Satiety-relevant sensory qualities enhance the satiating effects of
mixed carbohydrate-protein preloads1–3
Martin R Yeomans and Lucy Chambers
Background: Orosensory cues such as food texture and flavor have
been shown to play a role in satiation, but their role in satiety
remains less clear.
Objective: The objective was to determine whether satiety-relevant
orosensory cues enhance the satiating effects of energy in the con-
text of beverage preloads.
Design: The effects of 6 drink preloads that combined 2 amounts of
energy [high energy (HE): 279 kcal; low energy (LE): 78 kcal] and
3 satiety-relevant sensory contexts [low sensory (LS), medium sen-
sory (MS), and high sensory (HS)] on subsequent appetite and test
meal intake were assessed in 36 healthy nonobese volunteers.
Results: The ability of the preloads to modify appetite 30 min after
consumption depended on both energy content and sensory context
(P-interaction , 0.05), with hunger significantly being lower after
consumption of the HE than after the LE preload in the HS context
(P , 0.001), tending to be lower in the MS context (P = 0.08), but
not different in the LS context. Food intake at lunch was lower after
the HE than after the LE preloads (effect of energy P , 0.001), but
this effect depended on sensory context (P , 0.005). The degree to
which reduced test meal intake compensated for the added energy in
the HE preloads was 88% in the HS context, which was significantly
greater than in the MS (47%) and LS (18%) contexts.
Conclusion: Small changes in the sensory characteristics of drinks
altered the degree to which added energy was satiating, which
implies that nutrients become more satiating when they are pre-
dicted by relevant sensory cues such as thickness and creaminess.
This trial was registered at http://www.controlled-trials.com as
ISRCTN36258511.Am J Clin Nutr 2011;94:1410–7.
The worldwide increase in prevalence of obesity and over-
weight (see references 1 and 2) has resulted in an urgent need to
better understand the nature of satiety (3). Several lines of evi-
dence suggest that the context in which energy is ingested may
be critical in determining whether effective satiety is generated.
In particular, energy consumed as a beverage tends to generate
weak satiety (4, 5), whereas energy consumed in soup generates
much stronger satiety (6, 7). Understanding why satiety should
depend on these subtle differences in physical and sensory qual-
ities should greatly aid the future development of more satiating
products that could be incorporated into weight-management
One possible explanation for differences in satiating effects of
energy between food contexts might be that the characteristics of
the food or drink have different capacities to generate satiety
expectations. People have clear expectations about how satiating
different foods and drinks will be (8). These satiety expectations
could modify the response to nutrient-generated satiety signals
(9). If so, then the presence of satiety-relevant sensory cues
should lead to more effective satiety than would be seen in the
absence of such cues, and the overall objective of the study re-
ported here was to test this idea. Several lines of evidence lead to
satiety, particularly in studies that compare oral with post-oral
nutrient administration (10–12). Second, orosensory experience
modifies satiation, with foods that provide longer sensory ex-
posure being more satiating (13) and drinks that have sensory-
relevant qualities (eg, yogurt flavor and texture) being more
satiating than the same energy in juice form (14). Third, although
as a beverage, beverage energy added as protein was more sa-
tiating than when added as carbohydrate (15), and protein was
most effective when it enhanced perceived creaminess (16).
Finally, although it has been suggested that the effects of drink
thickness (viscosity) on satiety may be due to postingestive
effects, such as modification of gastric emptying (17), dilution of
beverage ingredients in the stomach may minimize such effects
for normal foods (18). Thus, subtle differences in viscosity are
more likely to affect satiety through an orosensory than through
a postingestive mechanism (19).
The current study contrasted the short-term satiating effects of
beverages that differed in sensory quality and energy content. We
predicted that the satiating efficiency of covertly manipulated
energy (added as maltodextrin and whey protein) would increase
as the sensory quality of the beverage became more congruent
with satiety expectations. To add satiety-relevant sensory qual-
ities, the viscosity of the beverage was manipulated since ma-
nipulated viscosity affects satiation (20) and flavor elements that
enhanced creaminess were added (15).
1From the School of Psychology, University of Sussex, Brighton, United
2This study was conducted as part of research grant BB/H004645/1 from
the UK Biotechnology and Biological Sciences Research Council as part of
the Diet and Health Research Industry Club (DRINC) initiative.
3Address correspondence to MR Yeomans, School of Psychology, Uni-
versity of Sussex, Brighton, BN1 9QH, United Kingdom. E-mail: martin@
Received February 17, 2011. Accepted for publication September 27, 2011.
First published online October 26, 2011; doi: 10.3945/ajcn.111.011650.
Am J Clin Nutr 2011;94:1410–7. Printed in USA. ? 2011 American Society for Nutrition
by guest on November 24, 2015
SUBJECTS AND METHODS
effects of 6 drinks that combined 2 amounts of energy content
(HE4or LE) and 3 levels of sensory quality (LS, MS, or HS).
Test meal intake and subjective ratings of appetite were used to
Thirty-six healthy adults participated in the study. Sample size
was determined on the basis of the effect size needed to find
study in which sensory and protein content of a beverage had
been manipulated (15). On the basis of a medium effect size (f =
0.25), power calculations indicated that we needed a sample size
of 24 (power: 0.91), and this sample size was increased further
to allow contrast between sensory conditions. All participants
were staff or students at the University of Sussex, United
Kingdom, who had expressed an interest in participating in
appetite research by completing an online questionnaire. Pro-
spective participants were selected on the basis of data obtained
from this questionnaire and contacted by a recruitment e-mail,
which described the study as “an investigation of the effect of
food on mood.” The contact criteria included the following:
a score of ?7 on the Three-Factor Eating Questionnaire scale of
dietary restraint, smoking <5 cigarettes/d, and a BMI (in kg/m2)
<30. Respondents to this recruitment e-mail were enrolled in the
study if they met the following requirements: not taking pre-
scription medicine, not currently pregnant, without any previous or
current diagnosis of diabetes or an eating disorder, and without
aversions or allergies to the test foods. Written consent was ob-
tained before participation. The test cohort was made up of 18 men
and 18 women (age range: 19–33 y; mean 6 SD: 21.9 6 3.2) who
were not obese (BMI range: 18.9–29.9; mean 6 SD: 22.9 6 2.7)
nor diet restricted (Three-Factor Eating Questionnaire–restraint:
range, 0–7. mean 6 SD, 2.9 6 2.1).
Participants attended the Ingestive Behavior Unit at Sussex
University for 6 sessions, which were restricted to a maximum of
2/wk on nonconsecutivedays. The protocol was identical on each
test day, with only the preload varying; a summary of the daily
protocol can be seen in Figure 1. Participants arrived at
a scheduled time between 0830 and 1000, after having con-
sumed only water from 2300 the night before. They consumed
breakfast and were instructed to return exactly 3 h later for their
lunch session (only water consumption was allowed during this
period). To begin the lunch session, participants were taken to
a windowless air-conditioned testing cubicle where they tasted,
rated, and consumed 300 mL of 1 of the 6 drink preloads. The
preload was evaluated for sweetness, creaminess, pleasantness,
thickness, and familiarity by using SIPM software (University of
Sussex) on a PC computer, which presented randomized VASs
headed with the question “How [target rating] is the drink?” and
end-anchored with “not at all [target rating]” (scored as zero)
and “extremely [target rating]” (scored as 100). To be consistent
with the premise that the study examined effects of food on
mood, participants rated their mood and appetite, initially in the
absence of any food-related cues (the pre-preload baseline rat-
ing) and again immediately after consuming the preload (post-
preload). The mood ratings (alert, anxious, calm, clear-headed,
energetic, happy, headache, nauseous, and tired), which were
presented as 100-point computerized VASs end-anchored with
“not at all [mood]” and “extremely [mood],” were included as
distracters, and data were not analyzed. The critical VAS ratings
of hunger and fullness were embedded in these series of ques-
tions along with a rating of thirst.
A 30-min delay between the preload and lunch was used, as
this has been shown to be an optimal time period for detecting
the testing cubicle, participants completed their third set of mood
and appetite ratings in the absence of any food cues (the prelunch
ratings). Next, 500 g pasta with tomato sauce was served by an
experimenter who explained that the participant could eat as little
or as much as he or she liked. The SIPM prompted the partic-
ipants to taste the pasta and rate its pleasantness, savoriness,
saltiness, and familiarity followed by further ratings of hunger
and fullness to determine the appetizing effects of food pre-
sentation (21). The participants were then instructed to eat as
much as they desired. A hidden digital balance secured under
a placemat and linked to the SIPM recorded the weight of food
being eaten. If the participant consumed .400 g pasta, an au-
dible alert accompanied by an onscreen message prompted the
participant to call the experimenter. The experimenter then
served the participant another 500 g pasta; no limit was placed
on the number of refills permitted. To reduce the influence of
habit and portion-size effects on intake, participants were en-
couraged not to use the refill prompt as a cue to end that course.
At the end of the pasta course, the participants re-rated their
appetite and mood. Next, they were served 150 g chocolate ice
cream with the instruction that they could eat as much as they
liked after they had completed a taste test to evaluate the
pleasantness, sweetness, creaminess, and familiarity of the ice
cream followed by further ratings of hunger and fullness to
measure the appetizing effects of the second course. For this
course, refills were prompted when the participant had con-
sumed 100 g. When participants had confirmed that they had
FIGURE 1. Schematic representation of the timing of the fixed and test
meals and the key sets of appetite ratings on each test day.
4Abbreviations used: HE, high energy; HS, high sensory; LE, low energy;
LS, low sensory; MS, medium sensory; SIPM, Sussex Ingestion Pattern
Monitor; VAS, visual analog scale.
SENSORY QUALITY AND SATIETY
by guest on November 24, 2015
finished eating, they were asked a final series of mood and ap-
petite ratings; this completed the lunch session.
With 6 preloads, a full factorial counterbalanced design would
have required 720 participants. Instead, a Williams Latin square
design (22) was used, which ensures that first-order carryover
effects and the number of times a treatment follows another
treatment are balanced. This approach generated 6 preload
sequences, with each sequence used 6 times; sequences were
balanced across sex.
On the final test day, the participants answered a series of
questions about the study to assess their awareness of the true
rationale. Their height (m) and weight (kg) were measured, and
they were paid £50 for participating in the study. The study
protocol received ethical clearance from the School of Life
Sciences research governance committee (University of Sussex)
and adhered to recommendations of the Declaration of Helsinki
of 1975 as revised in 1983.
Breakfast and lunch
On the morning of each test day, participants consumed
a breakfast of 60 g of a proprietary breakfast cereal (Crunchy Nut
Cornflakes; Kellogg Co) plus 160 mL semi-skimmed milk
(Sainsbury’s) and 200 mL orange juice (Sainsbury’s). The
breakfast provided 390 kcal, 4.7 g fat, 8.7 g protein, and 77.0 g
carbohydrate. For the ad libitum lunch, each 500-g serving of
pastaconsisted of 250 g cookedweight of pasta [conchiglie (shell
pasta); Sainsbury’s] plus 250 g of a prepared pasta sauce (tomato
and basil; Sainsbury’s) served hot. The ice cream course was
served in a bowl and was chocolate flavored (Chocolate In-
spiration, Carte d’Or; Unilever).
Six preload drinks, which differed in energy and sensory
characteristics, were tested: LE-LS, LE-MS, LE-HS, HE-LS,
HE-MS, and HE-HS. Pilot tests (see Table 1) established that the
LE and HE versions were matched for pleasantness, creaminess,
sweetness, and novelty; and that the LS, MS, and HS versions
differed significantly in thickness, creaminess, and how filling
they were expected to be. Sensory and hedonic evaluations of
the final versions of these drinks from the pilot studies (shown in
Table 1) were based on ratings made by 20 healthy, normal-
weight volunteers who assessed all 6 drinks in a single tasting
The served portions (300 g) of the LE versions of the drinks
delivered 78 kcal; the HE versions delivered 279 kcal. The sen-
sory manipulations added no energy. Each serving consisted of
no-added-sugar pomegranate juice drink (Sainsbury’s; HE: 185
g; LE: 220 g), no-sugar orange and mango squash (Robinson’s;
both versions: 30 g each), 0.1%-fat fromage frais (Sainsbury’s;
HE: 25 g; LE: 50 g), rhubarb flavor (International Flavors and
Fragrances; both versions: 4 drops each), red color (Silver-
spoon; HE: 8 drops; LE: 6 drops), yellow color (Silverspoon;
both versions: 4 drops), and yogurt flavor (International Fla-
vors and Fragrances; LE: 10 drops). The energy difference was
achieved by adding 35 g carbohydrate in the form of malto-
dextrin (Cargill) and 25 g protein in the form of whey protein
isolate (Myprotein) to the HE versions. To match the thickness
of the high- and low-energy drinks in the LS version, 0.3 g tara
gum (Kalys) was added to the LE-LS drink. For the MS and
HS versions of the drink, sensory manipulations were achieved
by adding tara gum (LE-MS: 0.45 g; LE-HS: 0.9 g; HE-MS:
0.5 g; HE-HS: 1.5 g), milk caramel flavor (S Black; LE-MS:
0.1 g; LE-HS: 0.3 g; HE-MS: 0.1 g; HE-HS: 0.3 g), and vanilla
extract (Nielsen-Massey; LE-MS: 20 drops; LE-HS: 60 drops;
HE-MS: 20 drops; HE-HS: 60 drops).
The aim of the study was to test whether altering the sensory
context in which energy was delivered altered subsequent ap-
petite, measured both as intake at a subsequent test lunch and
through ratings of appetite both during the period between
consuming the preload and starting lunch and during the test
lunch. One of the key measures of satiety is the degree to which
increased energy intake before a meal results in reduced intake at
that meal (energy compensation). We therefore calculated the
degree ofcompensation forthe extra 201 kcal in theHE drinksby
calculating the difference in energy intake at the test meal be-
tween HE and LE conditions in each sensory context and
expressing that change as a percentage of the difference in
preload energy to produce a compensation measure. Initial
not normally distributed, and this was not corrected by data
transformation. Outlier analysis indicated that 4 compensation
values from male participants were .2 SDs from the relevant
condition mean, and consequently data from these 4 subjects
Sensory and hedonic evaluations (on 100-point visual analog scales) of the
test preload drinks both in the pilot study and in the main intake test1
49.8 6 3.7a
61.3 6 3.0a
53.6 6 2.8a
62.8 6 3.5a
58.7 6 2.5b
68.3 6 2.6b
54.8 6 3.0a
64.2 6 3.1a
58.1 6 2.5a
66.3 6 3.0a
66.1 6 2.6b
70.5 6 2.3b
62.1 6 3.7
75.4 6 2.4
62.1 6 2.6
74.0 6 2.4
62.4 6 2.8
72.7 6 2.7
58.9 6 3.0
72.7 6 4.7
53.8 6 3.7
73.6 6 4.2
55.5 6 3.9
70.8 6 4.1
43.3 6 3.8a
54.9 6 1.9a
50.3 6 3.4b
56.3 6 2.3a
52.3 6 3.3b
60.9 6 3.5b
1All values are means 6 SEs. n = 20 for the pilot study; n = 32 for the
main study. In the pilot study, healthy normal-weight volunteers (12 women
and 8 men) assessed all 6 drinks in a single taste test. ANOVA was used to
analyze each rating depending on both sensory context and energy content.
No significant main effects or interactions were found for energy content
with any of these ratings, but significant main effects of sensory context were
found for thickness, creaminess, and how filling the drinks were perceived to
be. For these evaluations, mean ratings with different superscript letters
differed significantly (P ? 0.05) on the basis of Bonferroni-protected con-
trasts between the 3 sensory conditions.
YEOMANS AND CHAMBERS
by guest on November 24, 2015
were omitted from all further analyses. Compensation data
distributed normally once the outliers had been removed. All
analyses were based on the remaining sample of 32 participants.
Initial analyses contrasted compensation data between the 3
sensory contexts with the use of a repeated-measures ANOVA,
with protected contrasts between conditions. Overall energy
intakes at lunch were also contrasted by using ANOVA with
sensory context and preload energy as within-subject factors
because compensation measures provided an index of the dif-
ference between energy conditions butdid not determinewhether
it was intake in the HE or LE conditions that varied between
sensory contexts. Finally, because the lunch consisted of 2
courses, intake at each course was analyzed, because this pro-
vided an indication of whether energy caused a general sup-
pression of intake or altered decisions on when to end the one
course and start the next.
Appetite ratings were made before and after preload con-
sumption, 30 min after the preload, and periodically throughout
the test meal. A full analysis of these data was beyond the scope
of this study. The key question was whether the energy content
and sensory context of the drink modified appetite. To test this,
ratings at 5 key points were used: before and after preload
consumption, 30 min after preload, once the first course had been
tasted, and at the end of the test meal. These ratings were
contrasted by using ANOVA, with rating time, sensory context,
and energy content as within-subject factors.
Toensure thatpreloadsdifferedinkeysensory dimensions,but
were similar on other dimensions, we analyzed ratings of the
sensory and hedonic quality of the preloads when they were first
tasted by using 2-factor ANOVA. To test how preloading altered
the evaluation of the 2 lunch foods, we conducted similar
analyses on pasta and ice cream ratings.
To control for order effects, the 6 preload sequences were
included as a between-subject factor in all analyses. Sex was also
included as a between-subject factor in all analyses; however,
because sex was not a focus of the study and did not alter the
primary findings, for brevity full details of all effects of sex have
Total energy consumed at lunch depended on both the energy
content and sensory quality of the preload drink (Figure 2), with
the difference in lunch intake between LE and HE preload
conditions increasing as the drinks increased in thickness and
Although the adoption of a 2-course meal model made the
eatingtest more realistic than previous single-food lunch tests, one
potential concern was that the high palatability of the ice cream
might have ameliorated effects of the energy and sensory
manipulations, particularly because lunch palatability has been
shown to counteract the satiating effects of energy preloads (23).
We therefore also analyzed intake separately for each course. As
shown in Figure 2, intake of the pasta course did not differ sig-
significantly less was consumed in HE than in LE conditions in
both the MS and HS contexts. Less ice cream was also consumed
after the HE than after the LE preloads, and ice cream intake
decreased from LS to HS contexts. These 2 effects combined
meant that most of the ice cream was consumed in the LE-LS
condition and the least was consumed in the HE-HS condition.
In relation to the study hypotheses, the critical question was to
what extent the change in intake at lunch compensated for the
extra energy in the HE preloads. To test this, the difference in
energy intake between equivalent HE and LE conditions for each
sensory level was calculated and expressed as a percentage of the
actual energy difference between LE and HE conditions such that
a value of 100% would reflect full compensatory eating (24).
Compensation was significantly greater in the HS (87%) than
in the LS (18%) context (P , 0.005); the MS context was
As would be expected, men tended to eat more overall at lunch
than did women (P = 0.051). There were also complex signifi-
cant interactions between the 2 test conditions and effects of
order and sex, with a 4-way interaction between sensory context,
FIGURE 2. Mean (6SE) effects of the preload manipulations on intake
during the 2 separate courses and overall at the test meal for each of the 3
sensory contexts. Open bars represent low-energy preloads; filled bars
represent high-energy preloads. n = 32. ANOVA showed significant 2-way
interactions between preload energy content and sensory context for both
intake at the first course (P , 0.05) and overall (P , 0.05).
*,***Significant differences between energy conditions in each sensory
context are indicated: *P , 0.05, ***P , 0.001.
SENSORY QUALITY AND SATIETY
by guest on November 24, 2015
energy content, test order, and sex (P = 0.03). Across all con-
ditions, participants consumed the least at the last session but
similar amounts on other days, suggesting possible boredom
effects at the last session. Inspection of these data suggested that
intake at lunch after the HE preload in the LS context was less
when this was experienced early in the study than at the end,
suggesting that energy became increasingly less effective at gen-
erating satiety in that context. In contrast, meal size after the HE
preloads was similar regardless of when they were experienced in
both the MS and HS contexts. Thus, the repeated-exposure effects
of the study seem to have reduced the effectiveness of energy in
the LS context, particularly in female participants, and this likely
explains the 4-way interaction.
Appetite ratings before lunch
We hypothesized that the extent to which the experience of
appetite changed in the period between consuming the drink and
starting the test lunch would depend both on the energy content
and sensory quality of these drinks. To test this, ratings of hunger
and fullness before and after consuming the drink, 30 min later
(just before lunch) and after first tasting the lunch were con-
trasted. To test whether participants also consumed their lunch to
the same level of satiety, ratings at the end of the meal were also
included. For hunger (Figure 3, left), analysis showed 2 key
interactions: a 2-way interaction between the energy content and
time of rating (P = 0.021) and a 3-way interaction between
energy, time of rating, and the sensory condition (P = 0.023 as
well as a main effect of time (P , 0.001), with hunger de-
creasing immediately after drink consumption and then tending
to increase over the subsequent 30 min before decreasing to
a low value at the end of the test lunch. There were no significant
differences between energy or sensory conditions before or
immediately after the drinks were consumed. However, whereas
hunger before lunch being served did not differ significantly
between HE and LE conditions in the LS context, these ratings
tended to be lower in HE than in LE conditions in the MS
context (P = 0.08) and were significantly lower after HE than
after LE preloads in the HS context (P = 0.007: see Figure 3).
Because hunger ratings once food has been tasted have been
shown to have the strongest correlation to meal size (25), we
also looked at hunger ratings immediately after tasting the pasta,
which again depended both on energy content and sensory
quality of the preload. After tasting the lunch, there was still no
significant difference in hunger between LE and HE preloads in
the LS context, there was a marginally significant difference in
MS context (P = 0.07), and there was a significant difference
between preload energy conditions in the HS context (P =
0.002). There were no differences in hunger between conditions
at the end of the meal despite differences in lunch intake.
Analysis of fullness ratings (Figure 3, right) also found the
expected significant main effect of time of rating (P , 0.001),
with fullness mirroring the changes in hunger. There were also
a significant 2-way interaction between energy and sensory (P =
0.002) and a significant 3-way energy · sensory · time in-
teraction (P = 0.015). Surprisingly, fullness ratings were sig-
nificantly lower after the HE than after the LE preload in the LS
context immediately after the drink was consumed (P = 0.048)
but did not differ between energy conditions either before lunch
or once food had been tasted in the LS context. In contrast,
fullness was significantly greater after the HE than after the LE
preload in both the MS and HS contexts before lunch (MS =
0.039, HS = 0.001) and after lunch had been tasted (MS, P =
0.048; HS, P = 0.002). Fullness did not differ significantly be-
tween conditions at the end of the test lunch.
Evaluations of the drink preloads and test lunch
The drink preloads were designed to be matched in sensory
quality between the LE and HE formulations within each sensory
context but to increase in thickness and creaminess from the LS
through the MS and HS contexts. Pilot data suggested that the
preloads had these qualities, and the actual evaluations of the
drinks when they were first tasted on each preload session (Table
1) were analyzed to ensure this. As expected, the perceived
contextsbutwasunaffected byenergycontent. Thedrinks didnot
differ significantly in terms of pleasantness, but there was an
unexpected effect of energy content on perceived sweetness, with
HE drinks being rated slightly sweeter than LE drinks regardless
of sensory context, an effect that was not evident in the pilot
There were no significant effects of the preload manipulations
on the rated pleasantness of either the pasta or ice cream con-
sumed in the test lunch.
Our main finding was that the degree to which a beverage
quality: participants consumed less after an HE preload in the HS
the extra energy in the HE drink was 87%, compared with 47% in
importance of this outcome, future studies should aim to replicate
this finding and to determine whether these effects are sustained
beyond the immediate test meal.
How might sensory context modify the response to added
nutrients? One possibility is that orosensory cues generate ex-
pectations about satiety that prime the appetite system to re-
spond to subsequent nutrient-derived satiety cues. Accordingly,
when the cues generated lower satiety expectations (our LS
context), the lack of preparatory changes reduced the efficiency
of nutrient processing. This view is consistent with evidence that
food-related cues result in learned preparatory responses (9, 26)
and thatsensoryqualitiesgenerateclear satietyexpectations(27).
Our findings may explain differences in the satiating effects of
foods and drinks varying in sensory quality (14, 28) and suggest
that caution is needed in interpretation of studies that report
differences in satiety where nutrient and sensory differences are
confounded: thus, for example, the finding that a breakfast with
isocaloric juice (29) could be attributed to sensory or nutrient
A corollary of the idea that sensory cues that predict energy
increase the efficiencyof nutrient processing is that experience of
such cues followed by an absence of nutrient ingestion could
leave the body prepared for nutrients, resulting in rebound
hunger. The higher prelunch hunger ratings in the LE-HS con-
YEOMANS AND CHAMBERS
by guest on November 24, 2015
LE preload in the HS than in the LS context. These findings need
substantiation but imply that diet-related products that generate
satiety expectations but fail to deliver nutrients may lead to
subsequent increased appetite and overeating.
A potential alternative explanation for the enhanced satiating
effects of energy in the HS context might be that the constituents
used to modify sensory characteristics enhanced the satiating
effects of added nutrients through a postingestive effect. The
FIGURE 3. Mean (6SE) ratings of hunger (left panels) and fullness (right panels) across the course of the test session for both the low-energy (h) and
high-energy ( ) preloads in the low (A), medium (B), and high (C) sensory contexts. n = 32. ANOVA showed a significant 3-way interaction between energy,
time of rating, and sensory condition (P , 0.05). *,**Significant differences between high- and low-energy conditions at each time point within each sensory
context are indicated: *P , 0.05, **P , 0.01.
SENSORY QUALITY AND SATIETY
by guest on November 24, 2015
addition of tara gum increased viscosity, and viscosity has been
reported to enhance satiation (20) and satiety (19, 30–32), per-
haps by altering the gastric emptying rate. However, studies to
date cannot dissociate whether viscosity operates through an
orosensory or a postingestive effect. If increased viscosity alone
produced more satiety than a drink with the same energy but
lower gum content (LE-HS compared with LE-LS). However,
this was not the case. Amounts of tara gum used were very low:
1.5 g in total in the most satiating drink (HE-HS) compared with
0.3 g added in the control (LE-LS). Studies that report significant
example, 12 g guar gum increased the satiating efficiency of
a high-fat soup (33). Likewise, enhanced satiety was reported
after addition of 12 g inulin in a protein-rich beverage (34),
although other studies failed to find fiber and protein more sa-
tiating than protein alone (28). Taken together, these various
findings cannot fully discount an explanation in terms of possible
postingestive effects of tara gum, but they do suggest that an
orosensory explanation is more plausible.
Previous studies have reported variable effects of drink pre-
loads on appetite (see reference 35). In particular, studies that
use recognizable beverages such as drinks sweetened with su-
crose or artificial sweeteners often report minimal compensatory
eating (36–40), but other studies report at least partial com-
pensation for energy consumed as drinks (15, 41). Our present
results may explain this variability. Drinks often lack the sensory
elements that generate satiety expectations. In the absence of
such expectations, the appetite system appears to be less able to
respond to nutrient signals. Previous studies that reported some
degree of compensation for energy in a beverage context had
sensory elements that should generate satiety expectations:
chocolate-flavored milk was more satiating than was cola (42)
and a yogurt/fruit drink was more satiating than an equally ca-
loric fruit juice (14). Thus, although drinks tend to be less sa-
tiating than solid foods (35), drinks may be satiating when they
have sensory qualities that lead to satiety expectations.
In addition to the role of sensory cues in satiety indicated by
the present results, similar cues reduced meal size (ie, enhanced
satiation) irrespective of nutrient content (13, 43). As a conse-
quence, because our participants consumed all of each drink it
might have been expected that we would see greater satiation
immediately after consuming the HS and MS drinks relative to
the LS context. However, no such effect was evident.
The outcome of the present study relied on the successful
manipulation of the sensory characteristics of the drinks. Eval-
uations of preloads both in the pilot and main study confirmed
that the HS drinks were thicker, creamier, and more filling than
the LS drinks; the MS drinks were intermediate in these char-
acteristics. However, ratings in the main study were substantially
different from those in the pilot study. The likely explanation is
that contrast effects made subtle differences more evident when
drinkswererated at thesametime (the pilot)than when evaluated
on separate days. The HE and LE versions were matched in
thickness, creaminess, how filling they were perceived to be, and
pleasantness, but a significant increase in sweetness for all HE
versions was seen in the main study. Because the key findings
were for different effects of added energy across contexts, it is
unlikely that the effects on satiety could be explained by small
differences in preload sweetness.
One limitation of the study was that it tested only normal-
weight, unrestrained participants. Accuracy of compensatory
eating (44, 45) and responses to sensory quality (46, 47) vary
across individuals depending on body size, physical activity,
age, and restraint. Whether the same outcome would be seen in
restrained and overweight or obese participants needs to be
explored. The study looked only at effects of single exposures
to each preload, and the interaction between test order and
preload effect in women might suggest some changes in the
relation between sensory quality and nutrients through
learning (48). Finally, both the breakfast and test lunch were
relatively rich in carbohydrate. Whether similar effects would
be seen with test meals with higher fat or protein content
should be explored. In particular, it is clear that different
macronutrients selectively modify physiologic satiety cues
such as polypeptide YY, cholecystokinin, and glucagon-like
peptide 1 (see reference 49), and if one effect of sensory cues
was to modify release of these or related hormones, then the
macronutrient specificity of the sensory context effect could
be modified accordingly.
Overall, the present study provides evidence that the sensory
context in which nutrients are consumed modifies subsequent
satiety, with higher-energy beverages being more satiating when
experienced in a thicker, creamier flavored drink. This outcome
effects of subtle differences in orosensory cues on the outcome of
preload-style studies of satiety.
We thank Maria Cruz Barbero Bejarano for her assistance during the pilot
work on drink formulation and assistance with running the study and Ken
Woodward and Harvey Els from the School of Service Management at the
University of Brighton for their help in formulating the drink preloads.
The authors’ responsibilities were as follows—MRYand LC: contributed
equally to the design of the study; MRY: had primary responsibility for data
analysis and writing of the manuscript; and LC: was responsible for the ex-
perimental work and data collection and contributed to the data analysis and
manuscript writing. Neither of the authors had a conflict of interest related to
the conduct or reporting of this study.
1. Popkin BM, Doak CM. The obesity epidemic is a worldwide phe-
nomenon. Nutr Rev 1998;56:106–14.
2. James WP. WHO recognition of the global obesity epidemic. Int J Obes
(Lond) 2008;32(suppl 7):S120–6.
3. Drewnowski A, Darmon N. The economics of obesity: dietary energy
density and energy cost. Am J Clin Nutr 2005;82:265S–73S.
4. DiMeglio DP, Mattes RD. Liquid versus solid carbohydrate: effects on
food intake and body weight. Int J Obes Relat Metab Disord 2000;24:
5. Mattes R. Fluid calories and energy balance: the good, the bad, and the
uncertain. Physiol Behav 2006;89:66–70.
6. Mattes R. Soup and satiety. Physiol Behav 2005;83:739–47.
7. Himaya A, Louis-Sylvestre J. The effect of soup on satiation. Appetite
8. Brunstrom JM, Shakeshaft NG, Scott-Samuel NE. Measuring ’ex-
pected satiety’ in a range of common foods using a method of constant
stimuli. Appetite 2008;51:604–14.
9. Woods SC. The eating paradox: how we tolerate food. Psychol Rev
10. Cecil JE, Castiglione K, French S, Francis J, Read N. Effects of in-
tragastric infusions of fat and carbohydrate on appetite ratings and food
intake from a test meal. Appetite 1998;30:65–77.
11. Cecil JE, Francis J, Read NW. Relative contributions of intestinal,
gastric, oro-sensory influences and information to changes in appetite
induced by the same liquid meal. Appetite 1998;31:377–90.
YEOMANS AND CHAMBERS
by guest on November 24, 2015
12. Lavin JH, French SJ, Ruxton CH, Read NW. An investigation of the Download full-text
role of oro-sensory stimulation in sugar satiety? Int J Obes Relat Metab
13. de Graaf C, Kok FJ. Slow food, fast food and the control of food intake.
Nat Rev Endocrinol 2010;6:290–3.
14. Tsuchiya A, Almiron-Roig E, Lluch A, Guyonnet D, Drewnowski A.
Higher satiety ratings following yogurt consumption relative to fruit
drink or dairy fruit drink. J Am Diet Assoc 2006;106:550–7.
15. Bertenshaw EJ, Lluch A, Yeomans MR. Satiating effects of protein but
not carbohydrate consumed in a between meal beverage context.
Physiol Behav 2008;93:427–36.
16. Bertenshaw EJ. Beverages and satiety: the effects of nutrients, sensory
qualities and learning. Brighton, United Kingdom: Psychology, 2008.
17. Slavin J, Green H. Dietary fibre and satiety. Nutr Bull 2007;32:S32–42.
18. Marciani L, Gowland PA, Spiller RC, Manoj P, Moore RJ, Young P,
Al-Sahab S, Bush D, Wright J, Fillery-Travis AJ. Gastric response to
increased meal viscosity assessed by echo-planar magnetic resonance
imaging in humans. J Nutr 2000;130:122–7.
19. Mattes RD, Rothacker D. Beverage viscosity is inversely related to
postprandial hunger in humans. Physiol Behav 2001;74:551–7.
20. de Wijk RA, Zijlstra N, Mars M, de Graaf C, Prinz JF. The effects of
food viscosity on bite size, bite effort and food intake. Physiol Behav
21. Yeomans MR. Palatability and the microstructure of eating in humans:
the appetiser effect. Appetite 1996;27:119–33.
22. Williams EJ. Experimental designs balanced for the estimation of re-
sidual effects of treatment. Austr J Sci Res, Ser A 1949;2:149–68.
23. Yeomans MR, Lee MD, Gray RW, French SJ. Effects of test-meal
palatability on compensatory eating following disguised fat and car-
bohydrate preloads. Int J Obes 2001;25:1215–24.
24. Shide DJ, Cabellero B, Reidelgerger R, Rolls BJ. Accurate energy
compensation for intragastric and oral nutrients in lean males. Am J
Clin Nutr 1995;61:754–64.
25. Yeomans MR, Bertenshaw EJ. Appetite and food intake: a human
experimental perspective. In: Harris RBS, Mattes RD, eds. Appetite
and food intake: behavioral and physiological considerations. Boca
Raton, FL: Taylor and Francis, 2008:61–80.
26. Woods SC, Vasselli JR, Kaestner E, Szakmary GA, Milburn P, Vitiello
MV. Conditioned insulin secretion and meal feeding in rats. J Comp
Physiol Psychol 1977;91:128–33.
27. Brunstrom JM, Shakeshaft NG. Measuring affective (liking) and
non-affective (expected satiety) determinants of portion size and food
reward. Appetite 2009;52:108–14.
28. Almiron-Roig E, Grathwohl D, Green H, Erkner A. Impact of some
isoenergetic snacks on satiety and next meal intake in healthy adults.
J Hum Nutr Diet 2009;22:469–74.
29. Dove ER, Hodgson JM, Puddey IB, Beilin LJ, Lee YP, Mori TA. Skim
milk compared with a fruit drink acutely reduces appetite and energy
intake in overweight men and women. Am J Clin Nutr 2009;90:70–5.
30. Zijlstra N, Mars M, de Wijk RA, Westerterp-Plantenga MS, de Graaf
C. The effect of viscosity on ad libitum food intake. Int J Obes 2008;
31. Zijlstra N, Mars M, de Wijk RA, Westerterp-Plantenga MS, Holst JJ,
de Graaf C. Effect of viscosity on appetite and gastro-intestinal hor-
mones. Physiol Behav 2009;97:68–75.
32. Russell K, Delahunty C. The effect of viscosity and volume on
pleasantness and satiating power of rice milk. Food Qual Prefer 2004;
33. French SJ, Read NW. Effect of guar gum on hunger and satiety after
meals of differing fat content: relationship with gastric emptying. Am J
Clin Nutr 1994;59:87–91.
34. Perrigue MM, Monsivais P, Drewnowski A. Added soluble fiber en-
hances the satiating power of low-energy-density liquid yogurts. J Am
Diet Assoc 2009;109:1862–8.
35. de Graaf C. Why liquid energy results in overconsumption. Proc Nutr
36. Rolls BJ, Kim S, Fedoroff IC. Effects of drinks sweetened with sucrose
or aspartame on hunger, thirst and food intake in men. Physiol Behav
37. Almiron-Roig E, Drewnowski A. Hunger, thirst, and energy intakes
following consumption of caloric beverages. Physiol Behav 2003;79:
38. Canty DJ, Chan MM. Effects of consumption of caloric versus non-
caloric sweet drinks on indices of hunger and food consumption in
normal adults. Am J Clin Nutr 1991;53:1159–64.
39. Green SM, Burley VJ, Blundell JE. The role of sweetness and carbo-
hydrate in the satiating capacity of sucrose. Int J Obes 1992;16:54
40. Soenen S, Westerterp-Plantenga MS. No differences in satiety or en-
ergy intake after high-fructose corn syrup, sucrose, or milk preloads.
Am J Clin Nutr 2007;86:1586–94.
41. Almiron-Roig E, Flores SY, Drewnowski A. No difference in satiety or
in subsequent energy intakes between a beverage and a solid food.
Physiol Behav 2004;82:671–7.
42. Harper A, James A, Flint A, Astrup A. Increased satiety after intake of
a chocolate milk drink compared with a carbonated beverage, but no
difference in subsequent ad libitum lunch intake. Br J Nutr 2007;97:
43. Zijlstra N, de Wijk RA, Mars M, Stafleu A, de Graaf C. Effect of bite
size and oral processing time of a semisolid food on satiation. Am J
Clin Nutr 2009;90:269–75.
44. Long SJ, Hart K, Morgan LM. The ability of habitual exercise to in-
fluence appetite and food intake in response to high- and low-energy
preloads in man. Br J Nutr 2002;87:517–23.
45. Appleton KM, Martins C, Morgan LM. Age and experience predict
accurate short-term energy compensation in adults. Appetite 2011;56:
46. Tepper BJ, Farkas BK. Reliability of the sensory responder classifi-
cation to learned flavor cues: a test-retest study. Physiol Behav 1994;
47. Yeomans MR, Tovey HM, Tinley EM, Haynes CL. Effects of manip-
ulated palatability on appetite depend on restraint and disinhibition
scores from the Three Factor Eating Questionnaire. Int J Obes 2004;28:
48. Brunstrom JM. Associative learning and the control of human dietary
behavior. Appetite 2007;49:268–71.
49. Geraedts MCP, Troost FJ, Saris WHM. Gastrointestinal targets to
modulate satiety and food intake. Obes Rev 2011;12:470–7.
SENSORY QUALITY AND SATIETY
by guest on November 24, 2015