Design and testing of foods differing in protein to energy ratios.
ABSTRACT Our aim was to design a selection of foods with differing proportions of protein but equal palatability in two settings, Sydney Australia and Kingston Jamaica. The foods were manipulated to contain 10, 15 or 25% E as protein with reciprocal changes in carbohydrate to 60, 55 or 45% E and dietary fat was kept constant at 30%. Naïve participants did not identify a difference in protein between the versions. On average, the versions were rated equal in pleasantness (Sydney-10%: 44±2, 15%: 49±2 and 25%: 49±2 Kingston-10%: 41±3, 15%: 41±3 and 25%: 37±3).
- SourceAvailable from: Miguel A Iglesias[Show abstract] [Hide abstract]
ABSTRACT: A significant contributor to the rising rates of human obesity is an increase in energy intake. The 'protein leverage hypothesis' proposes that a dominant appetite for protein in conjunction with a decline in the ratio of protein to fat and carbohydrate in the diet drives excess energy intake and could therefore promote the development of obesity. Our aim was to test the 'protein leverage hypothesis' in lean humans by disguising the macronutrient composition of foods offered to subjects under ad libitum feeding conditions. Energy intakes and hunger ratings were measured for 22 lean subjects studied over three 4-day periods of in-house dietary manipulation. Subjects were restricted to fixed menus in random order comprising 28 foods designed to be similar in palatability, availability, variety and sensory quality and providing 10%, 15% or 25% energy as protein. Nutrient and energy intake was calculated as the product of the amount of each food eaten and its composition. Lowering the percent protein of the diet from 15% to 10% resulted in higher (+12±4.5%, p = 0.02) total energy intake, predominantly from savoury-flavoured foods available between meals. This increased energy intake was not sufficient to maintain protein intake constant, indicating that protein leverage is incomplete. Urinary urea on the 10% and 15% protein diets did not differ statistically, nor did they differ from habitual values prior to the study. In contrast, increasing protein from 15% to 25% did not alter energy intake. On the fourth day of the trial, however, there was a greater increase in the hunger score between 1-2 h after the 10% protein breakfast versus the 25% protein breakfast (1.6±0.4 vs 25%: 0.5±0.3, p = 0.005). In our study population a change in the nutritional environment that dilutes dietary protein with carbohydrate and fat promotes overconsumption, enhancing the risk for potential weight gain.PLoS ONE 10/2011; 6(10):e25929. · 3.53 Impact Factor
Article: Protein leverage and energy intake.[Show abstract] [Hide abstract]
ABSTRACT: Increased energy intakes are contributing to overweight and obesity. Growing evidence supports the role of protein appetite in driving excess intake when dietary protein is diluted (the protein leverage hypothesis). Understanding the interactions between dietary macronutrient balance and nutrient-specific appetite systems will be required for designing dietary interventions that work with, rather than against, basic regulatory physiology. Data were collected from 38 published experimental trials measuring ad libitum intake in subjects confined to menus differing in macronutrient composition. Collectively, these trials encompassed considerable variation in percent protein (spanning 8-54% of total energy), carbohydrate (1.6-72%) and fat (11-66%). The data provide an opportunity to describe the individual and interactive effects of dietary protein, carbohydrate and fat on the control of total energy intake. Percent dietary protein was negatively associated with total energy intake (F = 6.9, P < 0.0001) irrespective of whether carbohydrate (F = 0, P = 0.7) or fat (F = 0, P = 0.5) were the diluents of protein. The analysis strongly supports a role for protein leverage in lean, overweight and obese humans. A better appreciation of the targets and regulatory priorities for protein, carbohydrate and fat intake will inform the design of effective and health-promoting weight loss diets, food labelling policies, food production systems and regulatory frameworks.Obesity Reviews 03/2014; 15(3):183-91. · 7.86 Impact Factor
Design and testing of foods differing in protein to energy ratios§
Alison K. Gosbya,*,1, Suzanne Soares-Wynterb, Claudia Campbellb, Asha Badaloob, Marion Antonellic,
Arthur D. Conigravec, Terrence G. Forresterb, David Raubenheimerg, Stephen J. Simpsona
aSchool of Biological Sciences, The University of Sydney, Rm 322, Heydon Laurence Building, A08, Sydney, NSW 2006, Australia
bTropical Metabolism Research Unit, Tropical Medicine Research Institute, University of the West Indies, Kingston, Jamaica
cSchool of Molecular Bioscience, The University of Sydney, NSW 2006, Australia
dMedical Research Council Human Nutrition Research, Cambridge United Kingdom
eThe Liggins Institute, The University of Auckland, Auckland, New Zealand
fBoden Institute of Obesity, Nutrition and Exercise, University of Sydney, NSW 2006, Australia
gInstitute for Natural Resources, Massey University, Auckland, New Zealand
To test feedback effects of macronutrients on energy intake, it is
important to disguise the macronutritional composition of foods.
This can be done by minimising differences in sensory aspects and
in pleasantness, sweetness and savouriness. Stubbs et al. (2001)
successfully disguised the fat content of foods, producing covert
manipulation in low fat and high fat foods. Overall participants
were unable to detect differences with the exception of dairy-
based foods. There have been no studies that have attempted a
covert manipulation of the protein to energy ratio of foods,
however, and that was the primary aim of the present study.
Experimental data indicates that protein is the most satiating
macronutrient group for humans and may therefore protect
against over-consumption (Barkeling, Rossner, & Bjorvell, 1990;
Rolls, Hetherington, & Burley, 1988). To date, experiments testing
the effect of the percentageof proteinenergy(E) in the dieton total
E intake suggest that protein feedback may exert an influence on
total E intake over a matter of 1–2 days. Such studies have not
however controlled for sensory differences between foods and
effects of prior exposure (Simpson, Batley, & Raubenheimer, 2003;
Simpson & Raubenheimer, 2005; Weigle et al., 2005). To test
whether feedbacks associated with protein act in the absence of
such cues, it is necessary to first test whether the protein to energy
ratiosof a varietyof sweetand savoury testfoods canbe concealed.
This study presents the design and palatability testing of such
foods, in two culturally diverse settings—in Sydney, Australia and
Experimental methods and materials
Approval for the study was obtained from the Sydney South
West Area Health Service Ethics Review Committee and The
University of West Indies, Mona, Ethics Committee. Males and
females aged 18–65 were recruited using the casual employment
website at the University of Sydney and via email to School of
Biological Sciences. Participants were not paid. Ten male (BMI:
23 ? 0.5 kg/m2(mean ? sem)) and six female (BMI: 21 ? 1.0 kg/m2)
participants were included in the first taste test trials in Sydney,
Appetite 55 (2010) 367–370
A R T I C L EI N F O
Received 2 March 2010
Received in revised form 12 May 2010
Accepted 14 June 2010
Protein leverage hypothesis
A B S T R A C T
Our aim was to design a selection of foods with differing proportions of protein but equal palatability in
two settings, Sydney Australia and Kingston Jamaica. The foods were manipulated to contain 10, 15 or
25% E as protein with reciprocal changes in carbohydrate to 60, 55 or 45% E and dietary fat was kept
constant at 30%. Naı ¨ve participants did not identify a difference in protein between the versions. On
average, the versions were rated equal in pleasantness (Sydney-10%: 44 ? 2, 15%: 49 ? 2 and 25%: 49 ? 2
Kingston-10%: 41 ? 3, 15%: 41 ? 3 and 25%: 37 ? 3).
Crown Copyright ? 2010 Published by Elsevier Ltd. All rights reserved.
§We would like to acknowledge Gerry Quinn of Deakin University for his adivce
on statistical analysis. This project was funded by a National Health and Medical
Research Council of Australia project grant. There is no conflict of interest.
* Corresponding author.
E-mail address: firstname.lastname@example.org (A.K. Gosby).
1The recipe can be obtained by contacting the corresponding author.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/appet
0195-6663/$ – see front matter. Crown Copyright ? 2010 Published by Elsevier Ltd. All rights reserved.
Australia, held over 2 days in a metabolic kitchen. 24 h food diaries
provided estimates for the average intake on the day prior to the
study of 10950 ? 940 kJ (19.7% protein, 38.7% fat and 41.6%
carbohydrate) for males and 7630 ? 750 (18.2% protein, 32.7% fat
and 49.1% carbohydrate) for females indicating adequate nutrition
prior to the trials and thus minimized the likelihood of any effects of
nutritional state on food preferences (Hill & Blundell, 1982). Nine of
these participants (5 males and 4 females) were chosen as being
knowledgeable, i.e. they had explained to them that the nutrient of
interest was protein, whereas the remaining 7 (5 males and 2
females) were kept uninformed about the purpose of the study.
Following analysis, 12 participants, 5 males (BMI: 25.6 ? 0.8 kg/m2)
and7females(BMI:21.3 ? 1.2 kg/m2)wererecruitedtoretest5foods
which had undergone further manipulations to further reduce any
differences in pleasantness. The average estimated intake for the 24 h
prior to this retesting was 11040 ? 2350 kJ (17.5% protein, 29.7% fat
and 52.8% carbohydrate) for males and 9090 ? 680 kJ (17.6% protein,
30.7% fat and 51.6% carbohydrate) for females. In Kingston, Jamaica,
14 participants with no prior knowledge of the manipulation
attended the Tropical Metabolism Research Unit, University of West
Indies Mona. All participants in Sydney and Jamaica were instructed
to consume a typical breakfast on the day of testing 4 h prior to
arriving and instructed not to eat or drink anything, except water,
following this breakfast. Sydney participants were also asked to
complete a 24 h food diary on the day prior to the taste testing study,
to ensure that subjects were not energy-deprived prior to testing. 12
Subjects give a power of 0.8 to find an effect size of (approximately
10 mm on the VAS scale, f = 0.35) with a significance of 0.05 within
subjects (G*Power 3.1, ANOVA, effect of treatment within subjects).
Recipe manipulation and study design
Recipes were sourced from recipe websites, cookbooks and food
packages and then modified to contain 10, 15 or 25% E as protein.
Carbohydrate was adjusted tobe 60, 55or45% E and dietary fatwas
kept constant at 30% E. Individuals would be able to meet protein
requirements on each of these compositions, whether lean, obese,
male or female. Energy density (kJ/g) was similar between the 10%,
foods (i.e. the energy density of the 10% version of one food differed
from the energy density of the 10% version of another food).
Macronutrient information (protein, fat and carbohydrate), in
grams, for each ingredient was collected either from the food label
orfromNutrientTablesforuse inAustralia 2006(NUTTAB 2006; for
the Sydney foods) or the USDA nutrient database (for the Kingston
energy values for each macronutrient were: protein (17 kJ/g),
carbohydrate (sugars: 16 kJ/g and complex carbohydrates 17 kJ/g)
and fat (37 kJ/g). Nutritional information for fibre, sugars and
foods. In some cases food ingredients, an experimental protein mix,
and/or maltodextrin (Sydney: Polyjoule, Nutricia Australia Pty Ltd.;
Jamaica: Polycal,Nutricia),a polysaccharideconsisting of D-glucose,
were used to manipulate the protein to energy ratio of the recipes.
The experimental protein mix was a 1:1:1 mix of whey protein
concentrate, calcium caseinate and egg white powder [Sydney:
Muscle Brand Pty Ltd., Myopure, Greenwich, Australia; Kingston:
Now, IL, USA (egg white powder and calcium caseinate) and Bobs
Red Mill, OR, USA (whey protein concentrate)]. All essential amino
g protein (‘‘Protein and amino acid requirements in human
nutrition,’’ 2007). The proportion of each essential amino acid per
g of protein expressed as a percent of total essential amino acids is
by the FAO/WHO/UNU technical report (‘‘Protein and amino acid
requirementsinhumannutrition,’’ 2007). The sameprotein sources
protein mix in Jamaica and similar amino acid profiles were
Supplementary Tables 2 and 3 Appendix A for Sydney and Kingston
foods, respectively. Energy, protein, fat and carbohydrate are
expressed as kJ/100 g of prepared food. Fibre and sugars as g/
100 g prepared food and salt as mg/100 g prepared food.
In Sydney, participants tested 9 foods (1. Raspberry Muesli, 2.
Banana Bread, 3. Savoury Breakfast Muffins, 4. Apricot Yoghurt
Muffins, 5. Sushi, 6. Sweet Potato Wrap, 7. Orange and Poppyseed
Cake, 8. Mushroom Pasta and 9. Chow Mein Mince); in Kingston
Soup, 4. Rice and Peas, 5. Curry, 6. Seasoned Rice, 7. Chicken Chow
Participants were presented with the 10%, 15% and 25% protein
versions of each of these food items simultaneously. They were
instructed to sample each version and then complete the question-
naire. The questionnaire was a modification of a previous
in pleasantness, sweetness and savouriness between the 10, 15 and
to 100 mm. The three questions asked were; ‘‘How sweet did you
find the food?’’, ‘‘How savoury did you find the food?’’ and ‘‘How
25% versions of each food could be distinguished by asking
participants to comment on differences in appearance, smell, taste
ortexture. Participants could alsocomment further onthe natureof
the difference. The subjects were then asked to comment on how
they perceived the foods with respect to protein, carbohydrate,
sugar, fibre, salt, fat and energy. Permissible responses were ‘‘high
in’’, ‘‘low in’’, ‘‘no difference’’ and ‘‘don’t know’’.
Data are presented as means ? standard errors unless otherwise
specified. Repeated measures analysis of variance was used to test for
and protein level for pleasantness, sweetness and savouriness ratings
given to foods via the visual analogue scale, before and after
modifications. The Greenhouse–Geiser correction was used for
repeated measures to correct for any sphericity violations (&:
Greenhouse-Geiser correction factor). Effects of the protein to energy
manipulation on distinguishability are presented as a percentage of
subjects identifying with a ‘‘yes’’ response, that on the whole, the
versions(10%, 15% and 25%)of eachfoodweredifferent inappearance,
smell, tasteand texture. An average of these responses to allfoodswas
calculated and is presented as means (%) ? standard errors. Partici-
pants could also comment on the nature of the difference. Although
these comments were not included in the analysis they provided a
measure of the differences identified. Nutritional perceptions are
presented as the sum of ‘‘high in’’, ‘‘low in’’, ‘‘no difference’’ and ‘‘don’t
know’’responses foreachproteinlevel(10%,15% and25%)forallfoods
and for foods with (Sydney foods 3, 5, 6, 8 and 9) and without (Sydney
differencesweretestedusinga3 ? 4contingencytableandchisquared
analysis (where p < 0.05 degrees of freedom = 6, x2> 12.59 corre-
sponds to statistically significant differences).
Effect of the protein to energy manipulation on pleasantness,
sweetness and savouriness
Each version (i.e. 10, 15 and 25%) of each food was tested for
pleasantness, sweetness and savouriness in Sydney and in
Kingston (Appendix A: Supplementary Figs. 1 and 2). Overall
A.K. Gosby et al./Appetite 55 (2010) 367–370
there was no main effect of percent protein level on pleasantness
ratings of test foods in Sydney (10%: 44 ? 2, 15%: 49 ? 2 and 25%:
49 ? 2, F(2, 28)= 3.307, p = 0.1 (e = 0.703) or in Kingston (10%: 41 ? 3,
15%: 41 ? 3 and 25%: 37 ? 3, F(2, 12)= 2.802, p = 0.1, Greenhouse-
Geiser (G–G) corrected: e = 0.604). There were, however, differences
among the foods in pleasantness ratings in Sydney (F(8, 112)= 3.018,
p < 0.05, (G–G) corrected; e = 0.446) and in Kingston (F(9, 54)= 3.096,
p < 0.05, G–G corrected;e = 0.396). A significant interaction between
food type and percent protein was also found on the pleasantness
ratings in Sydney (F(16,
e = 0.381) and in Kingston (F(18, 108)= 4.129, p < 0.05, G–G corrected;
e = 0.193). Supplementary Fig. 1a (Appendix A) shows that in Sydney
the pleasantness rating of foods 1 and 2 increased with decreasing
content. Modifications to foods 1–5 were undertaken in Sydney in an
attempttoenhancetheirpleasantness ratingsandreducethe effectof
percent protein. Upon further testing of these modified foods the
significant interaction between food type and protein was reduced
but remained significant (F(8, 80)= 4.205, p < 0.01, G–G corrected:
e = 0.491). The pleasantness ratings of the five modified foods,
although statistically significant, were quantitatively small (Appen-
dix A: Supplementary Fig. 1b).
The percent protein in the test food did not affect the extent to
which foods were considered to be sweet or savoury. This is
indicated by non-significant interactions between percent protein
and food type for sweetness both in Sydney (F(16, 224)= 2.166,
p = 0.1, G–G corrected;e = 0.286) and in Kingston (F(18, 108)= 0.868,
p = 0.5, G–G corrected; e = 0.187), as well as savouriness in Sydney
(F(16, 224)= 1.919, p = 0.1, G–G corrected;e = 0.358) and in Kingston
(F(18, 108)= 1.148, p = 0.4, G–G corrected; e = 0.180). The non-
significant interactions were maintained after the retesting trials
ofmodifiedfoods1–5inSydneyforbothsweetness(F(8, 80)= 2.204,
p = 0.1, G–G corrected; e = 0.461) and savouriness (F(8, 80)= 2.413,
p = 0.1, G–G corrected; e = 0.307).
224)= 5.183, p < 0.0005, G–G corrected;
Effects of protein to energy manipulations on distinguishability
Overall, participants could distinguish between the three
versions of the foods as a result of the protein to energy
manipulation (Appendix A: Supplementary Fig. 3). On average,
87 ? 8% of subjects detected a difference in taste, 58 ? 22% in
appearance, 56 ? 18% in texture and 41 ? 16% in smell. After
modifications to the Sydney foods 1–5 the distinguishability was
reducedso that 70 ? 19% of subjectsdetected a differencein taste but
fewer than 50% detected a difference in appearance (40 ? 17%),
texture (47 ? 15%) and smell (21 ? 8%). In Kingston, 83 ? 3% of
subjects detected a difference in taste, 72 ? 7% in appearance,
60 ? 9% in texture and 38 ? 4% in smell.
The effect of the protein manipulation on perceptions of nutritional
Participants in Sydney and Kingston were asked to report on
their perceptions of the nutritional compositions of the 10, 15 and
25% versions of each food (Appendix A: Supplementary Table 1).
In Kingston, no statistically significant differences were
observed for subjects reported perceptions for differences in
protein (x2= 11.1, df = 6, p = 0.1), carbohydrate (x2= 4.35, df = 6,
p = 0.6) or fat (x2= 1.53, df = 6, p = 1.0) contents or for any of the
other parameters tested.
On the other hand, Sydney participants detected directional
differences in protein (x2= 45.03, df = 6, p < 0.0001) that were
statistically significant in both knowledgeable (x2= 30.80, df = 6,
p < 0.0001) and naı ¨ve subjects (x2= 15.75, df = 6, p < 0.02). After
grouping foods into meat (foods 3, 5, 6, 8 and 9) and non-meat
(foods 1, 2, 4 and 7) containing categories, the detection of
differences in protein contents were only significant in foods
containing meat (x2= 40.98, df = 6, p < 0.0001) but not in those
that were manipulated by the addition of the protein powder
supplements (x2= 11.47, df = 6, p = 0.07). Sydney participants did
not detect differences in carbohydrate content (all: x2= 2.67,
df = 6, p = 0.8; knowledgeable: x2= 1.69, df = 6, p = 0.9; naı ¨ve:
x2= 2.01, df = 6, p = 0.9) or in fat content (all: x2= 2.74, df = 6,
p = 0.8; knowledgeable: x2= 5.36, df = 6, p = 0.5; naı ¨ve: x2= 2.64,
df = 6, p = 0.8). However, salt was perceived to be different
(x2= 22.67, df = 6, p < 0.001). After grouping foods into meat
categories the perception of a difference in salt content was only
statistically significant in the meat group (meat: x2= 20.11, df = 6,
p < 0.0001; non-meat: x2= 4.74, df = 6, p = 0.1). In general, the
foods to be higher in salt than the 10% and 15% versions with a
greater number of ‘‘high in’’ responses and fewer ‘‘low in’’
responses. Consistent with these results, the average sodium
content (mg/100 g food) tended to be higher in the 25% meat foods
than in the 10% and 15% foods whereas there was a more modest
difference in salt (mg/100 g food) between the versions of sweet
foods. None of the participants could detect differences in sugar,
fibre or energy. Subsequent to the modifications and retest,
participants did not detect a difference in protein (x2= 10.29,
df = 6, p < 0.1), carbohydrate, sugar, fibre, salt, fat or energy
between the versions.
This study has described the successful design and testing of
near equally palatable versions of sweet and savoury foods with
protein contents of 10, 15 and 25% with respect to total energy.
Although considered to be closely similar in pleasantness (the
main aim of the study), the 10%, 15% and 25% versions of each food
could be distinguished by the subjects when presented side by
side. Simultaneous presentation of foods was used to directly test
changing the percent protein of foods. After modifications to five
foods, taste was the only sensory quality by which those foods
could be distinguished by more than 50% of Sydney participants.
Both cohorts were given the option to comment on the nature of
the sensory differences between versions of food items. These
comments were not included in the statistical analysis but indicate
that the participants were meticulous in assessing differences, as
they were asked to do. Although not tested, subjects may well have
claimed to be able to detect differences even if they had been
presented with 3 samples of the same food and version in one
sitting. Nonetheless, these are subtle distinctions that differ
markedly from, for example, texture specific satiety where studies
compare large differences in texture, e.g. an apple to applesauce
(Guinard & Brun, 1998).
Participants from the Kingston test group could not identify
differences among different versions of the foods in any nutritional
variable (protein, carbohydrate, fat or salt). In contrast, the Sydney
cohort correctly identified differences in protein content between
versions of the same foods, although prior knowledge was an
important factor in this. The Sydney ‘‘knowledgeable’’ group who
detected the higher protein versions more frequently than the
‘‘naı ¨ve’’ group were aware that the nutrient of interest was protein.
Analysis in which foods were grouped into meat and non-meat
categories showed that familiar high-protein ingredients made it
It is interesting that participants were able to detect sensory
differences in the foods manipulated with the experimental
protein mix but were generally unable to associate these changes
with an increase in protein. This suggests that humans find it
A.K. Gosby et al./Appetite 55 (2010) 367–370
difficult to detect protein per se, but can detect it in a form with
which they are familiar (e.g. meat). This may have potential
importance in a study of the protein leverage effect if the absence
of experiential cues delays the appearance of regulatory feeding
responses for protein. In contrast to protein, participants were
unable to perceive differences in the carbohydrate levels between
the food versions. A possible explanation for this may be that all
versions tested were relatively high-carbohydrate foods and
therefore the relatively small change in carbohydrate content
between the 10% and 25% protein versions (15% from 60% to 45%
carbohydrate) may have been difficult to detect. A change in
sweetness is likely to indicate a change in sugar content, therefore
the similarities in sugar content and sweetness between different
versions of the recipes may also help to explain why participants
were unable to perceive a difference in carbohydrate contents.
Results from a study testing for differences between high and low
fat diets also found that differences in carbohydrate content in
solid foods are difficult to detect (Stubbs et al., 2001). Participants
correctly perceived an increase in saltiness with an increase in
the 25% protein foods is due to the use of high-protein high salt
foods such as cheese and processed meats. After the modifications
and retest, participants did not perceive any nutritional differ-
ences. We chose this hierarchical approach, whereby foods were
modified as a function of how successfully their nutrient content
was disguised in the first trial, as a means of achieving a set of well
disguised foods for use in future experiments on the role of protein
feedbacks on energy intake. However, because foods were not
analysed separately and not all foods were retested, there is the
possibility that improvements in how well the retested foods were
disguised came from the modification or the nature of the specific
Modifications and retests were not done in Jamaica. It was
difficult to formulate additional modifications that would
improve foods further without broaching cultural expectations
about the foods. An interesting observation made during the taste
form of the presentation of ‘‘traditional’’ foods affected their
pleasantness ratings. The pleasantness rating given to a food by a
participant may be affected by expectations of the manner of
studies by using alternative names to those used for traditional
The subject groups represented were predominantly lean
university populations. This is the population that will initially
be used in studies testing the role of protein in energy intake.
Further testing of the foods before use in overweight/obese
populations would be necessary.
The current study provides a mix of complete protein sources
such as meat and cheese and an experimental protein mix to
manipulate the protein to energy ratio. This is in accord with
studies demonstrating a variety of protein sources may be
important for efficient postprandial utilisation (Lacroix et al.,
2006). Energy density was kept constant across the three versions
of each food to avoid changes in intake due to differing weights of
food (Rolls, 2009). A consistent variety across the 10%, 15% and 25%
sweet and savoury foods was also important in the design of an
experimental menu. This is because changes in intake that might
occur as a result of sensory-specificsatiety occurringnot as a result
of fullness (Brondel et al., 2009) but rather of boredom because
palatability drops as more of one food is consumed (Rolls, 1986)
and/or limited variety (Hetherington, Foster, Newman, Anderson,
& Norton, 2006) need to be ameliorated.
In summary, there were no pronounced or consistent direc-
tional differences in pleasantness, sweetness or savouriness
between manipulated versions of experimental foods. The
differences that remain, particularly in the perception of foods
with differing protein contents are likely to be smaller than in
other studies where no formal testing has been done. Studies like
the present are important in the design of experimental diets. They
not only provide an insight into nutritional perceptions of foods
and food palatability but also remove areas of uncertainty when
interpreting the outcomes of subsequent trials. Our aim next is to
use these foods in experiments in which the subjects are confined
for periods on diets comprising our test foods, such that they
experience a period of 10%, 15% or 25% protein diets. We will then
be able to test for effects of dietary protein on total energy intake,
as well as the influence of nutritional state on sensory responses to
foods varying in sensory quality (sweet or savoury).
Barkeling, B.,Rossner,S., &Bjorvell, H.(1990). Effects ofa high-protein meal(meat) and
a high-carbohydrate meal (vegetarian) on satiety measured by automated com-
puterized monitoring of subsequent food intake, motivation to eat and food
preferences. International Journal of Obesity, 14(9), 743–751.
Brondel, L., Romer, M., Van Wymelbeke, V., Pineau, N., Jiang, T., Hanus, C., et al. (2009).
Behavior, 97(1), 44–51.
Guinard, J. X., & Brun, P. (1998). Sensory-specific satiety: comparison of taste and
texture effects. Appetite, 31(2), 141–157.
Hetherington, M. M., Foster, R., Newman, T., Anderson, A. S., & Norton, G. (2006).
Understanding variety: tasting different foods delays satiation. Physiology Behav-
ior, 87(2), 263–271.
Hill, A. J., & Blundell, J. E. (1982). Nutrients and behaviour: research strategies for the
investigation of taste characteristics, food preferences, hunger sensations and
eating patterns in man. Journal of Psychiatric Research, 17(2), 203–212.
Lacroix, M., Bos, C., Leonil, J., Airinei, G., Luengo, C., Dare, S., et al. (2006). Compared
with casein or total milk protein, digestion of milk soluble proteins is too rapid to
sustain the anabolic postprandial amino acid requirement. American Journal of
Clinical Nutrition, 84(5), 1070–1079.
Protein and amino acid requirements in human nutrition. (2007). World Health
Organization Technical Report Series, 935, 1–265, back cover.
Rolls, B. J. (1986). Sensory-specific satiety. Nutrition Reviews, 44(3), 93–101.
Rolls, B. J. (2009). The relationship between dietary energy density and energy intake.
Physiology Behavior .
Rolls, B. J., Hetherington, M., & Burley, V. J. (1988). The specificity of satiety: the
influence of foods of different macronutrient content on the development of
satiety. Physiology Behavior, 43(2), 145–153.
Simpson, S. J., Batley, R., & Raubenheimer, D. (2003). Geometric analysis of macronu-
trient intake in humans: the power of protein? Appetite, 41(2), 123–140.
Simpson, S. J., & Raubenheimer, D. (2005). Obesity: the protein leverage hypothesis.
Obesity Reviews, 6(2), 133–142.
Stubbs, R. J., Mullen, S., Johnstone, A. M., Rist, M., Kracht, A., & Reid, C. (2001). How
covert are covertly manipulated diets? International Journal of Obesity and Related
Metabolic Disorders, 25(4), 567–573.
Weigle, D. S., Breen, P. A., Matthys, C. C., Callahan, H. S., Meeuws, K. E., Burden, V. R.,
et al. (2005). A high-protein diet induces sustained reductions in appetite, ad
libitum caloric intake, and body weight despite compensatory changes in diurnal
plasma leptin and ghrelin concentrations. The American Journal of Clinical Nutrition,
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.appet.2010.06.009.
A.K. Gosby et al./Appetite 55 (2010) 367–370