Satiety after preloads with different amounts of fat and carbohydrate: implications for obesity13

476 Amis J C/in Nuir 1994:60:476-87. Printed in USA. © 1994 American Society for Clinical Nutrition

Satiety after preloads with different amounts of fat andcarbohydrate: implications for obesity13

Barbara J Rolls, Sion Kim-Harris, Marian W Fischman, Richard W Foltin, Timothy H Moran, and

Susan A Stoner

ABSTRACT High intake of dietary fat may be key in both

the etiology and maintenance of obesity. Because a reduction in

the proportion of energy derived from fat will be accompanied

by an increase in the proportion of energy derived from carbo-

hydrate, this study compared the effects of these macronutnients

on eating behavior in obese and lean individuals. The effects of

difTerent amounts of fat and carbohydrate, covertly incorporated

into yogurt preloads, on subsequent food intake, hunger, and sa-

tiety were assessed. A group of I 2 normal-weight men, uncon-

cerned about eating and body weight (unrestrained), accurately

compensated for the energy in the pneloads regardless of the nu-

tnient composition. Other groups (ml = 12 per group), including

normal-weight restrained men and normal-weight and obese re-

strained and unrestrained females, did not show such orderly en-

ergy compensation; joule-for-joule, the high-fat preloads

suppressed intake at lunch less than did high-carbohydrate pre-

loads. These results suggest that a relative insensitivity to the

satiating effect of fat could be involved in the development and

maintenance of obesity. Am J C/in Nutr l994;60:476-87.

KEY WORDS Body weight, carbohydrate, energy regula-

tion, sex, hunger, obesity, fat, restraint, satiety

Introduction

High intake of fat may be important in both the etiology and

the maintenance of obesity. Obese individuals have an elevated

preference for fat in foods ( I , 2) and they consume a higher

proportion of fat in their diets than do normal-weight individuals

(3, 4). There is also evidence that they have an increased ten-

dency to deposit dietary fat as body fat (5). Because of the as-

sociation between dietary fat and obesity, a reduction in fat intake

is suggested as a key factor in most weight-control programs. To

aid this dietary change, the food industry has introduced a wide

range of no-fat, low-fat, and reduced-fat foods. The widespread

availability of reduced-fat foods raises many questions. First,

what impact will these foods have on diet composition and en-

ergy intake? Second, how will changes in diet composition affect

hunger and satiety? Third, do obese individuals and those who

are concerned with their food intake and body weight (ie, those

with high dietary restraint) respond to reduced-fat foods in the

same way as do normal-weight individuals? This study was de-

signed to address each of these questions.

A reduction in the proportion of energy derived from fat in the

diet is likely to be accompanied by an increase in the proportion

of energy derived from carbohydrate. Thus, it is particularly im-

portant to compare the effects of these two macronutrients on

food intake, hunger. and satiety. To date, the conclusions reached

in studies comparing fat and carbohydrate have varied widely.

Some studies suggest that the two nutrients have different effects

and that carbohydrate is more satiating than is fat (6-8). Other

studies indicate that fat and carbohydrate are equally satiating in

that changes in the energy content of foods associated with mod-

ification of either the fat or carbohydrate content are followed by

accurate energy compensation (9- 13). Although it is not yet

clear why these studies differ, it is likely that the results were

influenced by many variables, including the characteristics of the

individuals being tested, the types of foods used for both the

nutrient manipulation and the subsequent test meal, and the in-

terval between the pneload and the subsequent meal.

The experimental design should be determined by the question

being asked. For example, when, as in this study, the question is

whether the physiological effects of carbohydrate and fat affect

hunger and energy compensation differently, the manipulation

must be covert because both palatability (14) and beliefs about

fat content ( I 5 ) can affect subsequent intake. Thus, the sensory

properties ofthe manipulated foods must be matched. In addition,

the volume eaten and the protein content of the manipulated

foods should be the same because these variables also could in-

fluence intake. In this study we used yogurts, matched along all

of these dimensions, that varied systematically in fat and carbo-

hydrate contents.

Kissileff ( I 6) proposed that by systematically varying specific

dimensions of preloads (ie, energy on macronutrient content) the

satiating efficiency of that dimension can be calculated by relat-

ing intake in the subsequent test meal to the magnitude of the

preload along the chosen dimension. An important feature of our

study is that the relative satiating efficiencies of yogurts differing

in fat and carbohydrate contents were determined.

One feature of the experimental design that can influence in-

take after a preload is the time interval between the preload and

� From the Department of Psychiatry and Behavioral Sciences, The

Johns Hopkins University School of Medicine, Baltimore.2 Supported by the National Institute of Diabetes and Digestive and

Kidney Diseases grant DK-39 177.

1 Address reprint requests to BJ Rolls, Nutrition Department. Penn

State University. l()4 Benedict House, University Park. PA 16802.

Received November 24. 1993.

Accepted for publication April 4, 1994.

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

ROLLS ET AL 477

TABLE ISubjcct charactcristtcs’

Group Age (/ IBW2 BMI’

Eating inventory

EAT” Zung1CR4 DI’ Hunger

5. (7(

Males. normal-weight unrestrained 26 ± 4,S 1(34 ± 6 24 ± I 4 ± 2 4 ± 3 4 ± 4 7 ± 2’ 34 ± 4

Males. normal-weight restrained 28 ± 6 1(X) ± 7 23 ± I 12 ± 3” 5 ± 4 3 ± 2 14 ± 6 34 ± S

Females. normal-weight unrestrained 27 ± 5” 93 ± 9 21 ± 2 6 ± 2 5 ± S 6 ± 3 9 ± 4 36 ± 8

Fen#{236}ales. tiormal-weight restrained 27 ± 4,’ 97 ± 9 22 ± 2 13 ± 3” 7 ± 3 S ± 3 IS ± 8� 39 ± 7

Females. obese unrestrained 32 ± S 60 ± 32” 36 ± 7” 6 ± 2 It) ± S 7 ± 4 2 ± 8 41 ± 7

Females. obese restrained 33 ± 5 134 ± 22” 31 ± 5” 13 ± 3” 9 ± 4 4 ± 2 14 ± 5 38 ± 6

/ .f� ± SD: is = I 2 per group.

2 Percentage of ideal body weight based on the 1983 Metropolitan Life Tables (24).

./ Weight (kg)fheight2 (m).

4 Cognitive Restraint Scale.

6Disinhihition Scale.n Eating Attitudes Test.

7 Zung Self-Rating Scale.

,� Significantly different from obese restrained females. P < 0.05.

9Significantly different from females, normal-weight restrained, P < 0.05./0 Significantly different from unmarked values in same column, P < 0.05.

// Significantly different from every other group. P < 0.05.

the test meal. The interval of 30 mm used in this study was based

on a previous study ( 10) in which the time course of the effects

of the high-fat and high-carbohydrate yogurt pneloads on sub-

sequent lunch intake was determined. The effects of the two pre-

loads did not differ significantly and the interval at which the

best energy compensation was seen was 30 mm (other intervals

tested were 90 and I 80 mm).

In our previous study we tested only normal-weight unre-

strained men and women. It cannot be inferred that obese or

restrained individuals will respond in the same way to manipu-

lations of fat or carbohydrate. Although previous studies are not

in complete agreement ( I 7), data do indicate that both obese ( I 8,

19) and restrained individuals (20-22) are less responsive to the

energy content of preloads on meals than are normal-weight un-

restrained individuals. Of particular interest is the question of

whether overweight or restrained individuals are relatively in-

sensitive to the satiating effect of fat. If so, this could lead them

to consume more energy and a higher proportion of fat in their

diets (3, 4), increasing their susceptibility to obesity.

Thus, this study used a dose-response preloading paradigm to

determine the satiating efficiencies of different amounts of fat

and carbohydrate in yogurt. Further, the study assessed how sex,

dietary restraint, and obesity influence the responses to these

macnonutnients. A preliminary analysis of these data was pre-

sented previously (23).

Subjects and methods

Subjects

Volunteers were recruited from the Johns Hopkins Hospital

community through posters and advertisements. Individuals who

called the laboratory were given a standard telephone interview

assessing age, height, weight, smoking status, and food restnic-

tions. Those who did not regularly eat breakfast or were dieting,

taking medication (except oral contraceptives), pregnant, or

breast feeding were not included in the study. All study partici-

pants were healthy nonsmokers between the ages of 2 1 and 40 y

with no relevant reported food restrictions (Table I).

After the initial phone interview, potential subjects completed

screening questionnaires including the Eating Inventory, which

measures degrees of dietary restraint (25). the Eating Attitudes

Test, which detects symptoms of an eating disorder (26). and the

Zung Self-Rating Scale (27), which detects depression. Only

those individuals scoring < 30 on the Eating Attitudes Test and

< 50% on the Zung Scale were asked to participate in the study

(Table 1 ). Subjects were grouped according to sex, dietary re-

straint, and body weight to form the following six groups of I 2

subjects each: I) normal-weight, unrestrained males; 2) normal-

weight, restrained males; 3) normal-weight, unrestrained fe-

males; 4) normal-weight, restrained females: 5) obese, unre-

strained females; and 6) obese, restrained females. The number

of subjects for each experimental group was chosen to achieve

the desired power of � 0.80, with r levels of � 0.05. Previous

investigations in this laboratory have produced significant results

with similar numbers of subjects. Normal weight and obesity

were defined as between 85% and I 15% and > 120% of ideal

body weight, respectively, according to the Metropolitan Life

Insurance weight tables (24). Dietary restraint was defined as a

score of > 10 for females and > 9 for males on the Cognitive

Restraint Scale of the Stunkard Eating Inventory. Examination

of a large number of scores on the Cognitive Restraint Scale

administered during screening for previous studies in our labo-

ratory revealed that, overall, males tended to score lower than

females. Thus, the different criteria for men and women for di-

etary restraint reflect the differing distributions of scores.

Finally, potential subjects were asked to complete a trial ses-

sion before starting the experiment to ensure that they liked and

would eat the foods to be served in the test meal. During this

session, subjects rated the pleasantness of the lunch foods on I ()0-

mm visual analog scales (VAS), which were anchored on the left

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

478 FAT, CARBOHYDRATE, AND SATIETY

2 Analysis performed by Medallion Labs. Minneapolis.

by ‘ ‘not at all pleasant’ ‘ and on the right by ‘ ‘extremely pleas-

ant. ‘ ‘ Only those subjects rating � 80% of the foods at � 50 mm

were asked to continue in the study. They were then served the

self-selection lunch. This trial session served to familiarize sub-

jects with the experimental protocol and measures (ie, VAS) and

minimized any novelty effect that might confound the first ex-

penimental session.

Overall, the resulting groups of subjects accurately represented

the racial and ethnic distribution of the Johns Hopkins commu-

nity. However, there was some disparity in racial makeup among

groups; in particular, the groups of obese females had a higher

proportion of African-American individuals than did the other

groups.

All subjects signed a consent form explaining the experimental

procedure. To ensure that subjects were naive to the true purpose

of the study, the consent form stated that the study would ex-

amine the perceptions people had of different types of foods. The

consent form and experimental protocol were approved by the

Johns Hopkins Joint Committee on Clinical Investigation.

Foods and macronutrients

Five raspberry-flavored yogurts that varied in energy and mac-

ronutnient content were developed for use as preloads in this

study (Table 2). A medium-fat yogurt and a high-fat yogurt de-

rived most of their energy from fat in the form of a corn oil and

cream emulsion, with carbohydrate contents kept as similar as

possible. A medium-carbohydrate yogurt and a high-carbohy-

drate yogurt derived most of their energy from the carbohydrate

maltose, with fat contents kept as similar as possible. A control

yogurt contained a minimal amount of fat and carbohydrate and

was relatively low in energy. The protein content was kept as

similar as possible across the five yogurts, and the same volume

of yogurt was always given to the subjects. In a previous study

(10), males received 500-g and females received 350-g yogurt

preload with no adverse effects, such as nausea. The same

amounts of yogurt were used in this study. The high-fat and high-

carbohydrate yogurts were matched for energy and were more

energy dense than the medium-fat and medium-carbohydrate yo-

gulls (also matched for energy), which in turn were more energy

dense than the control yogurt (Table 2). The energy and macro-

nutrient contents of the yogurts were determined from sample

analyses performed by Medallion Labs, Minneapolis.

Lunch was a buffet-style meal that allowed subjects ad libitum

self-selection of a variety of foods (Table 3). The foods presented

in the lunch had varying amounts of fat, carbohydrate, and pro-

tein to allow subjects to vary not only energy intake but also the

proportions of macronutrients. All foods were weighed before

being served and reweighed after the subjects had eaten to obtain

the net amount consumed (g) of each food. Energy and macro-

nutrient intakes were calculated by using nutritional information

for the foods obtained from the manufacturer or Bowes and

Church ‘5 Food Values of Portions (‘ommonly Used (28).

TABLE 2Ingredients. macronutrient content, and energy density of the yogurt preloads’

Contents

Yogurt type

Control Medium CHO Medium fat High CHO High fat

Ingredients (g)

Colombo plain nonfat yogurt 270.2 240. 1 247.7 207.7 223.8

Gelatinsolution(l0()ggelatinand4O()gwater) 8.7 7.0 7.0 7.4 6.1

Aspartame/yogurt mix (3.5 g pure aspartame and 346.5 g Colombo

plain nonfat yogurt) 15.4 9.2 14.8 4.9 14.8

Raspberries, pureed 55.7 56.4 55.8 55.9 55.8

Oil and cream solution (120 g Mazola corn oil, 278.8 g Giant brand

heavy whipping cream, and I .2 g Tween 80) - - 24.7 - 49.5

Maltose stock solution (198.0 g Satin Sweet 65% maltose cornsyrup and 2.0 g Miragel 463 starch) - 37.3 - 74. 1 -

Energy from fat2

(kJ) 104.6 54.4 527.2 75.3 970.7

(kcal) 25.0 13.0 126.0 18.0 232.0(% ofenergy) 16 5 48 5 65

Energy from carbohydrate2

(kJ) 301.2 811.7 318.0 1209.2 297.1

(kcal) 72.0 194.0 76.0 289.0 71.0

(%ofenergy) 45 74 29 81 20Energy from protein2

(kJ) 267.8 230.1 251.0 209.2 225.9(kcal) 64.0 55.0 60.0 50.0 54.0

(% ofenergy) 39 21 23 14 15

Total energy2(kJ) 673.6 1098.3 1098.3 1493.7 1493.7

(kcal) 161.0 262.5 262.5 357.0 357.0

‘ All amounts are for one 350-g serving for females; males received 500 g. Mazola: Best Foods, Englewood Cliffs. NJ; Giant Food, Inc. Landover.

MD; Colombo Inc. Methuen, MA: Tween 80: Sigma Chemical Co. St Louis; Satin Sweet: Cargill Inc. Eddyville. IA: Miragel: EA Staley. Decatur.

IL.

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

ROLLS ET AL 479

TABLE 3Amounts served. energy density. and macronutrient content(pcr 1(8) g) of foods used in the self-selection lunch’

Food item Amount Energy Energy Fat Carbohydrate Protein

g k_I k(#{149}(l/ g g g

Eckrich oven roasted turkey breast 1St) 615.0 47.0 7.2 0.0 18.7

Tuna salad ( I 23. I g Sttrkist Fancy White

Albacore Tuna. 61.5 g Hellmann’s

tnayonnaise. and 15.4 g celery) 2(8) 1234.3 295.0 26.0 0.0 16.0

Doritos Nacho Cheese Tortilla Chips 60 2()66.2 493.8 24.7 63.5 3.5

Doritos Light Nacho Cheese Tortilla Chips 60 18()4.9 431.4 15.7 70.6 7.8

Hellmann’s mayonnaise 25 2987.4 714.0 78.6 0.0 1.4

Gulden’s spicy brown mustard 25 477.0 1 14.0 30.0 30.0 30.0

McCadam mild cheddar cheese. sliced 75 1644.3 393.0 32.1 3.6 25.0

Roman Meal whole wheat bread 130 1092.0 261.0 0.0 47.8 8.7

Iceberg lettuce. leaves 40 62.8 15.0 0.0 2.0 1.0

Tomato. sliced 1(8) 81.6 19.5 0.2 4.3 0.9

Cucumber, sliced SO 56.5 13.5 0.2 2.9 0.6

Naturally Fresh creamy Italian dressing 40 1924.6 460.0 47.3 10.0 0.7

Walden Farms reduced-calorie Italian dressing 35 2 13.4 S I .0 0.0 9.0 0.0

H#{228}agen-Dazs chocolate ice cream 2(8) 1 129.7 270.0 7.0 24.0 5.0

Dole raspberry sorbet I 68 535.6 128.0 0.0 32.6 0.4

Milky Way bars. sliced 122 1910.8 456.7 17.3 69.3 4.7

Red delicious apple. cored and sliced 16() 245.6 58.7 0.4 15.3 0.2

Water 5(8) - - - - -

‘ Eckrich, Oak Brook. IL: Starkist Seafood Co. Long Beach. CA: Hellmann�s: Best Foods. Englewood Cliffs. NJ: Doritos: Frito Lay. Inc. Atlanta:

Gulden’s American Home Foods, NY: McCadam Cheese Co. Inc. Heuvelton. NY: Roman Meal: Schmidt Baking Co. Baltimore: Naturally Fresh.

Atlanta: Walden Farms. Inc. Linden. NJ: The I-I#{228}agen-Dazs Co. lnc, Teaneck. NJ: Dole Packaged Foods Corp. San Francisco: Milky Way: M&M/

Mars. Hackettstown, NJ.

Procedures

The experiment had a balanced, within-subject, repeated-mea-

sures design. with each subject participating in six sessions. Each

subject served as his or her own control, and the order of the

presentation of the conditions was counterbalanced across sub-

jects. Subjects received a preload of a fixed volume of yogurt

before lunch during five of the six sessions. with the yogurts

covertly varying in energy and macronutrient contents from ses-

sion to session. On one occasion, subjects received no preload,

which served as the baseline condition. For each subject, sessions

were separated by at least 3 d and were conducted on weekdays

only, by appointment. Two females, one obese restrained and one

normal-weight restrained, were dropped from the study for fail-

ure to keep scheduled appointments.

Subjects were instructed to eat the same breakfast before each

session � 3 h before their session appointment. They were also

instructed not to eat any foods or drink any beverages except

water between breakfast and the session time. For each subject.

every session was scheduled for the same time every session day,

the subject’s normal lunch time. On arrival at the laboratory,

subjects presented food diaries in which they had recorded all

the foods they had eaten from dinnertime the previous night to

the time they came to the laboratory. The session proceeded only

when review of the food diaries indicated that the subjects had

been compliant. Subjects who were noncompliant were resched-

uled for another day.

Subjects were then seated in individual cubicles where they

completed a series of 100-mm visual analog scales (VAS #1 ) to

rate their degree of hunger, thirst, fullness, desire to eat, percep-

tion of how much they could eat (prospective consumption), and

nausea. The first hunger rating on the first session was considered

the baseline hunger rating. For each subsequent session, the hun-

ger rating on VAS #1 was compared with the baseline rating to

ensure similar degrees of hunger from session to session. If the

rating differed by > 15 mm from the baseline rating, the subject

was rescheduled for another day. Subjects whose ratings consis-

tently varied by > 15 mm (ie. those who had to be rescheduled

more than once) were dropped from the study. Four normal-

weight female subjects (two unrestrained and two restrained)

were dropped from the study for inconsistent hunger ratings.

Subjects then received a yogurt preload or no load, depending

on the counterbalanced condition. Subjects were instructed to

consume the entire preload within IS mm. The mean (±SD) time

to eat the yogurt, across all subjects and conditions, was 7 ± 2

mm. Yogurts were presented with a VAS booklet for subjects to

rate the sensory properties of the yogurt as well as their desire to

eat the yogurt. Subjects were instructed to complete this booklet

after their first spoonful of yogurt. These ratings were compared

across conditions to determine whether subjects perceived that

the yogurts were different.

After finishing the yogurt. subjects completed another series

of visual analog scales (VAS #2) to rate their hunger. etc. They

were then allowed to read in their cubicles until lunch was served,

30 mm after the preload was presented. In the no-load condition

subjects completed VAS #2 10 mm after VAS #1 and sat quietly

in their cubicles for 30 mm until lunch. Immediately before

lunch, subjects again rated their hunger, etc. on VAS #3. When

lunch was served, subjects were told that they could eat as much

as they wanted of any of the foods offered. Subjects were allowed

a minimum of 20 mm and a maximum of 30 mm for lunch. They

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

480 FAT, CARBOHYDRATE, AND SATIETY

were not allowed to read during lunch. After lunch, subjects corn-

pleted the session’s final set of hunger. etc. ratings (VAS #4) and

left the laboratory.

A debriefing questionnaire was completed by all subjects after

the completion of their final session. The questionnaire asked

subjects to state what they believed to be the purpose ofthe study

and to give comments relevant to the study.

Statistical (I?Ull\’SCS

Subject characteristics. Regression analyses were performed

to determine the significance of subject characteristics in deter-

mining lunch energy intake in each of the experimental condi-

tions. Forward stepwise regression analyses were performed by

using lunch energy intake and lunch macronutrient intake as the

dependent variables and age. sex, percent of ideal body weight,

each of the three Eating Inventory scale scores (cognitive re-

straint, disinhibition, and hunger), the Eating Attitudes Test

score, and the Zung Self-Rating Scale score as independent van-

ables. Regression analyses were performed for each experimental

condition separately.

Energy intake. Energy intake including and excluding the en-

ergy content of the preload was first analyzed by using repeated-

measures multivariate analyses of variance (MANOVA) with

group and condition as factors. A second set of analyses excluded

the no-load condition to test for significant differences between

the preload conditions.

When significant group-by-condition interactions were found,

data for each of the six groups were analyzed separately by using

repeated-measures analysis of variance (ANOVA). When any

statistically significant effects were encountered (Greenhouse-

Geisser probability � 0.05), post hoc comparisons ofmeans were

done by using Tukey’s honestly significant difference test to as-

certain which specific means differed.

The percent compensation at lunch for the preload energy was

determined for each group by subtracting the mean lunch intake

(kJ ) for each of the yogurt conditions from the mean lunch intake

in the no-preload condition (baseline intake), dividing the differ-

ence by the preload energy, and multiplying by 100.

Macromiutrient intake. Macronutrient intakes including and ex-

eluding the macronutrient content of the yogurt across the six

sessions were compared by using repeated-measures ANOVA

(BMDP, module 2V; BMDP Statistical Software, Los Angeles).

Macronutrient intakes were calculated as the percentage of en-

ergy consumed both at lunch and in total (lunch and yogurt)

derived from fat, carbohydrate, or protein.

Satiating efficiency. The satiating efficiency of the preloads

based on Kissileff’s model (16) was also examined. First, the

satiating efficiency across preloads for each subject was calcu-

lated separately by taking the negative of the slope generated by

plotting the preload energy against energy consumed at lunch,

for all six conditions. A satiating efficiency of I represents exact

compensation for the preload energy manipulations; > I repre-

sents a reduction of intake at lunch by > I I per preload joule

(ie, undereating or overcompensating), and < I represents a re-

duction of < I I for every preload joule (ie, overeating or un-

dercompensating). These satiating efficiencies were then corn-

pared across subject groups by using ANOVA.

In addition, satiating efficiencies were generated separately for

the fat yogurts (medium fat and high fat) and carbohydrate yo-

gurts (medium carbohydrate and high carbohydrate) by plotting

preload energy against energy consumed at lunch in the no-load,

control, medium-fat, and high-fat conditions and preload energy

against energy consumed at lunch in the no-load, control, me-

dium-carbohydrate, and high-carbohydrate conditions, respec-

tively. These satiating efficiencies were analyzed by using two-

way ANOVA with subject group and yogurt type as factors and

in two separate MANOVAs: one including only the females with

weight group and restraint as factors and another including only

the normal-weight groups with sex and restraint as factors.

Visual-analog-scale ratings. The self-reported baseline hunger

ratings (VAS #1 ) were analyzed by using repeated-measures

ANOVA with condition as the within-subject factor and sex,

weight group, and restraint as between-subject factors. In addi-

tion. hunger, thirst, and other analog ratings which were collected

on four different occasions during each session (VAS #1 -VAS

#4). were compared within groups across the five preload con-

ditions and across time by using two-way repeated-measures

ANOVA. These analyses assessed whether the manipulations of

the energy and macronutrient contents of the yogurts resulted in

different effects on hunger and satiety. To examine whether any

of the groups differed in their responses throughout the condi-

tions, absolute hunger ratings were compared in two MANOVAs

with condition and time (VAS #1 -VAS #4) as within-subject

factors: one including only the females with weight group and

restraint as factors and another including only the normal-weight

groups with sex and restraint as factors. Ratings of the yogurts

were compared across the five yogurt conditions by using one-

way repeated-measures ANOVA. Data were analyzed by using

SPSSIPC+ version 4.0. 1 (SPSS, Inc, Chicago) and BMDP, mod-

ule 2V statistical software for the personal computer. The Green-

house-Geisser probability level used to establish significance was

P < 0.05.

Results

Energy intake

When lunch intakes were compared across groups in a MAN-

OVA, with group and condition as factors, the main effect for

group was significant (F,5,,,,� = 3.66, P < 0.006). The mean lunch

intakes of normal-weight unrestrained males and obese unne-

strained females were significantly greater than those of the other

four groups (P < 0.05). As expected, the main effect of condition

was also significant (F1533�1 = 38. 1 1 , P < 0.001 ), such that across

groups, lunch intake in the no-preload condition was significantly

greater than lunch intake in every other condition (P < 0.05).

However, as shown in Figure 1, the interaction of group and

condition was also statistically significant (F125.3301 = 1.87, P

< 0.009), showing that groups responded differently across the

experimental conditions.

Lunch intakes were also compared across the five preload con-

ditions (ie, excluding the no-load condition) in a MANOVA with

group and condition as factors. Again, the main effect for group

was significant (F15,,,,,1 = 3.49, P < 0.008), such that the normal-

weight unrestrained males and obese unrestrained females con-

sumed more than the other female groups. The group-by-condi-

(ion interaction was also significant (Ft2o.2M1 = I .74, P < 0.03).Only the normal-weight, unrestrained males and females showed

differential lunch intakes across the preload conditions. Normal-

weight unrestrained males consumed significantly more in the

control yogurt condition than in either the high-carbohydrate or

the high-fat conditions (P < 0.05) whereas normal-weight Un-

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

A. Normal-weight unrestrained males

I± :�i.:;;;;.::.acJ

800 .�

No load Control Med CHO Med Fat High CHO High Fat

%C0MP� n/a 92% 99% 98% 99% 105%

�1.d

C

200

0

No toad Ccn�o1 Med CHO Med Fat High CHO High Fat%COMP. n/a 248% 1 22% 108% 1 35% 108%

E. Obese unrestrained females

No load Control Med CHO Med Fat High CHO High Fat%C0MP� n/a 140% 1 12% 76% 1 02% 63%

11 � 11 I I � �

No load Control Med CHO Med Fat High CHO High Fat

%COMP’ n/a 38% 1 14% 70% 80% 68%

C. Normal-weight unrestrained females

ROLLS ET AL 481

7531

6694

5858

5021

4184

3347

2510

1674

837

A

B. Normal-weight restrained males

No load Coo�o1 Med CHO Med Fat High CHO High Fat

%COMP: n/a 242% 121% 126% 104% 70%

D. Normal-weight restrained females

1800

1600

1400

1200

1000�

800

600

400

200

F. Obese restrained females

FIG I. Mean total intake (lunch, 0, and yogurt preload, #{149})and percent compensation (%COMP) in each condition.Dashed lines represent the energy intake required for 100% compensation. Means with same letters are significantly

different, (P < 0.05). CHO, carbohydrate. � ± SEM: n 12 per group.

1800

1600

1400

1200

1000�

800 .�

600

400

200

0

restrained females consumed significantly more after the me-

dium-fat yogurt than after the high-carbohydrate yogurt (P <

0.05).

Thus, all groups except the normal-weight unrestrained males

and females consumed essentially the same amounts after the

preloads, regardless of preload type. The normal-weight unre-

strained males, however, appropriately adjusted their intakes

down after the high-energy yogurts compared with their intake

after the control (low-energy) yogurt. As shown in Figure 1 . nor-

mal-weight unrestrained males exhibited remarkably accurate

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

482 FAT. CARBOHYDRATE. AND SATIETY

compensation at lunch for the preload energy in each of the five

preload conditions (98.4- 103.2c% of baseline). Other groups

compensated less well. Although the normal-weight unrestrained

kmales showed differential intake after some of the preload con-

ditions. their results were not as orderly as those of their male

counterparts. suggesting that they may have more difficulty in

making appropriate adjustments to energy manipulations.

When total intake (test meal plus yogurt energy) was examined

in a MANOVA with group and condition as factors. the main

effecl of group was significant (F,5,.,,1 = 5.33. P < 0.001): nor-

mal-weight males and obese unrestrained females consumed sig-

nificantly more energy than the other groups (P < 0.05). In

addition. the main effect of condition (F16 � = 6.72. P

< 0.0()l ) and the interaction of group-by-condition were also

significant(/�25 � = 1.65. P <� 0.03. Across all six conditions.

normal-weight unrestrained males. normal-weight restrained fe-

males. and obese unrestrained females showed no significant dif-

terences in total intake. Restrained males and normal-weight un-

restrained females. however. consumed significantly more total

energy in the no-load condition than in the control condition.

Thus. these groups showed an inappropriately large reduction in

lunch intake in the control condition. In addition. restrained

males ate significantly less total energy in the control condition

than in the high-carbohydrate and high-fat conditions (P < 0.05)

whereas normal-weight unrestrained females ate significantly

less total energy in the control condition than in the medium-fat

and high-fat conditions. Finally. obese restrained females showed

significantly higher total intake in the high-fat condition than in

the control condition (Fig I).

As detailed above. obese unrestrained females consistently

consumed more energy than the other female groups. However,

when total intakes were adjusted by body weight to the three-

fourths power (kilkg”4). as is often done in the nutritional sci-

ences. obese restrained females consumed significantly less en-

ergy than both groups of male subjects (P < 0.05), but adjusted

intakes of obese unrestrained females did not differ from those

of any other group (Fig 2).

Forward stepwise-regression analyses examining the signifi-

cance of subject characteristics in predicting energy intake

showed that the Eating Inventory Perceived Hunger Scale score

was an important variable in predicting intake. The Perceived

Hunger score was the only subject characteristic that was statis-

tically significant in every experimental condition. Subjects with

higher Perceived Hunger scores consumed more energy at lunch,

an effect seen regardless of the preload (P < 0.02).

%lacronutru’nt intake

The percent of energy consumed at lunch derived from fat and

carbohydrate did not differ significantly across the six conditions

for any group (Table 4). Consequently. the percentage of total

energy consumed in a session (lunch plus yogurt) coming from

fat and carbohydrate was significantly different across conditions

for all groups because the macronutrient content of the preloads

varied (P < 0.0()l ). Thus, within groups, although there is some

evidence for energy compensation. macronutnient-specific com-

pensation did not occur in any group. All significant differences

in macronutrient intakes were attributable to the manipulations

of the preloads. Comparison of fat, carbohydrate, and protein

intakes as percent of lunch energy among the six subject groups

revealed no statistically significant differences. Although not

listed in Table 4, protein intake remained relatively constant

across all six conditions for all groups. with and without the

preload protein included (�l5%).

Satiatielt�’ eJ’fl(i(’fl(l of preloads

When Kissileff’s model for the satiating efficiency of foods

was used ( 16). the preload energy was plotted against the energy

consumed at lunch to examine how satiating the yogurts were

within each subject group. Comparison of the satiating efficien-

cies of the two yogurt types (one in which the fat content was

varied and one in which the carbohydrate content was varied)

resulted in a significant main effect ofyogurt type (F117�1 = 9.08,

P < 0.004) but showed no significant differences across groups

nor an interaction effect for yogurt type and subject group. The

mean satiating efficiency for the yogurts that varied in fat content

was 0.82 whereas the mean satiating efficiency for the yogurts

varying in carbohydrate content was I .03. Thus, overall, the car-

bohydrate yogurts were more satiating than were the fat yogurts

(Fig 3.

To investigate this difference further, two additional repeated-

measures MANOVAs were performed. The first examined only

the female groups with the between-subjects factors of weight

group (normal vs obese) and restraint (restrained vs unrestrained)

and the within-subject factor of yogurt type. This analysis re-

vealed a main effect of yogurt type (F11.�1 = 7.56, P < 0.01)

with no other effects or interactions, such that in the females the

carbohydrate yogurts were more satiating than the fat yogurts. In

the second analysis. only the normal-weight groups were exam-

med with sex and restraint as between-subjects factors and yogurt

type as the only within-subject factor. Again, there was a signif-

icant effect of yogurt type (F11�.,, = 9.72, P < 0.004), with no

other significant effects or interactions. The main effect of re-

straint was nearly significant (F11��1 = 3.85, P = 0.056), such

that. overall, satiating efficiencies tended to be higher in the un-

restrained normal-weight groups than in the restrained normal-

weight groups. Comparison of the overall satiating efficiency of

the preloads across the six groups revealed no significant differ-

ences.

Visual-analog-scale ratings

Analysis of baseline hunger ratings (VAS #1 ) across the six

conditions and between groups revealed no significant effects or

interactions, showing that the subjects did not differ systemati-

cally from condition to condition, nor were there any differences

attributable to sex, weight group, or restraint; all groups started

at the same degree of hunger in each of the six sessions. Within

groups, absolute ratings of hunger. thirst, fullness, desire to eat,

prospective consumption, and nausea did not differ significantly

across the five preload conditions. As expected, there was a sig-

nificant effect of time in each of these analyses (P < 0.05), with

the exception of nausea (which remained unchanged), showing

that the subjects, including the obese females, are responsive to

the consumption of preloads and test meals with respect to hun-

ger, thirst, etc. When absolute hunger ratings throughout each

condition were analyzed between groups by using MANOVA,

there were no significant effects or interactions, with the excep-

tion of time. The effect of time was significant in both analyses:

in the normal-weight groups with sex and restraint as factors

(F1,,,71 57.07, P < 0.0()l ) and in the females with restraint and

weight-group as factors (F11.1521 = 64.29. P < 0.()01 ). Again, this

demonstrates that all groups responded similarly with respect to

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

MALES FEMALES

C’)

0)

-3

normal-wel ht normal-weight normal.wel ht normal-weight obese obese

unrestraln..d restrained unrestrain..d restrained unrestrained restrained

ROLLS ET AL 483

Co0

C.)

0)

Co0

FIG 2. Total intake across all conditions. A: Absolute intake; B: intake adjusted for body weight to the three-fourths

power (kJ/kg314). Means with same letters are significantly different. P < 0.05. � ± SEM: n 12 per group.

the consumption of the preloads and test meals, regardless of

condition (Fig 4).

Comparisons of the ratings of the five yogurt preloads pro-

vided evidence that the different yogurts were similar hedoni-

cally. The only statistically significant difference found was in

the rating of the creaminess of the yogurt; paradoxically. normal-

weight restrained males rated the creaminess of the high-carbo-

hydrate yogurt (76 mm) significantly higher than that of the high-

fat yogurt (61 mm) (P < 0.05). No differences among yogurts

were found in any of the ratings of the other five subject groups.

These data confirm that the energy and macronutrient manipu-

lations were not systematically detected by the subjects.

Debriefing

Subject responses to the debriefing questionnaire showed that,

whereas many subjects believed that intake after yogurt was be-

ing examined, none were able to determine that the yogurts van-

ied in energy and macronutrient contents. Consequently. no sub-

ject was able to state correctly the purpose of the study.

Discussion

The most remarkable finding in this study was the accurate

energy compensation after preloads shown by the normal-weight,

unrestrained men. This confirms and extends our previous study

in which we found that a similar group of young men accurately

adjusted their lunch energy intake 30 mm after both high-fat and

high-carbohydrate yogurts without a change in macronutnient in-

take. The intriguing question is how subjects are able to adjust

their lunch intake so soon after the preloads. The VAS ratings

indicate that the subjects could not distinguish the yogurts on the

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

Normal-weight Nonnal-weight Normal-weight Normal-weight Overweight OverweightUnrestrained Restrained Unrestrained Restrained Unrestrained Restrained

484 FAT. CARBOHYDRATE. AND SATIETY

TAI3I.E 4

Macronutrient i ntake from lunch’

Condition

Norrnal-wei gIlt males Nortnal-weig fit females Overweigh t females

Unrestrained Restrained Utirestrai tied Restrained tJnrestrai tied Restrained

F�tt

No load 45 + 2 47 :t I 46 ii 2 50 t I 54 if I 49 ± 2

Control 48 � 2 41 � 4 49 it 3 50 � 3 53 � I SI � 2

Medium carbohydrate 48 � 2 43 It 2 44 �t 2 49 t 2 51 � I 48 � 2

Mediurnfat 44’2 46t2 SOit2 52t2 54i�i2 50±3

High carbohydrate 49 � 3 42 ± 3 45 ± 2 47 ± 3 52 ± 2 50 ± 3

High fat 45 :� 3 46 ± I 46 ± 3 53 ± 3 50 ± 2 51 ± 3

Carbohydrate

Noload 39�2 37±1 39±3 35±2 27±2 34�2

Control 36 t 3 47 ± 6 36 ± 4 33 ± 3 30 ± 2 27 ± 3

Medium carbohydrate 34 ± 3 40 ± 2 40 ± 3 37 ± 3 31 it 2 32 ± 2

Medium fat 41 ± 4 38 ± 4 35 ± 3 30 ± 3 29 ± 2 29 ± 4

Highcarbohydrate 33i�3 45±5 37±2 35±4 31±2 28±4

Highiat 38if4 35±3 39±4 30±3 34±2 28±3

‘ j: � SEM: ii = I 2 per group. Protein intake stayed relatively constant regardless of condition.

basis of the sensory properties. In addition, the yogurt preloads

were all the same volume and had similar protein contents. Thus,

these results indicate that these subjects have mechanisms for the

rapid detection of the energy content of foods.

We (29) recently completed a study that shows that some mdi-

viduals have preabsorptive satiety mechanisms that respond rapidly

to both carbohydrate and fat. When infused intraga.stnically over 15

mm with 2()92 Id (500 kcal, the same amount as the preloads in the

present study) of either fat (Liposyn: Abbott Laboratories, Abbott

Park, IL) or carbcthydrate (dextrose) and offered lunch after 30 mm,

young, normal-weight. unrestrained male subjects accurately com-

pensated for the energy in the infusate. Compensation was less ac-

curate when the nutrients were infused at a slower rate intragastni-

cally (over 3.5 h), and no compensation was seen when the nutrients

were infused intravenously (30). Similar results showing accurate

energy compensation after carbohydrate or fat gastric preloads have

been seen in other species (31).

We hypothesize, based on the present study and previous find-

ings ( 10, 32), that there are multiple mechanisms for satiety that

may be active at different times after food intake. One set of

mechanisms is engaged rapidly. ie, 20-30 mm after the start of

eating. and therefore likely depends on preabsorptive changes.

These preahsorptive mechanisms, including gastric distension,

gastric emptying. and nutrient stimulation of the gastrointestinal

tract, may account for the results in this study. In many preload-

ing studies food was not offered this soon after a preload and

consequently there was no opportunity to observe rapid satiety.

In general. in studies in which food was offered 30 mm or sooner

after a preload to normal-weight, unrestrained males. energy

compensation was observed (10. 33, 34). Many other studies of-

fered food between I and 4 h after the preload and in these studies

poor energy compensation is observed (35-37). Furthermore,

depending on when subsequent food is offered, late satiety also may

not be recorded. For example. in a previous study in which the

C.)Ca)0

Lu

C

Co

COC/)

FIG 3. Satiating efficiencies of yogurts with various carbohydrate (U) contents compared with the satiatingefficiencies of yogurts with various fat (0) contents for each subject group. 5� ± SEM: n = 12 per group.

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

A. Normal-Weight Unrestrained Males B. Normal-Weight Restrained Males

100

90

80

70

60

0 10 20b.aclirie �x..L.yog�at

Time (mm)

C. Normal-Weight Unrestrained Females

20 30 40 50 60

Time (mm)

D. Normal-Weight Restrained Females

EE

C

CCC

E

CCC

C’CCC

EE

CCC

.=

‘C� 40

CCC

� 30

20

10

100

90

80

70

60

so

40

30

20

EE

C

IVDCC

EE

CC0

aI..a.C’DCC

EE

DCC

a

VDCC

Time (mm) Time (mm)

FIG 4. Mean visual analog scale hunger ratings (mm) for each of the five preload conditions: control, -4-- : medium

carbohydrate, -0-- ; medium fat, -U-- ; high carbohydrate, -0-- ; and high fat, ---- . Ratings are at four timepoints: baseline, or before preload; after preload; before lunch; and after lunch. ii 12 per group.

Time (mm)

E. Obese Unrestrained Females

10

Time (mm)

F. Obese Restrained Females

10 20 30 40 50 60 10 20 30 40 50 60

ROLLS ET AL 485

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

486 FAT, CARBOHYDRATE, AND SATIETY

energy manipulation was at breakfast, compensation was not ob-

served at lunch, 3.5 h later. but was observed at dinner, 7 h later:

thus, because no meal was offered shortly after breakfast, rapid sa-

tiety was missed, but compensation was observed in the evening

meal (32). Had this study not offered an evening meal, this corn-

pensation may have been missed altogether. The effects on foodintake at these later times probably depend on ingested foods or

nutrients becoming metabolic fuel. Oxidation ofthe fuel by the liver

may be one mechanism involved in this type of satiety (38).

An important feature of the present study was that the dose-

response preloading paradigm allowed the relative satiating ef-

ficiencies of the yogurts that varied in fat and carbohydrate con-

tent to be calculated. The normal-weight, unrestrained males re-

sponded similarly to the variations in fat or carbohydrate,

supporting the conclusion reached in our previous studies ( 10-

1 2) that the two nutrients do not differ in their effects on satiety.

However, when analyses were performed over all the groups to-

gether, fat in the yogurt was significantly less satiating than car-

bohydrate. In the restrained and overweight subjects there was a

tendency for the satiating efficiency of fat in yogurt to be < I .0,

indicating that in these groups there is some insensitivity to the

satiating effect of fat. It is possible that individuals who are over-

weight or concerned with their weight have different physiolog-

ical responses to fat in foods than do normal-weight unrestrained

individuals. In view of the importance of dietary fat in obesity

( 14). this preliminary observation requires confirmation.

Although we saw compensation for the different energy con-

tent of the preloads in some groups, there was no evidence that

the macronutrient composition of the lunch was adjusted in re-

lation to the composition of the preloads. The percentage of fat

or carbohydrate in the lunches did not vary significantly across

conditions. Thus, if subjects consumed the high-fat preload, the

overall proportion of fat in the preload and lunch combined was

higher than in the other conditions. These results confirm previ-

ous findings (10- 1 2, 32) and indicate that the consumption of

reduced-fat foods instead of the full-fat versions can help to re-

duce the daily percentage of energy from fat. Note, however, that

these results are in response to covert manipulation of fat content;

information about the fat content of foods may affect intake. In

a recent study the perception of the fat content of the test food

was manipulated (15). Normal-weight women who received yo-

gurt labeled as low fat consumed more energy during a subse-

quent lunch than after an isoenergetic yogurt labeled as high fat.

Thus, the benefit of consuming low-fat foods will depend on the

motivation of the consumers; if they use the consumption of a

low-fat food as an excuse to eat other high-fat foods, there may

be no overall reduction in fat intake.

Failure to separate subjects according to sex, dietary restraint,

and body weight could be part of the explanation for inconsis-

tencies among prior studies. None of the other groups tested

showed the orderly energy compensation observed in the normal-

weight, unrestrained men. The accurate energy compensation

shown by the young. normal-weight, unrestrained men confirms

our findings with similar subjects after other experimental ma-

nipulations of the energy and macronutnient contents of preloads

or meals (10- 1 2, 32). Thus, if the goal of an experiment is to

study the normal physiology underlying intake regulation, this

group of subjects is the optimal choice. It is of interest that, even

as children, boys show better energy compensation after preloads

than do girls (39). Future studies will determine whether sex dif-

ferences persist throughout the aging process.

In this study the quantity of yogurt administered to subjects

was dependent on the subject’s sex because in our laboratory

females typically consume 70�Yc of the energy intakes of males.

Thus, the preloads were adjusted for sex by that amount. As we

have seen previously. the normal-weight unrestrained and re-

strained females consumed �70% of the amount consumed by

the equivalent group of males in the no-load condition. This re-

lationship also held for the obese restrained group, but obese

unrestrained females ate almost as much as did unrestrained

males. Thus, this group received less energy in the preloads in

relation to their baseline intakes than the other groups. It is pos-

sible that female groups showed less precise energy compensa-

tion than unrestrained males because the energy differences be-

tween the preloads were smaller, although clearly this argument

would not apply to the restrained males.

The restrained subjects in this study ate similar amounts at

lunch regardless of the content of the preloads. We obtained a

similar result in a previous preloading study in which females

actively dieting to lose weight ate the same amount at lunch re-

gardless of the content of the preloads (33). Many studies have

indicated that individuals with high dietary restraint may impose

rules on their eating behavior that override physiological cues

about appropriate amounts to eat (40). Another possibility that

should be investigated is that restrained individuals, or those ac-

tively dieting, impose such rules about the amount they should

eat as a way of coping with impaired hunger mechanisms.

Categorizing subjects according to their degrees of dietary re-

straint yielded data that suggested different eating responses

within the normal-weight males and within the obese females,

which may have been obscured if dietary restraint had not been

taken into consideration. Restrained males have not been inves-

tigated previously but should be included in future studies be-

cause they showed poor energy compensation in contrast to the

unrestrained males. Note that this group was difficult to recruit

because young, normal-weight males tend to be unrestrained. We

found it even more difficult to recruit both unrestrained and re-

strained obese males and thus were unable to run these subjects

in this study. Obese women are not usually categorized according

to their restraint score, but this study indicates that restrained and

unrestrained obese women do behave differently; unrestrained

obese women ate more food at lunch and were significantly heav-

ier than the restrained obese women. Unrestrained obese females

appeared to show better energy compensation than the restrained

group, but this may in part be due to differences in their baseline

intakes. Note that many of the restrained and unrestrained obese

women were African Americans, which probably reflects their

greater cultural acceptance of obesity (4 1).

A major challenge in studies of food intake is to understand

why some individuals show better regulation than others. This

study indicates some characteristics of subjects that should be

considered when assessing energy compensation; these include

sex and dietary restraint. A clear effect of body weight indepen-

dent of restraint was not seen in this study.

In summary, this study shows that some individuals (ie, nor-

mal-weight. unrestrained males) can show accurate energy corn-

pensation after preloads that vary in fat and carbohydrate con-

tents. This compensation is seen 30 mm after the start of eating

the preloads and therefore probably depends primarily on preab-

sorptive satiety mechanisms or receptors in the gastrointestinal

tract. Further studies are needed to define the mechanisms in-

volved in this rapid, accurate energy compensation. The other

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from

ROLLS ET AL 487

groups tested showed less precise energy compensation and ap-

peared to have a relative insensitivity to the satiety value of fat.

Some investigators have suggested that dietary fat has only a

weak effect on satiety (6). Although the results of this study

indicate that in some individuals fat had a significantly lower

satiating efficiency than carbohydrate, this effect was small and

does not suggest that energy from fat is undetected by satiety

mechanisms. It will be important to determine whether this rel-

ative insensitivity to fat is associated with the overconsumption

of fat and energy and whether it could be important in the eti-

ology or maintenance of obesity. 0

We thank Stephen P Kaufman and Yoplait Research & Development

of General Mills. Inc. for their assistance in the development of the yo-

gurt preloads. We also thank David J Shide for his assistance in writing

the text and Jane Gamer. Paul Pirraglia. and Amy McNelis for their

assistance in data collection.

References

I . Drewnowski A. Brunzell JD, Sande K, Ivenius PH. Greenwood

MRC. Sweet tooth reconsidered: taste responsiveness in human obe-

sity. Physiol Behav l985:35:617-22.

2. Mela DJ. Sacchetti DA. Sensory preferences for fats: relationships

with diet and body composition. Am J Clin Nutr 1991:53:908-15.

3. Miller WC, Lindeman AK, Wallace J, Niederpruem M. Diet com-

position, energy intake. and exercise in relation to body fat in men

and women. Am J Clin Nutr 1990:52:426-30.

4. Strain GW, Hershcopf RJ, Zumoff B. Food intake of very obese

persons: quantitative and qualitative aspects. J Am Diet Assoc1992:92:199-203.

5. Thomas CD. Peters JC, Reed GW, Abumrad NA. Sun M. Hill JO.

Nutrient balance and energy expenditure during ad libitum feeding

of high-fat and high-carbohydrate diets in humans. Am J Clin Nutr

1992:55:934-42.

6. Blundell JE, Burley Vi. Cotton JR. Lawton CL. Dietary fat and the

control of energy intake: evaluating the effects of fat on meal size

and postmeal satiety. Am J CIin Nutr 1993:57(suppl):772S-8S.

7. Rolls BJ, Hethenington M. Burley VJ. The specificity of satiety: the

influence of foods of different macronutnient content on the devel-

opment of satiety. Physiol Behav 1988:43:145-53.

8. Gil KM. Skeie B, Kvetan V. Askanazi J, Friedman Ml. Parenteralnutrition and oral intake: effect of glucose and fat infusions. JPEN1991:15:426-32.

9. Geliebter AA. Effects of equicaloric loads of protein. fat and car-

bohydrate on food intake in the rat and man. Physiol Behav

I 979:22:267-73.10. Rolls BJ, Kim 5, McNelis AL, Fischman MW, Foltin RW, Moran

TH. Time course of effects of preloads high in fat or carbohydrate

on food intake and hunger ratings in humans. Am J Physiol1991 :26():R756-63.

I I. Foltin RW, Fischman MW, Moran Th. Rolls BJ, Kelly Th. Caloric

compensation for lunches varying in fat and carbohydrate content byhumans in a residential laboratory. Am J Clin Nutr l9�X):52:969-80.

12. Foltin RW, Rolls BJ, Moran TH, Kelly TH. McNelis AL, Fischman

MW. Caloric, but not macronutnient, compensation by humans for

required-eating occasions with meals and snack varying in fat and

carbohydrate. Am J Clin Nutr 1992:55:331 -42.

I 3. Caputo FA. Mattes RD. Human dietary responses to covert manip-

ulations of energy, fat, and carbohydrate in a midday meal. Am J

Clin Nutr l992;56:36-43.

14. Rolls BJ, Shide DJ. The influence of dietary fat on food intake and

body weight. Nutr Rev 1992:50:283-90.

15. Rolls BJ. Shide DJ. Hoeymans N. Jas P. Nichols A. Information

about fat content of preloads influences energy intake in women.

Appetite 1992:19:2 13 (abstr).

16. Kissileff HR. Satiating efficiency and a strategy for conducting food

loading experiments. Neurosci Biobehav Rev 1984:8:129-35.

I 7. Spitzer L. Rodin J. Effects of fructose and glucose preloads on sub-

sequent food intake. Appetite 1987:8:135-45.

18. Lissner L. Levitsky DA. Strupp BJ. Kalkwarf HJ. Roe DA. Dietary

fat and the regulation ofenergy intake in human subjects. Am J Clin

Nutr I 987:46:886-92.

19. Spiegel TA. Shrager EE. Stellar E. Respnses of lean and obese

subjects to preloads. deprivation. and palatability. Appetite

I989: I3:45-69.

20. Herman CP. Mack D. Restrained and unrestrained eating. J PersI 975:43:647-6().

2 1 . Ruderman AJ. Dietary restraint: a theoretical and empirical review.

Psychol Bull 1986;99:247-62.

22. Lowe MR. Whitlow JW, Bellwoar V. Eating regulation: the role of

restraint. dieting. and weight. tnt J Eat Disord l99l : 10:461 -7 I.

23. Rolls BJ. Kim S. Fischman MW, Ct al. Food intake following pre-

loads of fat and carbohydrate in normal weight and obese individ-

uals. Obes Res 1992:1:1 IS (abstr).

24. Metropolitan Life Insurance Company. l983 Metropolitan height

and weight tables. Stat Bull 1983:64:2-9.

25. Stunkard AJ, Messick S. The three-factor eating questionnaire to

measure dietary restraint, disinhibition, and hunger. J Psychosom

Res 1985:29:71-83.26. Garner DM. Gantinkel PE. The Eating Attitudes Test: an index of

the symptoms of anorexia nervosa. Psychol Med 1979:9:273-80.

27. Zung WWK. Zung Self-Rating Depression Scale and Depression

Status Inventory. In: Sartorius N. Ban TA, eds. Assessment of de-

pression. Berlin: Springer-Verlag. I 986:22 I -31.

28. Pennington JAT. Church HN. Bowes and Church’s fixxi values of por-

tions commonly used. l4th ed. Philadelphia: JB Lippincott. Co. 1985.

29. Shide DJ, Caballero B, Reidelberger R. Rolls B. Accurate energycompensation for intragastric and oral nutrients in lean males. Am

J Clin Nutr (in press).30. Shide DJ. Caballero B. Friedman M. Moran TH, Rolls BJ. Differ-

ential effects of intravenous or oral nutrients Ofl caloric intake in

healthy humans. Am J Clin Nutr l992:55suppI): I 19 (abstr).3 I . McHugh PR, Moran TH. Calories and gastric emptying: a regu-

latory capacity with implications for feeding. Am J Physiol

I 979:236:R254-60.

32. Rolls BJ. Pirraglia PA. Jones MB. Peters JC. Effects of olestra. a

non-caloric fat substitute, on daily energy and fat intake in lean men.

Am J Clin Nutr 1992:56:84-92.

33. Rolls BJ, Gnizak N. Summerfelt A, Laster II. Food intake in dieters

and non-dieters following a liquid meal containing medium chaintriglycerides. Am J Clin Nutr 1988:48:66-7 I.

34. Rolls BJ, Fedoroff IC, Guthrie J, Laster U. Foods with different

satiating effects in humans. Appetite 1990:15:1 15-26.

35. de Graaf C, Hulshof 1, Westrate JA, Jas P. Short-term effects of

different amounts of protein, fats, and carbohydrates on satiety. Am

J Clin Nutr 1992:55:33-8.

36. Rolls BJ, Hetherington M. Burley VJ. Sensory stimulation and en-

ergy density in the development of satiety. Physiol Behav

l988:44:727-33.

37. Louis-Sylvestre J. Tournier A, Verger P. Chabert M. Delorme B.

Hossenlopp J. Learned caloric adjustment of human intake. Appetite

1989:12:95-103.

38. Friedman Ml. Metabolic control of calorie intake. In: Friedman MI.

Tordoff MG. Kare MR. eds. Chemical senses. Vol 4. Appetite and

nutrition. New York: Marcel Dekker, Inc. 1991:19-65.39. Johnson SL. Birch LL. Gender differences in compensation patterns

of preschool boys and girls. lnt J Obes I992:15(suppl):4l (abstr).

40. Wardle J. Cognitive control of eating. J Psychosom Res

l988:32:6()7- 12.

41. Kumanyika S. Wilson JF. Guilford-Davenport M. Weight-related

attitudes and behaviors of black women. J Am Diet Assoc

l993:93:4l6-22.

by guest on Septem

ber 8, 2015ajcn.nutrition.org

Dow

nloaded from