The Effect of a Low-Energy Food Foam on Appetite Measures During a 1-Day Reduced-Energy Meal Plan

ORIGINAL ARTICLE The effect of a low-energy food foam on appetite measures during a 1-day reduced-energy meal plan

HPF Peters, WP Koppenol, EAH Schuring, SL Abrahamse and DJ Mela

BACKGROUND/OBJECTIVES: We have previously shown that 500 ml of a foamed drink (‘foam’) significantly improved appetite versus a non-foamed control. The objectives of this research were to assess the effect of smaller volumes of foams on appetite, and the potential benefits of foam ingestion and its timing on appetite measures in a reduced-energy context. SUBJECTS/METHODS: Two randomized, parallel design studies (pre- and main study) were conducted using healthy adult subjects. Pre-study: 133 subjects (age 18–50 years, body mass index (BMI) 20–32 kg m − 2) each consumed either 10, 25, 50, 100, 150 or 250 ml foamed meal replacer (~0.2 kcal ml − 1), 150 min after a fixed breakfast. Main study: four groups of subjects (n = 134; age 18–60 years, 2 BMI 22.5–35.0 kg m − ) consumed 200 ml/22 kcal foam (based on pre-study results) immediately after main meals (M), after snacks (S), in-between snacks and main meals (I) or not at all (control, C) within 1 day of a reduced-energy meal plan consisting of three main meals and three snacks. Measurements included self-reported appetite (six scales, reported as area under the curve (AUC)) and (main study only) end-of-day appetite questionnaire. RESULTS: Pre-study: the strongest effect on appetite was produced by 250 ml (consistent across scales), whereas 150 ml showed more pronounced effects than 100 and 50 ml in most scales. Volumes 10 and 25 ml had no effects on any scale. Main study: 200 ml foam reduced appetite AUC substantially in all treatments, particularly M (for example, hunger AUC reduced by 35% (P o0.001), 28% (P o0.05) and 20% (P = 0.11) for M, S and I, respectively versus C). A strong reduction in ‘appetite for a snack’ wasseenforalltimings(all P o0.05). The end-of-day appetite ratings confirmed these findings. CONCLUSIONS: Modest amounts of a low-energy foam can reduce appetite measures during a 1-day reduced-energy meal plan. International Journal of Obesity (2015) 39, 361–367; doi:10.1038/ijo.2014.151

INTRODUCTION reduction in hunger might come from foams taken between Many different dietary approaches can be effective for weight loss snacks and meals. in the context of an overall diet plan.1–4 However, failure to adhere For eventual use of foams in a realistic consumer weight control to these diets is a primary factor in the failure to achieve or context, questions arise as to whether meaningful effects can be maintain weight loss.2–5 Perceived hunger has been shown to be a achieved with relatively low foam volumes, and when these are significant predictor of the failure to lose weight in clinical trials.6 consumed with or between meals (for example, as a snack). The Delaying the post-ingestion return of hunger can potentially objective of the pre-study here was therefore to generate a – fi increase consumer satisfaction with weight control programs and dose response pro le for single-exposure effects of foams on reduced-energy food products and facilitate sustained compliance appetite ratings, focused on identifying lower volume limits for with a reduced-energy diet.7 increasing satiety. The general objectives of the main study were One proposed route toward enhancing satiety and possibly to test the effect of a relatively small amount of foam, selected on affecting satisfaction with a diet plan is the use of food foams. the basis of the pre-study results, on different appetite measures Osterholt et al.8 and Rolls et al.9 showed that entrapped air in corn when consumed with different timings during a 1-day reduced- energy meal plan. snacks and milk shakes, respectively, could enhance acute satiation and satiety. We have recently shown that aeration of a liquid meal replacer into a stable foam can dramatically heighten PRE-STUDY and prolong satiety effects compared with the corresponding non-foamed, equicaloric product.10 In our initial study large foam Subjects and Methods volumes (500 and 1000 ml) were given as breakfast, and it is Subjects. Male and female (n = 144) subjects were recruited from therefore important to establish the dose–response profile and a local area near the study site. Selection criteria were: age 18–50 − 2 effects in other use contexts. years, body mass index (BMI) 21–35 kg m , apparent good health A recent review11 suggests that increasing food volumes and (assessed by questionnaire) and no use of medicines that are thus satiety with water or gas may be promising for weight judged likely to influence the study results. In order to focus on management, although the timing of ingestion might be potential responsiveness to physiological effects, only normal and important. We hypothesized that, because meals and snacks low-restraint eaters were included.12 Any subject with a tendency already deliver a satiety effect and because the ingested foods toward diagnosable eating disorders was also excluded.13 could possibly destroy the foam in the stomach, the greatest Potential volunteers were trained on completion of visual analog

Unilever Research & Development, Vlaardingen, The Netherlands. Correspondence: Dr HPF Peters, Discover, Unilever Research & Development Vlaardingen, Olivier van Noortlaan 120, PO Box 114, 3130 AC, Vlaardingen, The Netherlands. E-mail: [email protected] Received 19 March 2014; revised 1 July 2014; accepted 27 July 2014; accepted article preview online 1 August 2014; advance online publication, 2 September 2014 Foams and appetite HPF Peters et al 362 scales (VAS) for appetite ratings, and familiarized with the test calculated using the Weibull modeling technique, which allows product and study design. The subjects were randomized into for quantitative estimation of the duration of appetite responses.17 groups of 24 per treatment, matched for gender mix, age and For this explorative pre-study, the sample size consideration body weight. was based on previous studies where TTRTB was determined.17 Per treatment 24 measurements were required to get an estimate Study design. The study used a random allocation parallel design, for TTRTB with a 95% confidence interval of ± 20 min, which was with the treatments balanced across test days. Instructions to needed to set up a dose–response profile. For six dose levels this volunteers on background diet, physical activity and dietary meant 144 subjects in total. restrictions before and during the test day were identical to 14 All parameters were analyzed using analysis of variance with previous studies (for example, Peters et al. ). subjects as blocks and treatment as the factor. This pre-study was After an overnight fast, each subject group was given a 250-kcal designed only to generate a dose–response profile for satiety breakfast at 0800 hours, followed by single exposure to 10, 25, 50, effects and to identify practical lower volume limits for the main 100, 150 or 250 ml foam as a mid-morning snack (at 1030 hours). study, and not to test for statistically significant differences Self-reported appetite ratings (six scales) were collected at six time between treatments. points before consumption of the test product and at seven time points ending at 180 min afterwards (details below). The pre-study was carried out at Leatherhead Food Interna- Results tional, Leatherhead, UK. Ethical approval was obtained from Kent Subjects. Ten subjects dropped out during or after the first Research Ethics Committee, UK. intervention day. Data from one subject were removed during blind review of the study data because she had a sore throat on Test foods. The breakfast consisted of a 250-kcal bar (45 g the day of the study. Dropouts were not replaced, therefore all Jordans Absolute Nut Luxury Bar, The Jordans & Ryvita Company, analyses were performed with 133 subjects (42 males), with a Biggleswade, UK). – • mean age of 35.8 (range 18 60) years and BMI of 24.8 (range The test products consisted of Slim Fast high-protein chocolate – − 2 shakes (0.59 kcal ml − 1, Unilever, Englewood Cliffs, NJ, USA), 21.0 34.6) kg m . aerated on site with N2O within a dispenser (iSi cream chargers and iSi Gourmet Whipper, Carrick Co. Ltd, Bury, UK). This aeration Appetite. The foams induced clear effects on eating motivational produced liquid food foams (hereafter called ‘foam’) containing ratings. An example of the effects of the six foam volumes is shown for fullness (Figure 1) and hunger (Figure 2). ~ 3 ml liquid shake, 7 ml N2O and 2 kcal per 10 ml. The unopened liquid shakes were stored at 5 °C for at least 24 h The effects on the other line scales are roughly similar to these (Supplementary Figures S1–S4). The results indicate a general before each test day, and the N2O chargers and dispensers were stored at room temperature. All test products were presented in a dose–response pattern, although rather inconsistent across the glass beaker. Foams were consumed with a 10-ml spoon and the different scales used. On all line scales, the strongest effects are subjects were instructed to eat all of the foam within 10 min. All consistently found for 250 ml and no effect for 10 or 25 ml, with test products were prepared fresh on test days according to a intermediate effects for 50, 100 and 150 ml. In four out of six line standard operating procedure. scales the effect of 150 ml on peak and duration is somewhat more pronounced than 50 and 100 ml. TTRTB and AUC were Measures. Self-report appetite ratings were collected at time consistent with this outcome, for example, for ‘appetite for a meal’ points of − 155, − 120, − 90, − 60, − 30, − 5, 15, 30, 60, 90, 120, 150 TTRTB were 96, 69, 45 and 41 min for the 250, 150, 100 and 50 ml, and 180 min (with test product consumption at 1030 hours respectively (Supplementary Tables S1 and S2). For 25 and 10 ml regarded as 0 min). Ratings were made on a handheld device these estimates could not be calculated, as the curves did not (iPAQ, Hewlett Packard, Palo Alto, CA, USA) by means of a mark on 15 cross the baseline. a 64-mm line using an Electronic VAS (EVAS ) anchored at the The appetite curves and TTRTB data in this pre-study suggest low end with the most negative or lowest intensity feelings (for that foam volumes as low as 50 ml consistently decrease hunger. example, not at all), and with opposing terms at the high end (for However, higher volumes might be needed to generate mean- example, very high), as described by Flint et al.16 The scale items ingful hunger reductions in other study circumstances, for were ‘appetite for a meal’, ‘appetite for a snack’, ‘hunger’, ‘how much do you want to eat’, ‘satiety’ and ‘fullness’. example, during a 1-day reduced-energy meal plan. As the Immediately before consumption of the breakfast and at consumption of meals and snacks over the day may disguise and/ time points − 155, − 5, 90 and 180 min, gastro intestinal (GI) or negatively impact the hunger effects of foams (vs when foams disturbances (nausea, heartburn, belching, abdominal bloating, are consumed in isolation), we elected to use 200 ml in the main flatulence and urge to defecate) and general bodily symptoms study, reflecting a modest foam volume that had reliable and (headache, dizziness and fatigue) were scored electronically using robust effects (on its own). Higher volumes were not regarded as EVAS on 4-point scales (0 = none, 1 = mild, 2 = moderate, and consumer- and commercially relevant proposals. 3 = severe). The questions were phrased to refer to symptoms that the subjects might possibly have experienced during the Symptoms. GI symptoms and headache did not differ between preceding hour. Consumption of only non-caloric drinks (max- treatments and most were rated ‘mild’ and already present before imum 150 ml water, coffee/tea without milk/sugar) was allowed the foam was given. In those cases where there was a slight each hour directly after recording EVAS scores (at − 60, 60, 120 increase in occurrence after consumption, this was independent of and 180 min). volume as it occurred both at higher and lower volumes. Statistical analysis. The 64- mm appetite EVAS scores were transformed and expressed on a 100 mm basis. MAIN STUDY To establish the effect of foam volume on appetite ratings, two different measures were used: incremental cumulative area under Subjects and methods the curve (AUC) and ‘time to return to baseline’ (TTRTB). AUC was Subjects. The number, characteristics and selection of subjects calculated using the trapezoid rule with the values at t = 0 min were similar to the pre-study, with a slightly broader age and BMI (just before consumption of foam) as covariate. TTRTB was range (18–60 years, 22.5–35.0 kg m − 2). Subjects were randomized

Figure 1. The effect of different foam volumes on fullness ratings (LSmeans, n = 133, in mm; s.e. is shown for 250 ml and 10 ml).

Figure 2. The effect of different foam volumes on hunger ratings (LSmeans, n = 133, in mm; s.e. is shown for 250 ml and 10 ml). into four groups of 36 subjects per treatment, with groups identical to the pre-study and other previous studies (for example matched for gender mix, age and body weight. Peters et al.14). The study was carried out at Leatherhead Food International, Study design. The study used a random allocation, parallel Leatherhead, UK. Ethical approval was obtained from Kent design, with the three experimental and one control treatments Research Ethics Committee, UK. balanced across test days. On their single test day, volunteers followed 1 day of a reduced-energy meal plan (1200 kcal per day Meal plan and test food ﬁ for women and 1500 kcal per day for men), consisting of xed Meal plan: The study meals were based on the FitnessQuest diet snacks and meals: breakfast at 0800 hours, mid-morning snack at plan.18 Details of the meals are given in Supplementary Table S3. 1000 hours, lunch at 1230 hours, mid-afternoon snack at The controls received the entire 1200/1500 kcal. By applying small 1400 hours, dinner at 1700 hours and evening snack at 1830 hours. adjustments to the main meals, subjects in the treatment groups The volunteers went home at 2100. The three experimental received 66 kcal less from the meals throughout the day to treatments consisted of the meal plan plus foams (three in total) compensate for the 66 kcal (three times 22 kcal) in the foams. As a consumed at different timings: (1) immediately after the main consequence all groups consumed the same amount of energy meals, (2) immediately after the snacks or (3) in between snacks and main meal. Control subjects followed the identical meal plan during the test day. but with no foam products. Appetite scales and symptoms were collected throughout the Test food. Participants were given the 200-ml foam in a glass and day at regular intervals and the overall appetite throughout the instructed to eat all of the foam with a spoon within 10 min. day was tested by an end-of-day questionnaire at 2100 hours Although the exact viscosity could not be measured reliably, the (schematic overview in Supplementary Figure S5). viscosity of the foams was evidently higher than of liquids like Instructions to volunteers on background diet, physical activity chocolate milk. The foam was thick but pourable, which would and dietary restrictions before and during the test day were allow (most) subjects to drink the foam. However, previous

© 2015 Macmillan Publishers Limited International Journal of Obesity (2015) 361 – 367 Foams and appetite HPF Peters et al 364 experience showed that not all subjects were able to drink the clearly visible effects on the patterns of appetite ratings relative to foam, so it was decided that all subjects would use a spoon. the control. An example of the effect of the foam products is Depending on the test group, the foam was provided to shown for hunger (Figure 3) and appetite for a snack (Figure 4). subjects either immediately after the main meals, immediately The effects on the other line scales are shown in the after the snacks, between the main meals and snacks or not at Supplementary Figures S6–S9. On the basis of these data, AUC all. The test products consisted of foams contained within and differences in AUC vs control were calculated for the entire pressurized canisters (produced by Gay Lea Foods Cooperative day period and shown in Table 1. Foams consumed immediately Ltd, Toronto, ON, Canada), and dispensed according to a after the main meals reduced appetite ratings very robustly (for standard procedure. The liquid part of the foam differed from example, hunger AUC reduced by ~ 35%, P o0.01; all other scales that used for the pre-study only in the type of chocolate flavor. P o0.05 to Po0.001). Foams consumed in between snacks and The ratio of liquid:gas used in the foam and therefore energy main meal also reduced appetite, but with a smaller effect size (for density was also somewhat lower than that used in the pre- example, hunger AUC reduced by 28%, P o0.05; two other scales o o study: ~ 2 ml liquid shake, 8 ml N2O and1kcalper10ml(41ml P 0.05, two other scales P 0.10). Foams consumed immedi- liquid shake, 159 ml N2O, 22 kcal per 200 ml). Compliance and ately after snacks also reduced appetite throughout the day, but mean foam consumption was calculated by measuring weight had the smallest effect (for example, hunger AUC reduced by 20%, before and after consumption. P = 0.11; one other scale P o0.05). A strong statistically significant reduction in ‘appetite for a snack’ was seen for all treatments Measures. Data were collected using EVAS by methods identical (varying from 22 to 37% decrease). to the pre-study except as described below. Study parameters were appetite ratings, GI symptoms and overall end-of-day End-of-day appetite and satisfaction. The effects of the foam appetite measures. Appetite and GI symptoms were rated snack on end-of-day ratings are shown in Table 2. Three foam immediately before consumption of the breakfast (0800 hours) servings appeared efficacious in reducing reported (retrospective) and at regular intervals thereafter (see Supplementary Figure S5). feelings of hunger and desire to eat another meal, snack or other Overall appetite and one question on satisfaction with the meal foods in that day, while increasing reported fullness feelings. The plan was rated on an end-of-day questionnaire with six scale items magnitude of these effects varied somewhat depending on the as described in Table 2. scale measure and test condition. The addition of foams did not, however, make it ‘less difficult to follow the meal plan’ for Statistical analysis. A sample size consideration indicated that 36 that day. subjects per group were required in this parallel study based on a subject variance of 85 for appetite AUC as determined in former Physical symptoms. Reported headache, heartburn, nausea, cross-over studies, an alpha of 0.05 (corrected for Dunnett’s stomach discomfort and bloating were in most cases rated as difference testing) and a power of 0.8 for a difference of 10 points ‘mild’ and none differed significantly between treatments. ‘ ’ AUC, and taking into account dropouts. Data were analyzed using However, belching (although also mostly rated mild ) was clearly Analysis of covariance with baseline values ( = questionnaires greater for foam treatments than the control, although dependent completed after arrival at the lab, 0745 hours) as covariates and on timing of the day (data not shown). Belching was more groups as factor. Dunnett’s difference test was used to make the apparent with foams consumed in between snack and main meal pre-planned comparisons of treatments to the control. The or immediately after the snack than when consumed immediately different foam treatments were not statistically compared with after the meal. each other, as the study was not planned or powered for these additional multiple comparisons. DISCUSSION The primary study parameter was the AUC of the appetite EVAS scores for the entire day (0800 till 2100 hours) and for smaller time The present results indicate that a foamed version of a liquid periods. Secondary study parameters were overall appetite and meal replacer, consumed as a snack, affects eating motivations in fl a roughly dose–response pattern in the range 10–250 ml satisfaction as re ected by the end-of-day questionnaire at – 2100 hours. AUC is described by LSmeans and s.e. LSmeans for (2 50 kcal), and robust effects are apparent at volumes much lower than previously tested. We have also clearly established total AUC and for smaller time periods were calculated using fi ANCOVA with baseline measurement per subject as covariate. that 200 ml (22 kcal) servings of this foam product are ef cacious for reducing hunger when consumed three times during 1 day of a hypocaloric meal plan. These effects appeared particularly RESULTS robust when the foams were consumed immediately after the main meals, although the different foam treatments were not Subjects statistically compared with each other. Effects were still apparent A total of 134 subjects (55 males) fully complied with the study though with smaller impact and less consistency among the six protocol, mean age 38.2 (range 18–60) years and BMI: 25.8 (range scales when the foam snacks were consumed immediately after − 21.7–34.8) kg m 2. the snack or when consumed in between snacks and the main meals. A strong reduction in ‘appetite for a snack’ was seen for all Compliance. Volunteers fully complied with the required foam foam timings. consumption. Compliance with the required food consumption We and others have shown that entrapped air can effect was also very high. Post-meal plate weights indicated 11 out of 79 an acute enhancement of satiety following single preloads of women consumed between 1 and 19 kcal less (mean 8 kcal) than milk shakes,9 bread rolls,19 ready-to-eat cereals20 and meal the prescribed 1200 kcal, whereas 1 woman consumed 97 kcal replacers.10 We now show that modest volumes of low-energy less. Seven out of 55 men consumed between 3 and 38 kcal less foams can be effective for reducing reported hunger when (mean 19 kcal) than the prescribed 1500 kcal. Study results with consumed three times during 1 day of a reduced-energy and without these ‘non-compliant’ subjects resulted in the same meal plan. conclusions. In the current study, foams were used to reduce hunger during a hypocaloric meal plan. As perceived hunger has been shown to Appetite ratings. The consumption of 200 ml foam in all test be a significant predictor of failure to lose weight in clinical trials,6 conditions, particularly immediately after the main meals, induced foams can potentially increase consumer tolerance and

Figure 3. The effect of repeated consumption of 200 ml foam on hunger ratings (LSmeans in mm) when the foam was provided during a 1-day reduced-energy meal plan either immediately after the main meals (n = 36), immediately after the snacks (n = 32), between the snacks and main meals (n = 32) or not at all (control, n = 34).

Figure 4. The effect of repeated consumption of 200 ml foam snacks on ratings of ‘appetite for a snack’ (LSmeans in mm) when the foam was provided during a meal plan either immediately after the main meals (n = 36), immediately after the snacks (n = 32), between the snacks and main meals (n = 32) or not at all (control, n = 34). satisfaction with weight control programs and reduced-energy (deﬁned as kcal g − 1), it may also affect food intake via the food products, and encourage long-term compliance with a volume increase. Because the daily intake of foods may be reduced-energy regime. However, it has yet to be proven whether driven in part by the amount (weight or volume)23,26–28 the hunger-suppressing effects of these foams persist when tested consumed, lower-energy content of the daily diet while longer-term and whether these foams can reduce dysphoria or maintaining the total volume might lead to weight loss. Indeed, improve compliance and weight loss when used as an adjunct to studies over a 12-week29 and a 12-month30 diet intervention sustained weight control programs. period showed a signiﬁcant effect of additional water consump- We have shown that foams reduce hunger, but foams might tion on weight loss. Whether this is also the case for gas remains also affect weight loss via their effect on energy intake and to be determined. energy density (energy displacement) analogous to the use of The results of the present study may be attributable to effect of water. Many (for example, Rolls et al.21,22 and Bell et al.23), but the foam on the gastric handling of the food or to the effect of not all (for example, Lappalainen et al. and Gray et al.24,25) foam itself on gastric volume. Stomach distension has been shown studies have shown that the incorporation of water into foods to have an inverse relationship with appetite,31–35 a relationship decreased energy intake via its effect on energy density or hypothesized to be based on the activation of gastric mechan- volume. Although aeration doesnotaffectenergydensity oreceptors. Gastric distension signaling has been shown to be fast

© 2015 Macmillan Publishers Limited International Journal of Obesity (2015) 361 – 367 Foams and appetite HPF Peters et al 366 acting and relatively short-lived,34,35 with increased satiety satiety are inconsistent (for example, Bolhuis et al. and Martin observed only during and not after distension.34,35 It has also et al.40,41). It is also not known to what extent these effects are been shown that significant changes in satiety scores occur only at specifically linked to the effect of aeration. Although obviously the higher intragastric volumes, for example, 400 or 600 ml.35 As such textural structure of the foam differs from liquids (for example, one might not expect that the consumption of 200 ml foam by somewhat higher viscosity), we have no data available to itself would be sufficient for satiety, yet such effects were seen at speculate on the role of oral processing. We would expect that even smaller volumes in the pre-study. It is possible that the effect any such effects would be relatively small and transient (certainly of foams on satiety was most apparent when consumed as compared to the volume effects), but this would need to be immediately after the main meals because foams were being confirmed experimentally. added to the highest pre-foam level of gastric fill. Our hypothesis An effect of timing was expected based on studies using water. that the meals and snacks might destroy the foam in the stomach For example, when 400 ml water was consumed with breakfast, 24 or that the satiety increase due to the meals and snacks would satiety increased during the meal, but not after the meal. In override the added satiety effect of the foams appeared not to be other studies, water with a meal did not affect satiety, but water fi incorporated in the meal did,22 and giving water before a meal con rmed. 42–44 Some part of the satiety effects of foams, next to gastric also increased satiety. Here we compared the effect of foam distension, might be attributed to cognitive effects. Brunstrom after meals and after snacks. It remains to be determined whether fi and coauthors,36,37 have shown that the amount of ‘expected’ foams would also be ef cacious when consumed before main satiety and satiation is related to the perceived volume of a meals or snacks. meal, although others also have hypothesized that volume Clear satiety effects were observed when foams were consumed might affect the beliefs of the satiating capacity of foods.22,38 periodically throughout a day-long period. Also, at the end of the Some part of the appetite effects of foams might also be day, retrospective ratings of hunger and appetite indicated the efficacy of the foams over the day. However, subjects did not attributed to sensory exposure and ingestive behavior during fi ‘ ’ consumption.11 Oral transit time, sip size and eating frequency report a change in perceived dif culty to follow the diet plan . have been linked to appetite, but these studies refer mostly to the This question may have been somewhat irrelevant as the test was carried out only over 1 day, which in any case would have been effects on satiation (within-meal) or ad libitum food intake (for relatively easy for the subjects. Proper testing of this benefit would example, Zijlstra et al. and Bolhuis et al.39,40), whereas effects on need to be in the context of an extended period of a reduced- energy regime or perhaps using a diet plan with even lower energy content. Table 1. AUC for the whole day (LSmeans, n = 134, in mm∙min, ± s.e.) Foams caused an increase in bloating and especially belching, ‘ ’ for all appetite scales when the foam was provided during a meal plan although these symptoms were mostly rated mild and were less either immediately after the main meals (n = 36), immediately after the prevalent when foams were consumed immediately after the snacks (n = 32), between the snacks and main meals (n = 32) or not at main meal. The relatively high incidence of belching, even though ‘ ’ all (control, n = 34) mostly rated mild , might raise some concerns for consumer acceptance. It must, however, be noted that the study was not Treatment s.e. specifically set up to identify whether, when and to what extent foams would cause GI discomfort. It is also uncertain whether this Control Meals In-between Snacks belching, and to a lesser extent bloating, actually bothered the +foam foam +foam volunteers. In conclusion, we have shown that relatively modest amounts of Hunger 40 26** 30* 32 2.8 a low-energy foam consistently reduced hunger feelings and Fullness 48 62** 57 55 2.8 appetite throughout the day in a reduced-energy context. Longer- Appetite for a meal 36 25* 27* 30 2.9 term studies are needed to show the efficacy of these foams over Appetite for a snack 46 29*** 36* 32** 3.2 Prospective food 43 31** 35 36 2.8 a longer time period and the potential of foams to affect weight consumption control programs, either via increased diet acceptance and Satiety 48 62** 55 55 2.9 compliance or via direct effects of foams on energy intake. Further research is also needed to establish whether foam Abbreviations: AUC, area under the curve; LSmeans, least squares means. volumes below 200 ml would also be efficacious in a reduced- *Po0.05; **Po0.01; ***Po0.001. energy context.

Table 2. The effects of the foam products on end-of day ratings (LSmeans, n = 134) as shown by different questions when the foam was provided during 1 day of a diet plan either immediately after the main meals (n = 36), immediately after the snacks (n = 32), between the snacks and main meals (n = 32) or not at all (control, n = 34)

Treatment s.e.

Control Meals+foam In-between foam Snacks+foam

How hungry have you felt today? 54 37* 41 47 4.0 How full have you felt today? 45 56 58 59* 4.2 How strong was your desire to eat another meal today? 50 35* 41 42 4.9 How strong was your desire to eat another snack today? 60 38** 45 36** 5.1 How difﬁcult was it to follow the diet plan today? 26 22 25 28 4.0 How strong was your desire to eat additional foods today? 55 33** 38* 36* 5.1 *Po0.05; **Po0.01.

International Journal of Obesity (2015) 361 – 367 © 2015 Macmillan Publishers Limited Foams and appetite HPF Peters et al 367 CONFLICT OF INTEREST 20 Irvine P, Livingstone B, Welch R. Bulk density of ready-to-eat cereals affects satiety This research was funded by Unilever R&D, Vlaardingen. All authors are employees of and subsequent intakes in young women. Int J Obesity 2007; 31: S154. the funding company. 21 Rolls BJ, Castellanos VH, Halford JC, Kilara A, Panyam D, Pelkman CL et al. Volume of food consumed affects satiety in men. Am J Clin Nutr 1998; 67: 1170–1177. ACKNOWLEDGEMENTS 22 Rolls BJ, Bell EA, Thorwart ML. Water incorporated into a food but not served with a food decreases energy intake in lean women. Am J Clin Nutr 1999; 70:448–455. We thank Sarah Hull and her colleagues at Leatherhead Food International, UK, for 23 Bell EA, Castellanos VH, Pelkman CL, Thorwart ML, Rolls BJ. Energy density of performing the study and Luben Arnaudov, Miranda Slotboom, Johan Schilt and foods affects energy intake in normal-weight women. Am J Clin Nutr 1998; 67: Jeroen Sterken for their help in this study. 412–420. 24 Lappalainen R, Mennen L, Vanweert L, Mykkanen H. Drinking-water with a REFERENCES meal—a simple method of coping with feelings of hunger, satiety and desire to eat. Eur J Clin Nutr 1993; 47: 815–819. 1 Anderson JW, Luan JY, Hoie LH. Structured weight-loss programs: meta-analysis of 25 Gray RW, French SJ, Robinson TM, Yeomans MR. Dissociation of the effects of weight loss at 24 weeks and assessment of effects of intervention intensity. Adv preload volume and energy content on subjective appetite and food intake. 21 – Ther 2004; :61 75. Physiol & Behav 2002; 76:57–64. fi 2 Dansinger ML, Gleason JA, Grif th JL, Selker HP, Schaefer EJ. Comparison of the 26 Duncan KH, Bacon JA, Weinsier RL. The effects of high and low-energy density Atkins, Ornish, Weight watchers, and Zone diets for weight loss and heart disease diets on satiety, energy-intake, and eating time of obese and non-obese subjects. 293 – risk reduction. JAMA 2005; :43 53. Am J Clin Nutr 1983; 37:763–767. fi 3 Heyms eld SB, Harp JB, Reitman ML, Beetsch JW, Schoeller DA, Erondu N et al. 27 Stubbs RJ, Ritz P, Coward WA, Prentice AM. Covert manipulation of the ratio of Why do obese patients not lose more weight when treated with low-calorie diets? dietary-fat to carbohydrate and energy density—effect on food-intake and A mechanistic perspective. Am J Clin Nutr 2007; 85:346–354. energy-balance in free-living men eating ad-libitum. Am J Clin Nutr 1995; 62: 4 Sacks FM, Bray GA, Carey VJ, Smith SR, Ryan DH, Anton SD et al. Comparison of 330–337. weight-loss diets with different compositions of fat, protein, and carbohydrates. 28 Marti-Henneberg C, Capdevila F, Arija V, Perez S, Cuco G, Vizmanos B et al. Energy New Engl J Med 2009; 360: 859–873. density of the diet, food volume and energy intake by age and sex in a healthy 5 Truby H, Baic S, Delooy A, Fox KR, Livingstone MBE, Logan CM et al. Randomised population. Eur J Clin Nutr 1999; 53:421–428. controlled trial of four commercial weight loss programmes in the UK: initial 29 Dennis EA, Dengo AL, Comber DL, Flack KD, Savla J, Davy KP et al. Water findings from the BBC ‘diet trials’. Br Med J 2006; 332: 1309–1311. consumption increases weight loss during a hypocaloric diet intervention in 6 Womble LG, Williamson DA, Greenway FL, Redmann SM. Psychological and middle-aged and older adults. Obesity 2010; 18:300–307. behavioral predictors of weight loss during drug treatment for obesity. Int J 30 Stookey JD, Constant F, Popkin BM, Gardner CD. Drinking water is associated with Obesity 2001; 25: 340–345. weight loss in overweight dieting women independent of diet and activity. 7 Johansson K, Neovius M, Hemmingsson E. Effects of anti-obesity drugs, diet, and Obesity 2008; 16: 2481–2488. exercise on weight-loss maintenance after a very-low-calorie diet or low-calorie 31 French SJ, Read NW. Effect of guar gum on hunger and satiety after meals of diet: a systematic review and meta-analysis of randomized controlled trials. Am J differing fat content: relationship with gastric emptying. Am J Clin Nutr 1994; 59: Clin Nutr 2014; 99:14–23. 87–91. 8 Osterholt KM, Roe DS, Rolls BJ. Incorporation of air into a snack food reduces 32 Khan MI, Read NW, Grundy D. Effect of varying the rate and pattern of gastric energy intake. Appetite 2007; 48:351–358. distension on its sensory perception and motor-activity. Am J Physiol 1993; 264: 9 Rolls BJ, Bell EA, Waugh BA. Increasing the volume of a food by incorporating air – affects satiety in men. Am J Clin Nutr 2000; 72:361–368. G824 G827. 10 Melnikov SM, Stoyanov SD, Kovacs EMR, Arnaudov L, Groot de PWN, Schuring 33 Tack J, Piessevaux H, Coulie B, Caenepeel P, Janssens J. Role of impaired gastric 115 EAH et al. Sustained hunger suppression from stable liquid food foams. Obesity (in accommodation to a meal in functional dyspepsia. Gastroenterol 1998; : – press) 2014 doi:10.1002/oby.20828. 1346 1352. 11 Welch RW. Satiety: have we neglected dietary non-nutrients? Proc Nutr Soc 2011; 34 Geliebter A, Westreich S, Gage D. Gastric distension by balloon and test-meal 48 – 70: 145–154. intake in obese and lean subjects. Am J Clin Nutr 1988; : 592 594. 12 Van Strien T, Frijters JER, Bergers GPA, Defares PB. The Dutch Eating Behavior 35 Oesch S, Ruegg C, Fischer B, Degen L, Beglinger C. Effect of gastric distension Questionnaire (Debq) for assessment of restrained, emotional, and external eating prior to eating on food intake and feelings of satiety in humans. Physiol & Behav 87 – behavior. Int J Eating Dis 1986; 5: 295–315. 2006; :903 910. 13 Morgan JF, Reid F, Lacey JH. The SCOFF questionnaire: assessment of a new 36 Brunstrom JM, Brown S, Hinton EC, Rogers PJ, Fay SH. 'Expected satiety' changes 56 – screening tool for eating disorders. Br Med J 1999; 319: 1467–1468. hunger and fullness in the inter-meal interval. Appetite 2011; : 310 315. 14 Peters HPF, Foltz M, Kovacs EMR, Mela DJ, Schuring EAH, Wiseman SA. The effect 37 Brunstrom JM, Collingwood J, Rogers PJ. Perceived volume, expected satiation, 55 – of protease inhibitors derived from potato formulated in a minidrink on appetite, and the energy content of self-selected meals. Appetite 2010; :25 29. fi food intake and plasma cholecystokinin levels in humans. Int J Obesity 2011; 35: 38 Bell EA, Roe LS, Rolls BJ. Sensory-speci c satiety is affected more by volume than 244–250. by energy content of a liquid food. Physiol & Behav 2003; 78:593–600. 15 Stratton RJ, Stubbs RJ, Hughes D, King N, Blundell JE, Elia M. Comparison of the 39 Zijlstra N, de Wijk RA, Mars M, Stafleu A, De Graaf C. Effect of bite size and oral traditional paper visual analogue scale questionnaire with an Apple Newton processing time of a semisolid food on satiation. Am J Clin Nutr 2009; 90:269–275. electronic appetite rating system (EARS) in free living subjects feeding ad libitum. 40 Bolhuis DP, Lakemond CMM, de Wijk RA, Luning PA, De Graaf C. Both a higher Eur J Clin Nutr 1998; 52: 737–741. number of sips and a longer oral transit time reduce ad libitum intake. Food Qual 16 Flint A, Raben A, Blundell JE, Astrup A. Reproducibility, power and validity of visual Pref 2014; 32:234–240. analogue scares in assessment of appetite sensations in single test meal studies. 41 Martin CK, Anton SD, Walden H, Arnett C, Greenway FL, Williamson DA. Slower Int J Obesity 2000; 24:38–48. eating rate reduces the food intake of men, but not women: implications for 17 Schuring E, Quadt F, Kovacs EM, Meullenet JF, Wiseman S, Mela DJ. A quantitative behavioral weight control. Behav Res Ther 2007; 45: 2349–2359. method for estimating and comparing the duration of human satiety responses: 42 Norton GNM, Anderson AS, Hetherington MM. Volume and variety: relative effects Statistical modeling and application to liquid meal replacers. Appetite 2012; 59: on food intake. Physiol & Behav 2006; 87: 714–722. 601–609. 43 Latner JD, Rosewall JK, Chisholm AM. Food volume effects on intake and appetite 18 Fitnessquest diet plan. Internet Communication https://www.iaff.org/HS/wfir in women with binge-eating disorder and weight-matched controls. Int J Eating esource/Fitness/Nutrition.html (accessed June 2012). Dis 2009; 42:68–75. 19 Irvine P, Welch R, Livingstone M, Robson P, McDwyer M, O'Connor G. Effects of 44 Gray RW, French SJ, Robinson TM, Yeomans MR. Increasing preload volume with varying the volume of bread in a breakfast meal on satiety and subsequent water, reduces rated appetite but not food intake in healthy men even with intakes. Proc Nutr Soc 2005; 64: 36A (abstract). minimum delay between preload and test meal. Nutr Neuroscience 2003; 6:29–37.

Supplementary Information accompanies this paper on International Journal of Obesity website (http://www.nature.com/ijo)