other endocrine disorder, hypersensitivity for food products under study, smoking, and being a vegetarian. A total of 50 participants would provide 80% power to detect a 20% difference in expectations regarding the satiating effect of the liquid food. In total, 53 participants (M/F: 12/41, age: 21 ± 2.9 y, BMI: 21.8 ± 2.0 kg/m2) completed the study. Participants were unaware of the exact aim of the study. In a screening session before the start of the study, they were informed that we were interested in taste sensations and acceptance of dairy products for breakfast. During this screening, body weight and height was measured and all participants signed an informed consent. Participants were debriefed after the study and received financial compensation. The present study was conducted according to the ethical standards of the
Table 2 Composition and characteristics of the liquid and semi-solid test foods: energy, macronutrientsa and energy percentage; ratingsb (mean ± SD) on pleasantness, familiarity and sensory attributes on day 1 (N = 53, M/F = 12/41).
Energy (kcal/kJ) Carbohydrates, g (% energy) Protein, g (% energy) Fat, g (% energy) Pleasantness Familiarityc Thickc Sweetc Sourc Intensity of tastec
Liquid
Semi-solid
96/402 14.0 (59) 3.1 (13) 3.0 (28) 58 ± 22 36 ± 24 15 ± 14 71 ± 19 40 ± 25 66 ± 14
98/410 14.5 (60) 3.1 (13) 3.0 (27) 55 ± 23 51 ± 25 85 ± 13 64 ± 17 14 ± 12 55 ± 18
A sensory profile of the test foods from a trained panel has been described elsewhere (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). a Shown per 100 g. Values calculated from recipe provided by the producer. b Ratings were performed on a 100-unit VAS scale after one spoonful of the test food. c Mean ratings differ significantly between liquid and semi-solid products (p < 0.05). Differences persisted over all test days, except for differences in familiarity ratings: these ratings increased over repeated consumption (F(1,376) = 44.02, p < 0.0001) and were not different on day 5.
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
Medical Ethical Committee (MEC) of Wageningen University and all procedures involving human subjects were approved by the MEC of Wageningen University (NL32016.081.10) and registered with the Dutch trial registration (NTR 2374). Test foods The test foods were selected on the basis of a preceding experiment (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011), in which 30 participants conducted a single measure of pleasantness and expected satiety effects of eight custards different in flavour (lemon or meringue) and texture (different structures with increasing viscosity). It was observed that texture, but not flavour, affected expectations. The products chosen were different in expected satiety effects and texture, but similar in liking. This was confirmed by ratings on perceived thickness and pleasantness in the current experiment (see Table 2 and ‘test food characteristics’ below). Meal duration was measured daily, and used to calculate eating rate (g/min). Both products were consumed with a spoon. The liquid product was flavoured with a lemon-aroma (Lemon Flavour JN-299-483-5; Givaudan SA Corp, Vernier, Switzerland), the semi-solid product with a meringue-aroma (Schuimgebak aroma DA53211, Givaudan SA Corp). A colouring agent (‘caramel’) was added to the semi-solid product to have the product appearance assumed to be congruent with flavour attributes (sulfiet-ammoniakkaramel E150d, ButterEssence B.V., Zaandam); no colouring agents were added to the liquid product. The test foods were milk based products specially developed for this study (Royal FrieslandCampina, Deventer) to obtain two isoenergetic products (with similar macronutrient content (Table 2). The degree of modification of the thickening agents (starch) was varied in the products to have a difference in their physical state: a liquid product, comparable with very thin custard, and a semi-solid product, comparable with firm pudding. The test foods were served in transparent bowls (1 L content) that were taken from the refrigerator 30 min before consumption. On days 1 and 5, all participants were offered 1000 g of the test food and instructed to eat until they were pleasantly satiated. Before and after consumption, the bowl was weighed to obtain ad libitum intake. On days 2, 3 and 4, participants consumed a fixed amount of the test food (Table 1). This amount provided 20% of the individual’s daily energy needs, estimated by means of the on Schofield equation 1 (WHO, 1985), taking into account gender, age, weight and a physical activity level of 1.6. Calculated amounts
421
were categorised per 25 g. Portions ranged from 400 to 575 g (average 468 ± 41 g) for women and from 550 to 650 g (average 594 ± 41 g) for men. There were no time constraints in consumption of this portion. Seven participants failed to consume the product in its entirety on day 2, of which six times in the semi-solid condition (leftovers: all <150 g). Two of these participants did not consume the entire amount of semi-solid custard on day 3 (leftovers: 54 g of 600 g served; and 83 g of 575 g). Appetite ratings after incomplete consumption were considered as missing values. Measurement expected satiety effects Expected satiety effects were measured using the ‘method of adjustment’ (based on (Brunstrom, Shakeshaft, & Scott-Samuel, 2008)). During the measurement, the fixed amount of test food was physically present and assessed against five commonly consumed ‘comparison foods’. The comparison foods were spaghetti Bolognese, penne and tomato sauce, rice curry (chicken tikka masala), oven fries, and cheese and tomato pizza. Pictures of each comparison food were shown on the screen of a laptop. The order of appearance of these foods was randomised across participants, and the trial started with a different and randomly selected amount for each comparison food. Participants were asked to ‘‘indicate the amount of food on the picture that would be equally satiating as the bowl of dairy product in front of you’’ based on a single spoonful of the test food. The amount on the screen could easily be adjusted using the arrow keys on the keyboard. The pictures were loaded with sufficient speed and continuous use of the arrow keys gave the impression that the change in amount of food was ‘animated’. The amount of food on the pictures ranged from 209 kJ (50 kcal) to 5225 kJ (1250 kcal). A total of 51 pictures was used to display these amounts, with a 1045-kJ portion (250 kcal) shown on picture number 25. Picture number 0 showed 0.2 times and picture number 50 displayed five times this amount. Procedures In each test condition, participants came to the research centre on five consecutive days and consumed the test food for breakfast at 7.30 (n = 24), 8.15 (n = 19) or 9.00 (n = 10). Participants were tested at in isolated sensory cabins during all sessions. Participants were asked to refrain from eating starting 23.00 the day before each test day. Non-caloric beverages were allowed up to 1 h before the test session. Participants were instructed not to eat anything for 3 h after each test session.
Fig. 1. Intake of the liquid and semi-solid product when offered ad libitum before (day 1) and after (day 5) repeated consumption of the fixed volume of food. The dotted line indicates the average of the fixed volume served on days 2, 3 and 4 (497 ± 67 g). N = 53, M/F = 12/41. ⁄For both products, intake on day 5 was higher than on day 1 (both (p < 0.01). #On both day 1 and day 5, intake of the liquid food was higher than intake of the semi-solid food (p < 0.0001).
422
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
On arrival, participants rated their appetite sensations (hunger, fullness, desire to eat, prospective consumption) (t = 0). Ratings were performed on a 100-mm visual analogue scale (VAS) (Stubbs et al., 2000), anchored ‘not at all’ to ‘extremely’. Participants then received the test food, consumed a single spoonful, and rated pleasantness, familiarity, and sensory attributes (sweetness, sourness, intensity of taste, and thickness). On days 1, 2 and 5, expected satiety effects were measured immediately after these ratings, before food consumption. On days 3 and 4, the test food was consumed immediately after rating its sensory attributes (Table 1). Next, participants rated appetite sensations at regular time intervals: immediately after consumption and at 30, 60, 90, 120, and 180 min after the start of the session. For five participants (n = 2 in liquid condition; n = 3 in semi-solid condition), data on appetite sensations in the second test condition were not available due to lack of compliance with the procedures, and considered as missing values. Statistical analyses Continuous variables are presented as means (±SD) and categorical variables are presented as frequencies. Ratings on pleasantness, familiarity and sensory attributes are reported as evaluated on day 1, and were analysed by means of a paired t-test (unless otherwise specified). Differences in ad libitum intake between product conditions, and before and after repeated consumption, as well as differences in eating rate, were tested using paired t-tests. The selected amounts (in kJ) of the five comparison foods were averaged for each participant. This average represents the expected satiety, i.e. the amount (in kJ) that is expected to be equally as filling as the fixed volume of the test food. Statistical analyses were conducted with log transformed data of the comparison foods to meet assumption of normality. For reasons of clarity, however, we converted the data into corresponding calories when mean values are presented. Expected satiety over time was tested by means of ANOVA (mixed model procedure) for an overall treatment (=product) effect, effect of repeated consumption (days) and their interaction. Tukey’s post hoc tests were used to test for differences between the products and test days, when the overall ANOVA indicated significant effects.
For all appetite sensations, total areas under the curve (AUC) from time points t = 0–180 were calculated (expand procedure, SAS). Differences in the AUCs between the liquid and semi-solid product were tested using ANOVA (mixed model procedure). The effect of repeated consumption and the product⁄ repeated consumption interaction effect on AUCs were tested separately for days 2, 3, and 4 (consuming the fixed volume) and day 1 and 5 (ad libitum consumption). Data were analysed using SAS (version 9.1; SAS Institute Inc.). Results at a p-value of <0.05 were considered significantly different. Results Ad libitum intake Average intake of the liquid product was higher compared with intake of the semi-solid product on both day 1 (t = 6.19; p < 0.0001) and day 5 (t = 5.64; p < 0.0001) (Fig. 1). On day 5, ad libitum intake was significantly higher than on day 1 for both products (both p < 0.01) (Fig. 1). Expected satiety On day 1, the semi-solid product (expected satiety: 1781 ± 924 kJ) was expected to be more satiating than the liquid product (1133 ± 472 kJ). This product effect persisted over all test days (F(1, 51) = 78.63; p < 0.0001). Overall, i.e. when data of both products were analysed together, expected satiety effects tended to increase over repeated consumption (F(1,258) = 3.15; p = 0.08). Post-hoc tests showed that expected satiety of the semi-solid product remained unchanged (F(1,102) = 0.94, p = 0.33) and that the increase in expected satiety of the liquid product (to 1216 ± 460 kJ on day 5) was borderline significant (F(1,63) = 2.11, p = 0.05). The product⁄ repeated consumption interaction was not statistically significant (F(1,258) = 0.28; p = 0.59). Appetite sensations Baseline (t = 0) ratings for hunger, fullness, desire to eat and prospective consumption did not differ between the product conditions or test sessions.
Fig. 2. Top: mean (±SEM) ratings for hunger on days 2, 3 and 4 on fixed time points up to 180 min and mean (±SEM total areas under the curve (AUC) after a fixed volume of liquid food (- -s- -) and semi-solid food (—j—) and N = 48, M/F = 9/39. Below: mean (±SEM) ratings for fullness on days 2, 3 and 4 on fixed time points up to 180 min and mean (±SEM total areas under the curve (AUC) after a fixed volume of liquid food (- -s- -) and semi-solid food (—j—). N = 48, M/F = 9/39.
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
After ad libitum intake, appetite sensations did not differ between liquid and semi-solid products (AUC not different), but ratings were lower for hunger, desire to eat and prospective consumption (all p < 0.0001) and higher for fullness (p = 0.0002) on day 5 compared with day 1. Areas under the curve after consumption of the fixed volume indicated that overall, the appetite response was stronger after the semi-solid product compared with the liquid product (product effects all appetite sensations: p < 0.0001). We also observed a significant product⁄ repeated consumption interaction for the AUCs of hunger (p = 0.005), fullness (p = 0.04) (Fig. 2) and desire to eat (p = 0.008); and a borderline significant interaction for prospective consumption (p = 0.06). Post-hoc tests showed that the semi-solid product decreased hunger, desire to eat, and prospective consumption and increased fullness more than the liquid product on day 2 (all p < 0.0001), but appetite sensations did not differ between the products on day 4, except for fullness (p = 0.03). Changes in hunger and fullness ratings were more pronounced after repeated consumption of the semi-solid foods than after the liquid foods (Fig. 2).
Test food characteristics Pleasantness ratings were not different between the test foods on any of the test days. The liquid food was rated as less familiar than the semi-solid food on day 1 (p < 0.0001) (Table 2). Ratings increased over repeated consumption (p < 0.0001) and were not different between the liquid and semi-solid foods on day 5. The sensory profile showed that the semi-solid product was evaluated to be thicker than the liquid product (p < 0.0001). The liquid product was evaluated to be sweeter (p = 0.04), more sour (p < 0.0001), and with a higher intensity of taste (p = 0.001) (Table 2). Overall, meal duration was longer for the fixed volume of semisolid (387 ± 140 s) compared with the liquid foods (295 ± 99 s) (p < 0.0001) and consequently, eating rate was higher for the liquid (114 ± 45 g/min) compared with the semi-solid products (88 ± 38 g/min) (p = 0.0004). Eating rate changed over repeated consumption (F(1,52) = 27.37; p < 0.0001): for both products, eating rate on day 5 was higher than on day 1. No differences were observed in meal duration between days 1 and 5.
Discussion Results of this study showed changes in ad libitum intake and appetite sensations after repeated consumption of iso-energetic foods with a difference in the sensory attributes and in the expected effect on satiety. However, expectations about the food’s effect on satiety did not change. Before repeated consumption (day 1), expected satiety effects of the semi-solid custards were higher than of the iso-energetic liquid custards; this was anticipated and consistent with the results of the preceding experiment with the same test foods (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). In line with these findings, ad libitum intake on day 1 was about 30% (100 g) higher for the liquid compared with the semi-solid product, while appetite sensations were similar. After repeated consumption (day 5), ad libitum intake increased for both the liquid and the semi-solid food, but expected satiation remained unchanged. Results showed that expected satiety of the liquid food tended to increase over repeated consumption, but the change was less than 100 kJ and this is considered not very meaningful. Participants reported a stronger satiety response after the fixed amount of semi-solid product compared with the liquid product on day 2. After repeated consumption, however, appetite sensations were similar: the foods were perceived to be equally
423
satiating. Still, ad libitum intake was higher, and expected satiety lower for the liquid compared with the semi-solid food on day 5. Pleasantness ratings of both foods were similar, which makes it unlikely that pleasantness accounts for these differences in expectations or intake between the test foods. In addition, findings of our preceding experiment showed that differences in flavour did not affect expectations about satiety effects for these specific test foods (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). Familiarity ratings were different between the liquid and semi-solid food differed at baseline, but not after repeated consumption. These ratings do not explain the differences in expected satiety effects and intake between the test foods. These differences may rely on other aspects: the texture attributes of the foods seem to play an important role in these outcome measures, also after repeated consumption of a food. These observations are in line with the satiety cascade (Blundell, Rogers, & Hill, 1987) that indicates that satiation and early satiety is driven by cognitive factors and sensory attributes, and late satiety by the post-ingestive and post-absorptive factors. The visual and oral cues differed greatly between the liquid and semi-solid test foods, and may explain the large differences between satiety expectations and intake. In addition, it seemed that the appetite sensations (early satiety outcomes) were initially based on the texture of the foods, i.e. on the differences in visual and oral cues. After repeated consumption of the liquid and semi-solid foods, however, participants experienced or learned that metabolic effects of the iso-energetic foods may not have been as different as expected. Visual inspection of Fig. 2 suggests that the loss of difference in these ratings was caused especially by adjustment in the ratings after consumption of the semi-solid food. This suggests that the ‘anticipated satiety’ effect (that was not reflected in the post-ingestive effects) was especially high for the relatively large portion of semisolid food. This shows the importance of cognitive cues in eating behaviour (Herman & Polivy, 2005; Wansink, 2004), and the role of texture in this (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). More important, it indicates that physiological ‘learning’ appears to occur more readily than adjustments of the corresponding expectations about the food’s satiety effect. The method used for measuring the expected satiety (based on (Brunstrom et al., 2008) was shown to be a reproducible measure and a precise means to assess expectations (Hogenkamp, Brunstrom, et al., 2011; Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). Moreover, previous studies showed that snack- or dessertspecific food items that were physically present, like fruit smoothies (Brunstrom, Brown, Hinton, Rogers, & Fay, 2011), sorbets (Wilkinson & Brunstrom, 2009), and dairy smoothies (Fay, Hinton, Rogers & Brunstrom et al., 2011) can be compared with meal shown on pictures successfully. A likely explanation of the absence of changes in expected satiety is that the satiety beliefs are still based on prior experiences with different food (texture) stimuli, therewith limiting the extent to which learning will result in behavioural changes (Hogenkamp, Brunstrom, et al., 2011; O’Sullivan et al., 2010). Both test foods were served as a breakfast replacer and consumed with a spoon, but the liquid food may still have been evaluated as a ‘drink’ rather than as a ‘food’. The different consumption context of liquids and (semi-)solid foods, i.e. quenching thirst vs. decreasing hunger feelings, contribute to the differences in intake and in expectations between the liquid and semisolids. The relatively low number of exposures to the test foods may also have limited behavioural changes. We do, however, not expect that prolonging the consumption period would have resulted in pronounced behavioural changes (Hogenkamp et al., 2010; Mars et al., 2009), such as expectations about the foods’ satiety effects. In addition, it was observed that e.g. viscosity tended to affect the time to lunch (behavioural satiety outcome) when consuming low-viscosity or high-viscosity rice breakfasts for 5 days
424
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
in a row (Russell & Delahunty, 2004), although intake at lunch was not reported. It would therefore be more interesting to investigate if the changes in appetite scores can be replicated and if this will result in changes in subsequent intake, rather than prolonging the consumption period. In addition, choice of the test foods, regarding issues of novelty or nutrient content, may be of greater methodological importance to observe changes in behaviour than the number of trials included in the design (Yeomans, 2012). We hypothesised that expected satiety plays a role in decisions on portion size (Brunstrom & Rogers, 2009), but the increase in ad libitum intake was not preceded by profound changes in expected satiety. This suggests that this role may be only small or easily overruled by food attributes, especially when one single food is consumed during the eating occasion. Results from previous studies show that, in the view of texture attributes, perceived volume (Brunstrom, Collingwood, & Rogers, 2010) and eating rate are also of great importance in ad libitum intake (Andrade, Greene, & Melanson, 2008; Martin et al., 2007; Zijlstra et al., 2008). Ad libitum intake increased after repeated consumption compared to baseline. This could be explained by the increase in the eating rate over repeated consumption (Andrade et al., 2008; Martin et al., 2007; Smit, Kemsley, Tapp, & Henry, 2011). Another explanation may be that participants habituated to consumption of the test foods, and adjusted ad libitum intake to the fixed volume they had been exposed to during the week. We observed that several participants struggled to finish the large volume of especially the semi-solid foods on day 2, but that they did not encounter this problem on day 4. It would be interesting to investigate if people can get used to portions larger or smaller than their self-selected ideal portion sizes within several days. This may contribute to limiting the ‘passive’ overconsumption, in which intake is higher than required to feel satiated (Prentice & Jebb, 2003). We conclude that the changes in hunger and fullness indicated that the fixed volume of liquid and solid food was perceived to be equally satiating after repeated consumption, but this did not result in the anticipated changes. Ad libitum intake increased over repeated consumption in the direction of the fixed portion size. Intake of the liquid foods, however, remained higher compared with the semi-solid foods. Texture remained important in ad libitum intake and expected satiety. These results show that decisions on meal size do not easily change, and that cognitive cues based on the food’s sensory attributes remain of great importance in short-term intake. Sensory attributes of a food, however, are not always congruent to its energy content. This may help to better understanding of both short and longer term regulation of energy intake. Liquid foods are not expected to be filling, and may therefore promote overconsumption, especially in single-item meals (e.g. drinks, shakes). In addition, habituation to the satiating effects of foods and a possible mismatch with ‘anticipated’ and ‘experienced’ satiety may e.g. explain why it is hard to keep up with e.g. consuming low-caloric weight loss products for prolonged periods of time.
References Almiron-Roig, E., Chen, Y., & Drewnowski, A. (2003). Liquid calories and the failure of satiety. How good is the evidence? Obesity Reviews, 4, 201–212. Andrade, A. M., Greene, G. W., & Melanson, K. J. (2008). Eating slowly led to decreases in energy intake within meals in healthy women. Journal of the American Dietetic Association, 108, 1186–1191. Blundell, J., Rogers, P. J., & Hill, A. J. (1987). Evaluating the satiating power of foods. Implications for acceptance and consumption. In J. Solms & D. A. Booth (Eds.), Food Acceptance and Nutrition (pp. 205–219). London: Academic Press. Booth, D. A., & Davis, J. D. (1973). Gastrointestinal factors in the acquisition of oral sensory control of satiation. Physiology & Behavior, 11, 23–29. Booth, D. A., & Grinker, J. A. (1993). Learned control of meal size in spontaneously obese and nonobese bonnet macaque monkeys. Physiology & Behavior, 53, 51–57.
Brunstrom, Jeffrey M. (2007). Associative learning and the control of human dietary behavior. Appetite, 49, 268–271. Brunstrom, J. M., Brown, S., Hinton, E. C., Rogers, P. J., & Fay, S. H. (2011). ‘Expected satiety’ changes hunger and fullness in the inter-meal interval. Appetite, 56, 310–315. Brunstrom, J. M., Collingwood, J., & Rogers, P. J. (2010). Perceived volume, expected satiation, and the energy content of self-selected meals. Appetite, 55, 25–29. Brunstrom, J. M., & Rogers, P. J. (2009). How many calories are on our plate? Expected fullness, not liking, determines meal-size selection. Obesity, 17, 1884–1890. Brunstrom, J. M., Shakeshaft, N. G., & Scott-Samuel, N. E. (2008). Measuring ‘expected satiety’ in a range of common foods using a method of constant stimuli. Appetite, 51, 604–614. Cardello, A. V., & Sawyer, F. M. (1992). Effects of disconfirmed consumer expectations on food acceptability. Journal of Sensory Studies, 7, 253–277. Herman, C. P., & Polivy, J. (2005). Normative influences on food intake. Physiology & Behavior, 86, 762–772. Hogenkamp, P. S., Brunstrom, J. M., Stafleu, A., Mars, M., & de Graaf, C. (2011). Expected satiation after repeated consumption of low- or high-energy-dense soup. British Journal of Nutrition. Hogenkamp, P. S., Mars, M., Stafleu, A., & de Graaf, C. (2010). Intake during repeated exposure to low- and high-energy-dense yogurts by different means of consumption. American Journal of Clinical Nutrition, 91, 841–847. Hogenkamp, P. S., Stafleu, A., Mars, M., Brunstrom, J. M., & de Graaf, C. (2011). Texture, not flavor, determines expected satiation of dairy products. Appetite, 57, 635–641. Hogenkamp, P. S., Stafleu, A., Mars, M., & de Graaf, C. (2011). Learning about the energy density of liquid and semi-solid foods. International Journal of Obesity. Hulshof, T., De Graaf, C., & Weststrate, J. A. (1993). The effects of preloads varying in physical state and fat content on satiety and energy intake. Appetite, 21, 273–286. Mars, M., Hogenkamp, P. S., Gosses, A. M., Stafleu, A., & De Graaf, C. (2009). Effect of viscosity on learned satiation. Physiology & Behavior, 98, 60–66. Martin, C. K., Anton, S. D., Walden, H., Arnett, C., Greenway, F. L., & Williamson, D. A. (2007). Slower eating rate reduces the food intake of men, but not women. Implications for behavioral weight control. Behaviour Research and Therapy, 45, 2349–2359. Mattes, R. D., & Rothacker, D. (2001). Beverage viscosity is inversely related to postprandial hunger in humans. Physiology & Behavior, 74, 551–557. O’Sullivan, H. L., Alexander, E., Ferriday, D., & Brunstrom, J. M. (2010). Effects of repeated exposure on liking for a reduced-energy-dense food. American Journal of Clinical Nutrition, 91, 1584–1589. Prentice, A. M., & Jebb, S. A. (2003). Fast foods, energy density and obesity. A possible mechanistic link. Obesity Reviews, 4, 187–194. Russell, K., & Delahunty, C. (2004). The effect of viscosity and volume on pleasantness and satiating power of rice milk. Food Quality and Preference, 15, 743–750. Smit, H. J., Kemsley, E. K., Tapp, H. S., & Henry, C. J. K. (2011). Does prolonged chewing reduce food intake? Fletcherism revisited. Appetite, 57, 295–298. Sorensen, L. B., Moller, P., Flint, A., Martens, M., & Raben, A. (2003). Effect of sensory perception of foods on appetite and food intake. A review of studies on humans. International Journal of Obesity and Related Metabolic Disorders, 27, 1152–1166. Stubbs, R. J., Hughes, D. A., Johnstone, A. M., Rowley, E., Reid, C., Elia, M., Stratton, R., Delargy, H., King, N., & Blundell, J. E. (2000). The use of visual analogue scales to assess motivation to eat in human subjects. A review of their reliability and validity with an evaluation of new hand-held computerized systems for temporal tracking of appetite ratings. British Journal of Nutrition, 84, 405–415. Tepper, B. J., Mattes, R. D., & Farkas, B. K. (1991). Learned flavor cues influence food intake in humans. Journal of Sensory Studies, 6, 89–100. Wansink, B. (2004). Environmental factors that increase the food intake and consumption volume of unknowing consumers. Annual Review of Nutrition, 24, 455–479. Wansink, B., Ittersum, K., & Painter, J. E. (2004). How diet and health labels influence taste and satiation. Journal of Food Science, 69, S340–S346. Warwick, Z. S., & Schiffman, S. S. (1991). Flavor-calorie relationships. Effect on weight gain in rats. Physiology & Behavior, 50, 465–470. WHO. (1985). Estimates of energy and protein requirements of adults and children. In Energy and protein requirements. Report of a Joint FAO/WHO/UNU Expert consultation (2 ed., pp. 71–112). Geneva: World Health Organization. Wilkinson, L. L., & Brunstrom, J. M. (2009). Conditioning ‘fullness expectations’ in a novel dessert. Appetite, 52, 780–783. Yeomans, Martin R. (2012). Flavour–nutrient learning in humans. An elusive phenomenon? Physiology & Behavior, 106, 345–355. Yeomans, M. R., Chambers, L., Blumenthal, H., & Blake, A. (2008). The role of expectancy in sensory and hedonic evaluation. The case of smoked salmon icecream. Food Quality and Preference, 19, 565–573. Yeomans, M. R., Weinberg, L., & James, S. (2005). Effects of palatability and learned satiety on energy density influences on breakfast intake in humans. Physiology & Behavior, 86, 487–499. Zandstra, E. H., Stubenitsky, K., De Graaf, C., & Mela, D. J. (2002). Effects of learned flavour cues on short-term regulation of food intake in a realistic setting. Physiology & Behavior, 75, 83–90. Zijlstra, N., Mars, M., de Wijk, R. A., Westerterp-Plantenga, M. S., & de Graaf, C. (2008). The effect of viscosity on ad libitum food intake. International Journal of Obesity, 32, 676–683.
Contents lists available at SciVerse ScienceDirect
Appetite journal homepage: www.elsevier.com/locate/appet
Research report
Repeated consumption of a large volume of liquid and semi-solid foods increases ad libitum intake, but does not change expected satiety q P.S. Hogenkamp a,b, M. Mars a,b, A. Stafleu a,c, C. de Graaf a,b,⇑ a
Top Institute Food and Nutrition, P.O. Box 557, NL-6700 AN, Wageningen, The Netherlands Division of Human Nutrition, Wageningen University, P.O. Box 8129, NL-6700 EV Wageningen, The Netherlands c TNO, P.O. Box 360, NL-3700 AJ, Zeist, The Netherlands b
a r t i c l e
i n f o
Article history: Received 25 November 2011 Received in revised form 23 May 2012 Accepted 11 June 2012 Available online 18 June 2012 Keywords: Food intake Texture Satiation Satiety Energy density Expectations Dietary learning
a b s t r a c t Food intake and a food’s expected satiating effect initially rely on sensory attributes. People will learn about the food’s satiating capacity by exposure. We investigated whether repeated consumption changed the expected satiety effects and intake of iso-energetic liquid and semi-solid foods. In a randomised cross-over study, participants (n = 53; age: 21 ± 2.9 y; BMI: 21.8 ± 2.0 kg/m2) consumed one of two isoenergetic dairy foods (liquid or semi-solid) for breakfast in each 5-day test condition. Expectations regarding satiety were measured on days 1, 2, and 5. Foods were offered ad libitum on days 1 and 5 and in a fixed volume on days 2–4. Appetite sensations were rated up to 180 min after the start of the session on fixed time points. Expected satiety effects of the semi-solid food were higher than of the liquid food on all days (p < 0.0001). Ad libitum intake of the liquid food was higher than of the semi-solid food on day 1 (liquid: 391 ± 177 g, semi-solid: 277 ± 98 g; p < 0.0001) and day 5 (liquid: 477 ± 161 g, semi-solid: 375 ± 148 g; p < 0.0001). On day 2, hunger was rated lower and fullness rated higher after the semi-solid compared with the liquid food; on day 4, no differences were observed (significant product⁄ exposure interaction AUC). Changes in hunger and fullness indicated that the fixed volumes of liquid and solid food were perceived to be equally satiating after repeated consumption, but this did not result in the anticipated changes: expected satiety effects remained lower, and ad libitum intake higher for the liquid compared with the semi-solid food. The effect of texture on a food’s expected satiety effects and its ad libitum intake appears to be large, also after repeated consumption. Expectations based on sensory cues are not easily changed. Ó 2012 Elsevier Ltd. All rights reserved.
Introduction The intake of a particular food is partly based on previous experience with the food. People may learn about the energy content of a food by exposure and link post-ingestive appetite sensations to the food’s sensory attributes (Yeomans, Weinberg, & James, 2005; Zandstra, Stubenitsky, De Graaf, & Mela, 2002). More precise, meal size will be largely controlled by a learned decrease in appetite towards the end of a meal, which has been defined as learned satiety (Booth & Davis, 1973; Booth & Grinker, 1993; Brunstrom, 2007). Learned q Acknowledgements: We kindly thank Jeffrey Brunstrom for sharing with us his computer task, Natasja Hück for research assistance, and Hans Meijer for technical support. The studies were supported by the Top Institute Food and Nutrition. The authors’ responsibilities were as follows—PSH: designed the study, collected and analysed the data, and wrote the manuscript; MM, AS, and CdG designed the study and provided significant advice and consultation. None of the authors had a personal or financial conflict of interest. ⇑ Corresponding author. E-mail address: [email protected] (C. de Graaf).
0195-6663/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.appet.2012.06.008
satiety will result in expectations about the food’s satiating capacity. These expected effects of a food on satiety may play a role in decisions on portion size (Brunstrom & Rogers, 2009), and consequently in energy intake. Learning about the energy content of a food may change these expected effects on satiety (Wilkinson & Brunstrom, 2009), with changes in portion size as a possible consequence. Expectations about the satiety effects of a specific food are among others based on its sensory attributes (Sorensen, Moller, Flint, Martens, & Raben, 2003). One of the sensory attributes that is important in the food’s expected satiety is texture. In a preceding experiment, we observed that an increase in perceived thickness of several dairy products resulted in a consistent increase in the expected satiating capacity of the foods (Hogenkamp, Stafleu, Mars, Brunstrom, & de Graaf, 2011). This is in line with the results of several studies that showed that solid foods produce higher satiety ratings than iso-energetic liquid foods (e.g. (Hulshof, De Graaf, & Weststrate, 1993; Mattes & Rothacker, 2001)), although the effect of texture on satiety and subsequent intake may critically depend on the volume of the liquid and solid preloads and on the time
420
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
Table 1 Overview of the study design. In this cross-over design, participants were served dairy foods for breakfast on five consecutive days in each condition.a Conditiona Liquid Semi-solid
Food intake Measurementsb:
Day 1
Day 2
Day 3
Day 4
Day 5
Ad libitum Expected satiety Sensory attributes Appetite sensations
Fixed Expected satiety Sensory attributes Appetite sensations
Fixed – Sensory attributes Appetite sensations
Fixed – Sensory attributes Appetite sensations
Ad libitum Expected satiety Sensory attributes Appetite sensations
a
The order of conditions was randomised and balanced within subjects. Conditions were separated by a period of at least one week. Expected satiety was measured using the ‘method of adjustment’ as described in the methods. Sensory attributes included pleasantness and familiarity, all rated on a visual analogue scale. Appetite sensations were rated on a visual analogue scale up to 180 min after the start of the session. b
interval to the next meal (Almiron-Roig, Chen, & Drewnowski, 2003; Hogenkamp, Stafleu, Mars, & de Graaf, 2011). In line with the (expected) satiety effects of liquid and solid foods, previous studies consistently showed that ad libitum intake of solid or semi-solid foods was lower than of liquid foods with similar macronutrient composition (Hogenkamp, Mars, Stafleu, & de Graaf, 2010; Mars, Hogenkamp, Gosses, Stafleu, & De Graaf, 2009; Zijlstra, Mars, de Wijk, Westerterp-Plantenga, & de Graaf, 2008). This suggests that learned satiety and expectations are indeed associated with satiation (portion size). Expected satiety of a specific food also depends on the energy content of the selected food item. Previous studies, however, showed that expectations did not change when repeatedly consuming low or high-energy-dense foods that were similar in appearance (Hogenkamp, Brunstrom, Stafleu, Mars, & de Graaf, 2011; O’Sullivan, Alexander, Ferriday, & Brunstrom, 2010). When satiety expectations based on the sensory attributes of a food are not congruent with the food’s energy content and its actual satiety sensations, this may disrupt energy intake regulation (Tepper, Mattes, & Farkas, 1991; Warwick & Schiffman, 1991). On the other hand, a large incongruence (between expected and experienced satiety) may promote changes in a food’s expected satiety effect, as is observed for hedonic expectations. Studies investigating hedonic expectations show that evaluation after consumption is generally adjusted towards expectations (Cardello & Sawyer, 1992; Wansink, Ittersum, & Painter, 2004). Only when there is a large contrast between expectations and evaluation (Yeomans, Chambers, Blumenthal, & Blake, 2008), e.g. when foods taste much better than expected (Wansink et al., 2004), future expectations will be adjusted. We hypothesised that a large contrast between expectations about a food’s effect on satiety and the experienced satiety value would result in adjustments in these expectations, and in changes in intake accordingly. Evaluation of the satiety process may, however, be more subtle than the (explicit) hedonic evaluation, and therefore requires repeated exposures to the food. In the current experiment, we investigated changes in expected satiety after repeated consumption of iso-energetic foods with different sensory attributes, and with different expected satiety at baseline (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011): the expected satiety of the test foods was either low (a liquid dairy food) or high (a semi-solid dairy food). We hypothesised that changes would be most pronounced in the condition with the liquid food, which was assumed to provide a higher perceived satiety than expected. Methods Design In a randomised cross-over design with two test conditions, participants were served dairy foods for breakfast on five consecutive days (Monday to Friday), see Table 1. The order of conditions was randomised and balanced within subjects. Conditions were separated by a period of at least one week. In each test condition,
participants consumed one of the two iso-energetic test foods: food with a low expected satiety at baseline (liquid ‘custard’) or food with a high expected satiety at baseline (semi-solid ‘pudding’). On days 1 and 5, test foods were offered ad libitum and intake was measured; on days 2, 3, and 4, a fixed volume of test food was consumed entirely. Expected satiety of the products was measured on days 1, 2 and 5. Preceding the study, the participants were trained in the procedures. Participants Healthy, normal weight participants, aged 18–40 y, were recruited from Wageningen and the surroundings. Exclusion criteria were: restrained eating [Dutch eating behaviour questionnaire: men score >2.25; women score >2.80 (26)], lack of appetite, following an energy-restricted diet or change in body weight >5 kg during the last 2 months, stomach or bowel diseases, diabetes, thyroid disease or any other endocrine disorder, hypersensitivity for food products under study, smoking, and being a vegetarian. A total of 50 participants would provide 80% power to detect a 20% difference in expectations regarding the satiating effect of the liquid food. In total, 53 participants (M/F: 12/41, age: 21 ± 2.9 y, BMI: 21.8 ± 2.0 kg/m2) completed the study. Participants were unaware of the exact aim of the study. In a screening session before the start of the study, they were informed that we were interested in taste sensations and acceptance of dairy products for breakfast. During this screening, body weight and height was measured and all participants signed an informed consent. Participants were debriefed after the study and received financial compensation. The present study was conducted according to the ethical standards of the
Table 2 Composition and characteristics of the liquid and semi-solid test foods: energy, macronutrientsa and energy percentage; ratingsb (mean ± SD) on pleasantness, familiarity and sensory attributes on day 1 (N = 53, M/F = 12/41).
Energy (kcal/kJ) Carbohydrates, g (% energy) Protein, g (% energy) Fat, g (% energy) Pleasantness Familiarityc Thickc Sweetc Sourc Intensity of tastec
Liquid
Semi-solid
96/402 14.0 (59) 3.1 (13) 3.0 (28) 58 ± 22 36 ± 24 15 ± 14 71 ± 19 40 ± 25 66 ± 14
98/410 14.5 (60) 3.1 (13) 3.0 (27) 55 ± 23 51 ± 25 85 ± 13 64 ± 17 14 ± 12 55 ± 18
A sensory profile of the test foods from a trained panel has been described elsewhere (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). a Shown per 100 g. Values calculated from recipe provided by the producer. b Ratings were performed on a 100-unit VAS scale after one spoonful of the test food. c Mean ratings differ significantly between liquid and semi-solid products (p < 0.05). Differences persisted over all test days, except for differences in familiarity ratings: these ratings increased over repeated consumption (F(1,376) = 44.02, p < 0.0001) and were not different on day 5.
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
Medical Ethical Committee (MEC) of Wageningen University and all procedures involving human subjects were approved by the MEC of Wageningen University (NL32016.081.10) and registered with the Dutch trial registration (NTR 2374). Test foods The test foods were selected on the basis of a preceding experiment (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011), in which 30 participants conducted a single measure of pleasantness and expected satiety effects of eight custards different in flavour (lemon or meringue) and texture (different structures with increasing viscosity). It was observed that texture, but not flavour, affected expectations. The products chosen were different in expected satiety effects and texture, but similar in liking. This was confirmed by ratings on perceived thickness and pleasantness in the current experiment (see Table 2 and ‘test food characteristics’ below). Meal duration was measured daily, and used to calculate eating rate (g/min). Both products were consumed with a spoon. The liquid product was flavoured with a lemon-aroma (Lemon Flavour JN-299-483-5; Givaudan SA Corp, Vernier, Switzerland), the semi-solid product with a meringue-aroma (Schuimgebak aroma DA53211, Givaudan SA Corp). A colouring agent (‘caramel’) was added to the semi-solid product to have the product appearance assumed to be congruent with flavour attributes (sulfiet-ammoniakkaramel E150d, ButterEssence B.V., Zaandam); no colouring agents were added to the liquid product. The test foods were milk based products specially developed for this study (Royal FrieslandCampina, Deventer) to obtain two isoenergetic products (with similar macronutrient content (Table 2). The degree of modification of the thickening agents (starch) was varied in the products to have a difference in their physical state: a liquid product, comparable with very thin custard, and a semi-solid product, comparable with firm pudding. The test foods were served in transparent bowls (1 L content) that were taken from the refrigerator 30 min before consumption. On days 1 and 5, all participants were offered 1000 g of the test food and instructed to eat until they were pleasantly satiated. Before and after consumption, the bowl was weighed to obtain ad libitum intake. On days 2, 3 and 4, participants consumed a fixed amount of the test food (Table 1). This amount provided 20% of the individual’s daily energy needs, estimated by means of the on Schofield equation 1 (WHO, 1985), taking into account gender, age, weight and a physical activity level of 1.6. Calculated amounts
421
were categorised per 25 g. Portions ranged from 400 to 575 g (average 468 ± 41 g) for women and from 550 to 650 g (average 594 ± 41 g) for men. There were no time constraints in consumption of this portion. Seven participants failed to consume the product in its entirety on day 2, of which six times in the semi-solid condition (leftovers: all <150 g). Two of these participants did not consume the entire amount of semi-solid custard on day 3 (leftovers: 54 g of 600 g served; and 83 g of 575 g). Appetite ratings after incomplete consumption were considered as missing values. Measurement expected satiety effects Expected satiety effects were measured using the ‘method of adjustment’ (based on (Brunstrom, Shakeshaft, & Scott-Samuel, 2008)). During the measurement, the fixed amount of test food was physically present and assessed against five commonly consumed ‘comparison foods’. The comparison foods were spaghetti Bolognese, penne and tomato sauce, rice curry (chicken tikka masala), oven fries, and cheese and tomato pizza. Pictures of each comparison food were shown on the screen of a laptop. The order of appearance of these foods was randomised across participants, and the trial started with a different and randomly selected amount for each comparison food. Participants were asked to ‘‘indicate the amount of food on the picture that would be equally satiating as the bowl of dairy product in front of you’’ based on a single spoonful of the test food. The amount on the screen could easily be adjusted using the arrow keys on the keyboard. The pictures were loaded with sufficient speed and continuous use of the arrow keys gave the impression that the change in amount of food was ‘animated’. The amount of food on the pictures ranged from 209 kJ (50 kcal) to 5225 kJ (1250 kcal). A total of 51 pictures was used to display these amounts, with a 1045-kJ portion (250 kcal) shown on picture number 25. Picture number 0 showed 0.2 times and picture number 50 displayed five times this amount. Procedures In each test condition, participants came to the research centre on five consecutive days and consumed the test food for breakfast at 7.30 (n = 24), 8.15 (n = 19) or 9.00 (n = 10). Participants were tested at in isolated sensory cabins during all sessions. Participants were asked to refrain from eating starting 23.00 the day before each test day. Non-caloric beverages were allowed up to 1 h before the test session. Participants were instructed not to eat anything for 3 h after each test session.
Fig. 1. Intake of the liquid and semi-solid product when offered ad libitum before (day 1) and after (day 5) repeated consumption of the fixed volume of food. The dotted line indicates the average of the fixed volume served on days 2, 3 and 4 (497 ± 67 g). N = 53, M/F = 12/41. ⁄For both products, intake on day 5 was higher than on day 1 (both (p < 0.01). #On both day 1 and day 5, intake of the liquid food was higher than intake of the semi-solid food (p < 0.0001).
422
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
On arrival, participants rated their appetite sensations (hunger, fullness, desire to eat, prospective consumption) (t = 0). Ratings were performed on a 100-mm visual analogue scale (VAS) (Stubbs et al., 2000), anchored ‘not at all’ to ‘extremely’. Participants then received the test food, consumed a single spoonful, and rated pleasantness, familiarity, and sensory attributes (sweetness, sourness, intensity of taste, and thickness). On days 1, 2 and 5, expected satiety effects were measured immediately after these ratings, before food consumption. On days 3 and 4, the test food was consumed immediately after rating its sensory attributes (Table 1). Next, participants rated appetite sensations at regular time intervals: immediately after consumption and at 30, 60, 90, 120, and 180 min after the start of the session. For five participants (n = 2 in liquid condition; n = 3 in semi-solid condition), data on appetite sensations in the second test condition were not available due to lack of compliance with the procedures, and considered as missing values. Statistical analyses Continuous variables are presented as means (±SD) and categorical variables are presented as frequencies. Ratings on pleasantness, familiarity and sensory attributes are reported as evaluated on day 1, and were analysed by means of a paired t-test (unless otherwise specified). Differences in ad libitum intake between product conditions, and before and after repeated consumption, as well as differences in eating rate, were tested using paired t-tests. The selected amounts (in kJ) of the five comparison foods were averaged for each participant. This average represents the expected satiety, i.e. the amount (in kJ) that is expected to be equally as filling as the fixed volume of the test food. Statistical analyses were conducted with log transformed data of the comparison foods to meet assumption of normality. For reasons of clarity, however, we converted the data into corresponding calories when mean values are presented. Expected satiety over time was tested by means of ANOVA (mixed model procedure) for an overall treatment (=product) effect, effect of repeated consumption (days) and their interaction. Tukey’s post hoc tests were used to test for differences between the products and test days, when the overall ANOVA indicated significant effects.
For all appetite sensations, total areas under the curve (AUC) from time points t = 0–180 were calculated (expand procedure, SAS). Differences in the AUCs between the liquid and semi-solid product were tested using ANOVA (mixed model procedure). The effect of repeated consumption and the product⁄ repeated consumption interaction effect on AUCs were tested separately for days 2, 3, and 4 (consuming the fixed volume) and day 1 and 5 (ad libitum consumption). Data were analysed using SAS (version 9.1; SAS Institute Inc.). Results at a p-value of <0.05 were considered significantly different. Results Ad libitum intake Average intake of the liquid product was higher compared with intake of the semi-solid product on both day 1 (t = 6.19; p < 0.0001) and day 5 (t = 5.64; p < 0.0001) (Fig. 1). On day 5, ad libitum intake was significantly higher than on day 1 for both products (both p < 0.01) (Fig. 1). Expected satiety On day 1, the semi-solid product (expected satiety: 1781 ± 924 kJ) was expected to be more satiating than the liquid product (1133 ± 472 kJ). This product effect persisted over all test days (F(1, 51) = 78.63; p < 0.0001). Overall, i.e. when data of both products were analysed together, expected satiety effects tended to increase over repeated consumption (F(1,258) = 3.15; p = 0.08). Post-hoc tests showed that expected satiety of the semi-solid product remained unchanged (F(1,102) = 0.94, p = 0.33) and that the increase in expected satiety of the liquid product (to 1216 ± 460 kJ on day 5) was borderline significant (F(1,63) = 2.11, p = 0.05). The product⁄ repeated consumption interaction was not statistically significant (F(1,258) = 0.28; p = 0.59). Appetite sensations Baseline (t = 0) ratings for hunger, fullness, desire to eat and prospective consumption did not differ between the product conditions or test sessions.
Fig. 2. Top: mean (±SEM) ratings for hunger on days 2, 3 and 4 on fixed time points up to 180 min and mean (±SEM total areas under the curve (AUC) after a fixed volume of liquid food (- -s- -) and semi-solid food (—j—) and N = 48, M/F = 9/39. Below: mean (±SEM) ratings for fullness on days 2, 3 and 4 on fixed time points up to 180 min and mean (±SEM total areas under the curve (AUC) after a fixed volume of liquid food (- -s- -) and semi-solid food (—j—). N = 48, M/F = 9/39.
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
After ad libitum intake, appetite sensations did not differ between liquid and semi-solid products (AUC not different), but ratings were lower for hunger, desire to eat and prospective consumption (all p < 0.0001) and higher for fullness (p = 0.0002) on day 5 compared with day 1. Areas under the curve after consumption of the fixed volume indicated that overall, the appetite response was stronger after the semi-solid product compared with the liquid product (product effects all appetite sensations: p < 0.0001). We also observed a significant product⁄ repeated consumption interaction for the AUCs of hunger (p = 0.005), fullness (p = 0.04) (Fig. 2) and desire to eat (p = 0.008); and a borderline significant interaction for prospective consumption (p = 0.06). Post-hoc tests showed that the semi-solid product decreased hunger, desire to eat, and prospective consumption and increased fullness more than the liquid product on day 2 (all p < 0.0001), but appetite sensations did not differ between the products on day 4, except for fullness (p = 0.03). Changes in hunger and fullness ratings were more pronounced after repeated consumption of the semi-solid foods than after the liquid foods (Fig. 2).
Test food characteristics Pleasantness ratings were not different between the test foods on any of the test days. The liquid food was rated as less familiar than the semi-solid food on day 1 (p < 0.0001) (Table 2). Ratings increased over repeated consumption (p < 0.0001) and were not different between the liquid and semi-solid foods on day 5. The sensory profile showed that the semi-solid product was evaluated to be thicker than the liquid product (p < 0.0001). The liquid product was evaluated to be sweeter (p = 0.04), more sour (p < 0.0001), and with a higher intensity of taste (p = 0.001) (Table 2). Overall, meal duration was longer for the fixed volume of semisolid (387 ± 140 s) compared with the liquid foods (295 ± 99 s) (p < 0.0001) and consequently, eating rate was higher for the liquid (114 ± 45 g/min) compared with the semi-solid products (88 ± 38 g/min) (p = 0.0004). Eating rate changed over repeated consumption (F(1,52) = 27.37; p < 0.0001): for both products, eating rate on day 5 was higher than on day 1. No differences were observed in meal duration between days 1 and 5.
Discussion Results of this study showed changes in ad libitum intake and appetite sensations after repeated consumption of iso-energetic foods with a difference in the sensory attributes and in the expected effect on satiety. However, expectations about the food’s effect on satiety did not change. Before repeated consumption (day 1), expected satiety effects of the semi-solid custards were higher than of the iso-energetic liquid custards; this was anticipated and consistent with the results of the preceding experiment with the same test foods (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). In line with these findings, ad libitum intake on day 1 was about 30% (100 g) higher for the liquid compared with the semi-solid product, while appetite sensations were similar. After repeated consumption (day 5), ad libitum intake increased for both the liquid and the semi-solid food, but expected satiation remained unchanged. Results showed that expected satiety of the liquid food tended to increase over repeated consumption, but the change was less than 100 kJ and this is considered not very meaningful. Participants reported a stronger satiety response after the fixed amount of semi-solid product compared with the liquid product on day 2. After repeated consumption, however, appetite sensations were similar: the foods were perceived to be equally
423
satiating. Still, ad libitum intake was higher, and expected satiety lower for the liquid compared with the semi-solid food on day 5. Pleasantness ratings of both foods were similar, which makes it unlikely that pleasantness accounts for these differences in expectations or intake between the test foods. In addition, findings of our preceding experiment showed that differences in flavour did not affect expectations about satiety effects for these specific test foods (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). Familiarity ratings were different between the liquid and semi-solid food differed at baseline, but not after repeated consumption. These ratings do not explain the differences in expected satiety effects and intake between the test foods. These differences may rely on other aspects: the texture attributes of the foods seem to play an important role in these outcome measures, also after repeated consumption of a food. These observations are in line with the satiety cascade (Blundell, Rogers, & Hill, 1987) that indicates that satiation and early satiety is driven by cognitive factors and sensory attributes, and late satiety by the post-ingestive and post-absorptive factors. The visual and oral cues differed greatly between the liquid and semi-solid test foods, and may explain the large differences between satiety expectations and intake. In addition, it seemed that the appetite sensations (early satiety outcomes) were initially based on the texture of the foods, i.e. on the differences in visual and oral cues. After repeated consumption of the liquid and semi-solid foods, however, participants experienced or learned that metabolic effects of the iso-energetic foods may not have been as different as expected. Visual inspection of Fig. 2 suggests that the loss of difference in these ratings was caused especially by adjustment in the ratings after consumption of the semi-solid food. This suggests that the ‘anticipated satiety’ effect (that was not reflected in the post-ingestive effects) was especially high for the relatively large portion of semisolid food. This shows the importance of cognitive cues in eating behaviour (Herman & Polivy, 2005; Wansink, 2004), and the role of texture in this (Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). More important, it indicates that physiological ‘learning’ appears to occur more readily than adjustments of the corresponding expectations about the food’s satiety effect. The method used for measuring the expected satiety (based on (Brunstrom et al., 2008) was shown to be a reproducible measure and a precise means to assess expectations (Hogenkamp, Brunstrom, et al., 2011; Hogenkamp, Stafleu, Mars, Brunstrom, et al., 2011). Moreover, previous studies showed that snack- or dessertspecific food items that were physically present, like fruit smoothies (Brunstrom, Brown, Hinton, Rogers, & Fay, 2011), sorbets (Wilkinson & Brunstrom, 2009), and dairy smoothies (Fay, Hinton, Rogers & Brunstrom et al., 2011) can be compared with meal shown on pictures successfully. A likely explanation of the absence of changes in expected satiety is that the satiety beliefs are still based on prior experiences with different food (texture) stimuli, therewith limiting the extent to which learning will result in behavioural changes (Hogenkamp, Brunstrom, et al., 2011; O’Sullivan et al., 2010). Both test foods were served as a breakfast replacer and consumed with a spoon, but the liquid food may still have been evaluated as a ‘drink’ rather than as a ‘food’. The different consumption context of liquids and (semi-)solid foods, i.e. quenching thirst vs. decreasing hunger feelings, contribute to the differences in intake and in expectations between the liquid and semisolids. The relatively low number of exposures to the test foods may also have limited behavioural changes. We do, however, not expect that prolonging the consumption period would have resulted in pronounced behavioural changes (Hogenkamp et al., 2010; Mars et al., 2009), such as expectations about the foods’ satiety effects. In addition, it was observed that e.g. viscosity tended to affect the time to lunch (behavioural satiety outcome) when consuming low-viscosity or high-viscosity rice breakfasts for 5 days
424
P.S. Hogenkamp et al. / Appetite 59 (2012) 419–424
in a row (Russell & Delahunty, 2004), although intake at lunch was not reported. It would therefore be more interesting to investigate if the changes in appetite scores can be replicated and if this will result in changes in subsequent intake, rather than prolonging the consumption period. In addition, choice of the test foods, regarding issues of novelty or nutrient content, may be of greater methodological importance to observe changes in behaviour than the number of trials included in the design (Yeomans, 2012). We hypothesised that expected satiety plays a role in decisions on portion size (Brunstrom & Rogers, 2009), but the increase in ad libitum intake was not preceded by profound changes in expected satiety. This suggests that this role may be only small or easily overruled by food attributes, especially when one single food is consumed during the eating occasion. Results from previous studies show that, in the view of texture attributes, perceived volume (Brunstrom, Collingwood, & Rogers, 2010) and eating rate are also of great importance in ad libitum intake (Andrade, Greene, & Melanson, 2008; Martin et al., 2007; Zijlstra et al., 2008). Ad libitum intake increased after repeated consumption compared to baseline. This could be explained by the increase in the eating rate over repeated consumption (Andrade et al., 2008; Martin et al., 2007; Smit, Kemsley, Tapp, & Henry, 2011). Another explanation may be that participants habituated to consumption of the test foods, and adjusted ad libitum intake to the fixed volume they had been exposed to during the week. We observed that several participants struggled to finish the large volume of especially the semi-solid foods on day 2, but that they did not encounter this problem on day 4. It would be interesting to investigate if people can get used to portions larger or smaller than their self-selected ideal portion sizes within several days. This may contribute to limiting the ‘passive’ overconsumption, in which intake is higher than required to feel satiated (Prentice & Jebb, 2003). We conclude that the changes in hunger and fullness indicated that the fixed volume of liquid and solid food was perceived to be equally satiating after repeated consumption, but this did not result in the anticipated changes. Ad libitum intake increased over repeated consumption in the direction of the fixed portion size. Intake of the liquid foods, however, remained higher compared with the semi-solid foods. Texture remained important in ad libitum intake and expected satiety. These results show that decisions on meal size do not easily change, and that cognitive cues based on the food’s sensory attributes remain of great importance in short-term intake. Sensory attributes of a food, however, are not always congruent to its energy content. This may help to better understanding of both short and longer term regulation of energy intake. Liquid foods are not expected to be filling, and may therefore promote overconsumption, especially in single-item meals (e.g. drinks, shakes). In addition, habituation to the satiating effects of foods and a possible mismatch with ‘anticipated’ and ‘experienced’ satiety may e.g. explain why it is hard to keep up with e.g. consuming low-caloric weight loss products for prolonged periods of time.
References Almiron-Roig, E., Chen, Y., & Drewnowski, A. (2003). Liquid calories and the failure of satiety. How good is the evidence? Obesity Reviews, 4, 201–212. Andrade, A. M., Greene, G. W., & Melanson, K. J. (2008). Eating slowly led to decreases in energy intake within meals in healthy women. Journal of the American Dietetic Association, 108, 1186–1191. Blundell, J., Rogers, P. J., & Hill, A. J. (1987). Evaluating the satiating power of foods. Implications for acceptance and consumption. In J. Solms & D. A. Booth (Eds.), Food Acceptance and Nutrition (pp. 205–219). London: Academic Press. Booth, D. A., & Davis, J. D. (1973). Gastrointestinal factors in the acquisition of oral sensory control of satiation. Physiology & Behavior, 11, 23–29. Booth, D. A., & Grinker, J. A. (1993). Learned control of meal size in spontaneously obese and nonobese bonnet macaque monkeys. Physiology & Behavior, 53, 51–57.
Brunstrom, Jeffrey M. (2007). Associative learning and the control of human dietary behavior. Appetite, 49, 268–271. Brunstrom, J. M., Brown, S., Hinton, E. C., Rogers, P. J., & Fay, S. H. (2011). ‘Expected satiety’ changes hunger and fullness in the inter-meal interval. Appetite, 56, 310–315. Brunstrom, J. M., Collingwood, J., & Rogers, P. J. (2010). Perceived volume, expected satiation, and the energy content of self-selected meals. Appetite, 55, 25–29. Brunstrom, J. M., & Rogers, P. J. (2009). How many calories are on our plate? Expected fullness, not liking, determines meal-size selection. Obesity, 17, 1884–1890. Brunstrom, J. M., Shakeshaft, N. G., & Scott-Samuel, N. E. (2008). Measuring ‘expected satiety’ in a range of common foods using a method of constant stimuli. Appetite, 51, 604–614. Cardello, A. V., & Sawyer, F. M. (1992). Effects of disconfirmed consumer expectations on food acceptability. Journal of Sensory Studies, 7, 253–277. Herman, C. P., & Polivy, J. (2005). Normative influences on food intake. Physiology & Behavior, 86, 762–772. Hogenkamp, P. S., Brunstrom, J. M., Stafleu, A., Mars, M., & de Graaf, C. (2011). Expected satiation after repeated consumption of low- or high-energy-dense soup. British Journal of Nutrition. Hogenkamp, P. S., Mars, M., Stafleu, A., & de Graaf, C. (2010). Intake during repeated exposure to low- and high-energy-dense yogurts by different means of consumption. American Journal of Clinical Nutrition, 91, 841–847. Hogenkamp, P. S., Stafleu, A., Mars, M., Brunstrom, J. M., & de Graaf, C. (2011). Texture, not flavor, determines expected satiation of dairy products. Appetite, 57, 635–641. Hogenkamp, P. S., Stafleu, A., Mars, M., & de Graaf, C. (2011). Learning about the energy density of liquid and semi-solid foods. International Journal of Obesity. Hulshof, T., De Graaf, C., & Weststrate, J. A. (1993). The effects of preloads varying in physical state and fat content on satiety and energy intake. Appetite, 21, 273–286. Mars, M., Hogenkamp, P. S., Gosses, A. M., Stafleu, A., & De Graaf, C. (2009). Effect of viscosity on learned satiation. Physiology & Behavior, 98, 60–66. Martin, C. K., Anton, S. D., Walden, H., Arnett, C., Greenway, F. L., & Williamson, D. A. (2007). Slower eating rate reduces the food intake of men, but not women. Implications for behavioral weight control. Behaviour Research and Therapy, 45, 2349–2359. Mattes, R. D., & Rothacker, D. (2001). Beverage viscosity is inversely related to postprandial hunger in humans. Physiology & Behavior, 74, 551–557. O’Sullivan, H. L., Alexander, E., Ferriday, D., & Brunstrom, J. M. (2010). Effects of repeated exposure on liking for a reduced-energy-dense food. American Journal of Clinical Nutrition, 91, 1584–1589. Prentice, A. M., & Jebb, S. A. (2003). Fast foods, energy density and obesity. A possible mechanistic link. Obesity Reviews, 4, 187–194. Russell, K., & Delahunty, C. (2004). The effect of viscosity and volume on pleasantness and satiating power of rice milk. Food Quality and Preference, 15, 743–750. Smit, H. J., Kemsley, E. K., Tapp, H. S., & Henry, C. J. K. (2011). Does prolonged chewing reduce food intake? Fletcherism revisited. Appetite, 57, 295–298. Sorensen, L. B., Moller, P., Flint, A., Martens, M., & Raben, A. (2003). Effect of sensory perception of foods on appetite and food intake. A review of studies on humans. International Journal of Obesity and Related Metabolic Disorders, 27, 1152–1166. Stubbs, R. J., Hughes, D. A., Johnstone, A. M., Rowley, E., Reid, C., Elia, M., Stratton, R., Delargy, H., King, N., & Blundell, J. E. (2000). The use of visual analogue scales to assess motivation to eat in human subjects. A review of their reliability and validity with an evaluation of new hand-held computerized systems for temporal tracking of appetite ratings. British Journal of Nutrition, 84, 405–415. Tepper, B. J., Mattes, R. D., & Farkas, B. K. (1991). Learned flavor cues influence food intake in humans. Journal of Sensory Studies, 6, 89–100. Wansink, B. (2004). Environmental factors that increase the food intake and consumption volume of unknowing consumers. Annual Review of Nutrition, 24, 455–479. Wansink, B., Ittersum, K., & Painter, J. E. (2004). How diet and health labels influence taste and satiation. Journal of Food Science, 69, S340–S346. Warwick, Z. S., & Schiffman, S. S. (1991). Flavor-calorie relationships. Effect on weight gain in rats. Physiology & Behavior, 50, 465–470. WHO. (1985). Estimates of energy and protein requirements of adults and children. In Energy and protein requirements. Report of a Joint FAO/WHO/UNU Expert consultation (2 ed., pp. 71–112). Geneva: World Health Organization. Wilkinson, L. L., & Brunstrom, J. M. (2009). Conditioning ‘fullness expectations’ in a novel dessert. Appetite, 52, 780–783. Yeomans, Martin R. (2012). Flavour–nutrient learning in humans. An elusive phenomenon? Physiology & Behavior, 106, 345–355. Yeomans, M. R., Chambers, L., Blumenthal, H., & Blake, A. (2008). The role of expectancy in sensory and hedonic evaluation. The case of smoked salmon icecream. Food Quality and Preference, 19, 565–573. Yeomans, M. R., Weinberg, L., & James, S. (2005). Effects of palatability and learned satiety on energy density influences on breakfast intake in humans. Physiology & Behavior, 86, 487–499. Zandstra, E. H., Stubenitsky, K., De Graaf, C., & Mela, D. J. (2002). Effects of learned flavour cues on short-term regulation of food intake in a realistic setting. Physiology & Behavior, 75, 83–90. Zijlstra, N., Mars, M., de Wijk, R. A., Westerterp-Plantenga, M. S., & de Graaf, C. (2008). The effect of viscosity on ad libitum food intake. International Journal of Obesity, 32, 676–683.