AJCN North Carolina Research Campus
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Marmonier, C.
Right arrow Articles by Louis-Sylvestre, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Marmonier, C.
Right arrow Articles by Louis-Sylvestre, J.
Agricola
Right arrow Articles by Marmonier, C.
Right arrow Articles by Louis-Sylvestre, J.
American Journal of Clinical Nutrition, Vol. 76, No. 3, 518-528, September 2002
© 2002 American Society for Clinical Nutrition

Original Research Communication

Snacks consumed in a nonhungry state have poor satiating efficiency: influence of snack composition on substrate utilization and hunger1,2

Corinne Marmonier, Didier Chapelot, Marc Fantino and Jeanine Louis-Sylvestre

1 From the Laboratoire de Physiologie du Comportement Alimentaire, Ecole Pratique des Hautes Etudes, Bobigny, France (CM, DC, and J L-S), and the Groupe Nutrition et Métabolisme Humain, Institut Européen des Sciences du goût et des Comportements Alimentaires, Dijon, France (MF).

2 Address reprint requests to J Louis-Sylvestre, Laboratoire de Physiologie du Comportement Alimentaire, Faculté Léonard De Vinci, 74 rue Marcel Cachin, 93017 Bobigny Cedex, France. E-mail: comp-alim{at}smbh.univ-paris13.fr.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Background: Several epidemiologic studies suggest that snacking may play an etiologic role in obesity.

Objective: We assessed the behavioral and metabolic consequences of a high-carbohydrate (HC) or high-protein (HP) snack consumed 215 min after lunch, thereby investigating ways that snacking in a nonhungry state could be involved in the etiology of obesity.

Design: Eight lean young men attended 3 sessions (basal, HP snack, and HC snack) in a counterbalanced order with 2 wk between sessions. During all sessions, subjects were time-blinded while we measured the temporal pattern of plasma glucose, insulin, and fatty acid concentrations; hunger ratings; substrate oxidation; and energy expenditure from 215 min after the beginning of lunch until the spontaneous dinner request.

Results: Compared with the basal (no snack) session, the HP snack delayed the spontaneous dinner request (by 38 ± 16 min, P < 0.05) but the HC snack did not. Energy and macronutrient intakes at dinner were unaffected by both snacks. After the HP snack, plasma fatty acid concentrations were lower (P < 0.05), but glucose and insulin were unchanged compared with the basal session. After the HC snack, plasma glucose and insulin concentrations were higher and plasma fatty acid concentrations were lower than those in the basal session (P < 0.05 for both). Both snacks promoted carbohydrate utilization (P < 0.05), and the HC snack depressed fat oxidation (P < 0.05).

Conclusion: This study showed that a snack consumed in a satiety state has poor satiating efficiency irrespective of its composition, which is evidence that snacking plays a role in obesity.

Key Words: Snack • appetite • satiety • substrate oxidation • metabolic response • hormonal response • behavioral response • insulin • glucose • fatty acids • obesity • weight gain


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Several epidemiologic studies suggest that snacking may play an etiologic role in obesity (1, 2). In 2 different studies, we investigated the behavioral and metabolic consequences of consuming 1-MJ snacks of various compositions (3, 4). In the first study, time-blinded subjects underwent continuous blood sampling to allow precise determination of plasma glucose, insulin, and fatty acid changes from lunch until the dinner request (3). Under these conditions, we found that a well-balanced snack consumed at various times during the intermeal interval (ie, when subjects were in a satiety state) did not influence the satiating power of the lunch; neither the time of the dinner request nor the energy intake at that meal were affected. Substrate and hormone profiles were indicative of carbohydrate storage when the snack was consumed early in the intermeal interval and of fat storage when the snack was ingested at later times.

In the second study, under the same conditions of time-blindness, we showed that snacks consumed 2 h before the time when dinner was requested in the basal session (ie, when no snack was provided) delayed the dinner request by 25, 60, and 34 min with high-fat, high-protein, and high-carbohydrate snacks, respectively (4). None of these snacks were associated with changes in energy or macronutrient intakes at dinner. Considering that consumption of 1 MJ food energy by young, healthy men should provide sufficient energy to match basal energy expenditure for 200 min, we concluded that these snacks had poor satiating efficiency.

These results were interesting, but were difficult to interpret for several reasons. First, the real-life, isoenergetic snacks that we used differed not only in their macronutrient composition but also in their energy density. Second, without metabolic and hormonal data, we could only assume that the weak satiating power of a snack consumed in a satiety state could be explained by the disposal of metabolic fuels.

The aim of the present study was to determine the behavioral and metabolic effects of a high-carbohydrate (HC) or high-protein (HP) snack offered 215 min after lunch to subjects who were accustomed to consuming 3 meals/d and no snacks. Measuring the duration of satiety is a useful way to determine the satiating power of food intake (5); therefore, we conducted the present study with time-blinded subjects who spontaneously requested the meal after the snack, which was dinner. We recorded the latency to the dinner request (ie, the interval of time from the beginning of lunch until the time when the subject requested dinner spontaneously), and we recorded the subject’s food intake at dinner. Also, we monitored various metabolic variables (plasma glucose, insulin, and fatty acid concentrations and substrate oxidation) continuously between the snack and dinner.


   SUBJECTS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Subjects
After the procedures were approved by the ethical committee of the Dijon University School of Medicine, subjects were recruited through advertisements posted at the School of Medicine. We excluded smokers, trained athletes, and persons who reported that they either skip meals occasionally, eat at unusual times, frequently eat > 3 meals/d, have food allergies, or have aversions to the foods that would be provided during the experiment. Potential subjects were also excluded if they reported a personal or family history of diabetes or another metabolic disease, a change in body weight during the 3 y before the study, or use of any medications. Subjects also kept a 7-d food diary, and any subjects consuming > 35% of energy as fat were excluded.

Eight subjects participated in this experiment. They all gave their written, informed consent and were financially compensated for completing the study. As shown in Table 1Go, the mean age was 22.6 ± 0.7 y (range: 20–25 y) and the body mass index values were within the normal range. The subjects’ scores on the Eating Attitude Test (6) and the Three-Factor Eating Questionnaire (7) indicated that they were not inclined to control food intake cognitively.


View this table:
[in this window]
[in a new window]
  		TABLE 1 Subject characteristics1
 
Study design
We used a within-subject design in which each subject served as his own control. Each subject attended 3 sessions (basal, HP, and HC) scheduled 2 wk apart in a counterbalanced order (ie, a Latin square). During the sessions, subjects were deprived of time cues to eliminate as much as possible their responses to habitual (time-determined) meal patterns. This allowed us to observe the subjects’ responses to primarily physiologic cues.

The basal session was designed to determine the following in each subject: 1) the temporal pattern of hunger ratings from lunch until the spontaneous dinner request, 2) the latency of the dinner request, 3) the temporal patterns of substrate oxidation and energy expenditure from 215 min after the beginning of lunch until the spontaneous dinner request, and 4) the temporal patterns of the plasma glucose, insulin, and fatty acid concentrations. The HP and HC sessions were designed to investigate the behavioral and metabolic consequences of consuming snacks of different compositions 215 min after the beginning of lunch.

Foods
The lunch was a 2-course meal consisting of spaghetti bolognaise and vanilla-flavored dessert cream (Table 2Go). For each session, each subject received a lunch with an energy content that corresponded to their usual lunch as calculated from their individual 7-d food diaries.


View this table:
[in this window]
[in a new window]
  		TABLE 2 Food items served at the lunch and free-choice food items offered at the ad libitum buffet-style dinner
 
The snack was either high in protein (971 kJ, 64% of energy from protein) or high in carbohydrate (996 kJ, 66% of energy from carbohydrate) (Table 3Go). The HP snack consisted of cooked chicken breast (160 g) served with a low-fat dressing (50 g) and the HC snack consisted of rice pudding (200 g). Special care was taken to ensure that the fat content and energy density of the HC and HP snacks were the same and that they only differed in terms of carbohydrate and protein content.


View this table:
[in this window]
[in a new window]
  		TABLE 3 Composition of the high-protein (HP) and high-carbohydrate (HC) snacks
 
The ad libitum dinner was a buffet-style meal. The foods offered were typical French food items, chosen to allow free selection of foods differing in macronutrient composition (Table 2Go).

Hunger ratings
Subjects were asked to assess their hunger sensations by rating them on a 100-mm visual analogue scale with the French words for none and extreme at the 2 ends. Similarly, subjects rated the pleasantness of each snack on a visual analogue scale with the French words for extremely unpleasant and extremely pleasant at the 2 ends of the scale.

Blood sampling and plasma assays
To obtain blood samples, we used a specially designed double-lumen catheter (MTB Medezintechnik, Amstetten, Germany) fit into a 21-gauge indwelling cannula inserted into an antecubital vein. Heparin was added to the catheter by using a peristaltic pump to push a sterile heparin-saline solution (200 U/mL) through the distal lumen of the catheter to the tip of the cannula at a flow rate of 24 µL/min. Heparin-containing blood was withdrawn continuously through the proximal lumen of the cannula at a flow rate of 400 µL/min. Thus, heparin-containing blood was sampled throughout the session without any infusion of heparin into the vein (8). Blood samples were collected in tubes every 5 min. The transit time of this continuous sampling line was {approx}10 min; this lag time was measured precisely for each test in every subject and was accounted for in the data analysis. Blood samples were centrifuged immediately at 2000 x g for 15 min at 4 °C. The plasma was pipetted into 4 different tubes and stored at -26 °C until analyzed.

Glucose concentrations were analyzed with an enzymatic assay that measured glucose on the basis of the glucose-oxidase reaction (Yellow Spring Instruments glucose analyzer 23A; Bioblock, Strasbourg, France; intraassay and interassay CVs of 1%). Insulin concentrations were measured with a single radioimmunologic assay (Insulin-CT kit; Cis Bio International, Gif-sur-Yvette, France; intraassay and interassay CVs of 7%) and fatty acid concentrations were measured with a colorimetric enzymatic method (C Wako kit; Unipath, Dardilly, France; intraassay and interassay CVs of 5%).

Gas exchange measurement
Tissot gasometers (Gautier, Paris) were used for the gas-exchange measurements. While lying down, the subject breathed into an oral-nasal mask equipped with a 2-way respiratory valve. The subject inspired atmospheric air through the inspiratory valve and expired through the expiratory valve. A 3-way tap allowed for continuous collection of expired air into 1 of 2 cylindrical bells (capacity, 200 L). Every 15 min, the 3-way tap was switched over to the other bell and the 15-min volume of expired air was measured. Then a 30-s continuous stream of expired air was drawn out from the bell by a pump. This air was dried and passed through gas analyzers that measured oxygen fractions with a paramagnetic analysis method and measured carbon dioxide fractions with an infrared method (Analyser Series 1400; Servomex, Paris). The bell was then emptied by pressing down on it gently. The gas analyzers were calibrated before every run by using atmospheric gas, standard gas mixtures, and N2 gas to set the zero points. To validate the system, an alcohol test was performed at the beginning of the experiment.

Urine was collected once and aliquots were frozen for measurement of urea excretion. We used a kinetic colorimetric enzymatic method (Hycel urea kit; Hycel, Pouilly en Auxois, France; intraassay and interassay CVs of 5%). Urinary nitrogen values (in g/min) were used to determine the nonprotein respiratory quotient (npRQ) (9). Carbohydrate and fat oxidation were calculated with the formula published by Jéquier et al (10). This formula uses measured values for O2 (mL/min), CO2 (mL/min), and excreted urinary nitrogen. Energy expenditure (EE) was calculated with the formula published by Ben-Porat et al (11).

Procedure
A preliminary session was performed before the experimental sessions to familiarize the subjects with all of the experimental procedures. Subjects were required to carefully weigh and report in a food diary all the foods and drinks that they consumed at dinner the night before the first test day and at breakfast on the first test day. They were instructed to eat exactly the same dinner and breakfast at the same hour of the day before the subsequent sessions. They were also told that they should not consume any food or drink, except water, between these meals. In addition, they were instructed to refrain from any unusual exercise and to try to sleep for the same duration on the nights preceding all the sessions. There was no evidence that the subjects failed to carry out these instructions.

On each test day, subjects arrived at the laboratory at 1145. The investigator confirmed that subjects were feeling well and had been compliant. Each subject was seated in 1 of 2 windowless, sound-attenuated rooms. Subjects were told that they would remain there until 2200 regardless of the time of their request for dinner. All time cues were then removed. No watches, clocks, radios, or television were in the room and all personnel were carefully trained to avoid references to time or time intervals. While isolated from temporal cues, subjects rapidly lost track of time (5, 12).

At 1215, after the premeal hunger rating, subjects were served a lunch that they had to consume in its entirety. For each subject, the energy content of this test lunch corresponded to that of their usual lunch as calculated from their 7-d food diary. On the first test day, the meal duration was measured; subjects were instructed to finish eating within 20 min. After eating, subjects were asked to rate their hunger. Additional hunger ratings were then recorded at 30-min intervals until the dinner request. To avoid references to time, we obtained additional mock ratings at irregular intervals; these ratings were not used in the analyses.

From lunch until 1415, no restrictions were placed on the subjects’ activities as long as the activities did not provide time cues. They were allowed to read and listen to music. At 1415, subjects were required to lie down for blood sampling; the cannula was inserted into an antecubital vein before the placement of the double-lumen catheter. Thereafter, subjects could only listen to music and were asked to remain quiet but stay awake throughout the session. Continuous blood sampling began at 1515.

At 1545, subjects were asked to empty their bladders to provide a urine specimen. For the HP and HC snack sessions, subjects were then given the snack at 1550 (215 min after the beginning of the lunch). They had to consume the entire snack within 10 min. Pleasantness ratings were recorded after consumption of the snack. At 1600, the mask system was placed on the subject’s face and continuous gas measurements began at 1605. These measurements continued until the spontaneous dinner request. Blood withdrawal continued until 10 min after the dinner request to take into account the time required for the blood to travel from the arm to the tubes. Then the blood sampling stopped and the subject was presented with the buffet-style dinner from which he could eat whatever he wanted. Upon completion of each test day, each subject was asked to estimate the actual time of day to confirm that he had been time-blinded successfully.

Calculations and statistical analyses
Mean temporal patterns of hunger ratings were calculated as in our previous studies (3, 4). Briefly, the temporal patterns of hunger ratings between lunch and dinner can be described in terms of a postlunch profile and a predinner profile. For each experimental condition, the postlunch profile was determined by calculating the mean of the subjects’ ratings for each time (every 30 min before and after lunch) up to the time of the earliest dinner request (390 min, see below) of the 3 sessions. For each session, the predinner profile was determined by establishing another time 0, which was the mean time of the dinner request. The mean time of the last hunger rating before the dinner request was then calculated along with the subjects’ hunger ratings at that time. The means of the ratings of the 8 subjects at 30, 60, and 120 min before the dinner request were then calculated.

All blood variables were expressed as changes from the basal value at 210 min. We obtained postsnack and predinner profiles of changes in plasma glucose, insulin, and fatty acid concentrations and substrate oxidation, npRQ, and EE by using the same methods that were used for the hunger rating profiles. Areas under the curve (AUCs) for glucose, insulin, and fatty acids were calculated over the postsnack and predinner periods by using the trapezoidal method. The AUCs were determined as areas over the basal value.

Hunger ratings at each eating occasion (lunch, snack, and dinner), latencies of the dinner request, food intakes, and AUCs were analyzed by repeated-measures analysis of variance (ANOVA) with session type as the within-subject factor. We used SSPS version 6.1 (SPSS Inc, Chicago). When these analyses revealed a significant session effect, a contrast analysis was performed with Scheffe’s procedure to determine which conditions differed. Hunger ratings, metabolic profiles, and hormonal profiles in the 3 sessions were compared by using repeated-measures ANOVA with time and session type as within-subject factors. We performed contrasts, as described above, when these analyses showed a significant session effect. Differences were considered significant if the observed F ratio exceeded the critical F value with P < 0.05.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Blinding to time of day
The average difference between the subject’s estimation of the time of day and the actual time at the end of the sessions was only -8 ± 20 min but ranged from -140 to 110 min, suggesting that the subjects were blinded to the time of day.

Hunger and pleasantness ratings
Hunger ratings just before the lunch, snack, or dinner were not significantly different across sessions. Mean hunger ratings obtained in the 3 different sessions are shown in Figure 1Go. The analyses showed a main effect of time for postlunch and predinner ratings (P < 0.001 for both) but no effect of session type for either. Therefore, hunger sensations were not significantly different between sessions from lunch until the dinner request. Mean pleasantness ratings were 58 ± 8 and 60 ± 9 mm on a 100-mm visual analogue scale for the HP and HC snacks, respectively; these values were not significantly different.



View larger version (15K):
[in this window]
[in a new window]
  		FIGURE 1. Mean (± SEM) postlunch and predinner hunger ratings in the basal session ({circ}), high-protein (HP) snack session ({triangleup}), and high-carbohydrate (HC) snack session ({square}); n = 8. There were no significant differences between sessions.

 
Latency of dinner request
Latencies of the dinner request were 433 ± 19, 471 ± 10, and 429 ± 13 min in the basal, HP-snack, and HC-snack sessions, respectively; these values were significantly different (P < 0.01). Contrast analysis showed that latency in the HP session was significantly longer than in the basal and HC sessions (P < 0.05 and P < 0.01, respectively). Latencies in the basal and HC sessions were not significantly different from each other.

Energy and macronutrient intakes
The group mean energy intake at lunch was 3763 ± 221 kJ (range: 3135–4843 kJ). Energy intake at the ad libitum dinner did not differ significantly between the 3 sessions (Figure 2Go). However, total energy intake (lunch, snack, and dinner combined) differed significantly (P < 0.01) and contrasts showed that it was higher in the HC session than in the basal and HP sessions (P < 0.02 for both). Macronutrient intakes were not significantly different across sessions.



View larger version (46K):
[in this window]
[in a new window]
  		FIGURE 2. Mean (± SEM) total energy intake from dinner ({blacksquare}), a snack ({square}), and lunch ({blacksquare}) in the basal session, high-protein (HP) snack session, and high-carbohydrate (HC) snack session; n = 8. aSignificantly different from the basal session, P < 0.01. bSignificantly different from the HP session, P < 0.05.

 
Temporal metabolic and hormonal changes from snack time to dinner time
As expected, the ANOVAs for each set of postsnack and predinner profiles (ie, metabolic, hormonal, and macronutrient oxidation profiles) showed a main effect of time (P < 0.05 for all postsnack profiles and P < 0.01 for all predinner profiles), except for the postsnack profiles for npRQ and fat oxidation (NS).

Blood variables
Basal values for plasma glucose concentrations did not differ significantly across sessions. Analyses of the postsnack and predinner changes in plasma glucose profiles showed a main effect of session type (P < 0.05) and a time-by-session interaction in the postsnack period (P < 0.005). Comparisons revealed that the postsnack profile obtained in the HC session was significantly different from that obtained in the basal session (P < 0.05). The predinner plasma glucose profile in the HC session was significantly different from those in the basal and HP sessions (P < 0.05 for both; Figure 3Go).



View larger version (17K):
[in this window]
[in a new window]
  		FIGURE 3. Mean (± SEM) postsnack and predinner temporal profiles of changes in plasma glucose concentrations in the basal session ({circ}), high-protein (HP) snack session ({triangleup}), and high-carbohydrate (HC) snack session ({square}); n = 8. For clarity, SEMs are shown every 20 min only. Because blood was drawn continuously and collected every 5 min, the width of each symbol equals 5 min. The open double arrow above the predinner curves indicates significant differences between sessions: a,bthe HC session was significantly different from the basal and HP sessions, P < 0.05 for both. Filled symbols indicate significant differences between time points: {blacksquare}, differences between the HC and basal sessions, P < 0.05.

 
Basal values for plasma insulin concentrations did not differ significantly across sessions. Analyses of the postsnack and the predinner changes in plasma insulin profiles showed a main effect of session type (P < 0.005 and P < 0.05, respectively). There was also a time-by-session interaction in the postsnack period (P < 0.01). Comparisons revealed that more insulin was secreted in the HC session than in the basal and HP sessions during the postsnack period (P < 0.01 and P < 0.05, respectively) and during the predinner period (P < 0.05 for both) (Figure 4Go).



View larger version (16K):
[in this window]
[in a new window]
  		FIGURE 4. Mean (± SEM) postsnack and predinner temporal profiles of changes in plasma insulin concentrations in the basal session ({circ}), high-protein (HP) snack session ({triangleup}), and high-carbohydrate (HC) snack session ({square}); n = 8. For clarity, SEMs are shown every 20 min only. Because blood was drawn continuously and collected every 5 min, the width of each symbol equals 5 min. The open double arrow above the predinner curves indicates significant differences between sessions: a,bthe HC session was significantly different from the basal and HP sessions, P < 0.05 for both. Filled symbols indicate significant differences between time points: {blacksquare}, differences between the HC and basal sessions; {blacktriangleup}, differences between the HC and HP sessions, P < 0.05 for all.

 
Basal plasma fatty acid concentrations did not differ significantly across sessions. Analyses of the postsnack and predinner changes in plasma fatty acid profiles showed main effects of session type for both periods (P < 0.05 and P < 0.01, respectively). A time-by-session interaction occurred in the postsnack period (P < 0.05). Also during the postsnack period, changes in plasma fatty acid concentrations in the HP session were significantly smaller than those in the basal session (P < 0.05). In contrast, during the predinner period, changes in plasma fatty acid concentrations in the HC session were smaller than those in the basal and HP sessions (P < 0.05 for both) (Figure 5Go).



View larger version (18K):
[in this window]
[in a new window]
  		FIGURE 5. Mean (± SEM) postsnack and predinner temporal profiles of changes in plasma fatty acid concentrations in the basal session ({circ}), high-protein (HP) snack session ({triangleup}), and high-carbohydrate (HC) snack session ({square}); n = 8. For clarity, SEMs are shown every 20 min only. Because blood was drawn continuously and collected every 5 min, the width of each symbol equals 5 min. The open double arrows above the predinner curves indicate significant differences between sessions: a,bthe HC session was significantly different from the basal and HP sessions; cthe HP session was significantly different from the basal session, P < 0.05 for all.

 
The AUCs for glucose, insulin, and fatty acids are shown in Figure 6Go. In most cases, the differences between the AUCs for different sessions were in agreement with the differences observed for the blood profiles.



View larger version (12K):
[in this window]
[in a new window]
  		FIGURE 6. Mean (± SEM) postsnack and predinner areas under the curve (AUCs) for glucose, insulin, and fatty acids in the basal ({square}) session, high-protein (HP; {blacksquare}) snack session, and high-carbohydrate (HC; {blacksquare}) snack session; n = 8. a,cSignificantly different from the basal session, aP < 0.05, cP < 0.01.

bSignificantly different from the HP session, P < 0.05.

 
Macronutrient oxidation, npRQ, and EE
Analyses of the postsnack and predinner carbohydrate oxidation profiles showed main effects of session type for both time periods (P < 0.01 for both; Figure 7Go). Compared with the basal session, the HC and HP snacks induced significant increases in carbohydrate oxidation in the postsnack period (P < 0.05 for both) and in the predinner period (P < 0.001 for both). The HC snack induced greater carbohydrate oxidation than did the HP snack in the predinner period (P < 0.05).



View larger version (15K):
[in this window]
[in a new window]
  		FIGURE 7. Mean (± SEM) postsnack and predinner temporal profiles of carbohydrate oxidation in the basal session ({circ}), high-protein (HP) snack session ({triangleup}), and high-carbohydrate (HC) snack session ({square}); n = 8. The open double arrows above the curves indicate significant differences between sessions: a,aathe HC session was significantly different from the basal session; bthe HC session was significantly different from the HP session; c,ccthe HP session was significantly different from the basal session, a,b,cP < 0.05, aa,ccP < 0.001.

 
Initial analyses of the postsnack and predinner fat oxidation profiles showed no significant effect of session type, but in the predinner period this difference nearly reached significance (P < 0.07). Therefore, further comparisons were conducted and they showed significantly lower fat oxidation in the HC session than in the basal and HP sessions (P < 0.05 for both; Figure 8Go).



View larger version (15K):
[in this window]
[in a new window]
  		FIGURE 8. Mean (± SEM) postsnack and predinner temporal profiles of fat oxidation in the basal session ({circ}), high-protein (HP) snack session ({triangleup}), and high-carbohydrate (HC) snack session ({square}); n = 8. The open double arrow above the predinner curves indicates significant differences between sessions: a,bthe HC session was significantly different from the basal and HP sessions, P < 0.05 for both.

 
Analyses of the npRQ profiles showed no effect of session type in the postsnack period but a significant effect of session type in the predinner period (P < 0.01). Compared with the basal and HP sessions, the HC session was associated with a significantly lower npRQ (P < 0.01 and P < 0.001, respectively; Figure 9Go).



View larger version (14K):
[in this window]
[in a new window]
  		FIGURE 9. Mean (± SEM) postsnack and predinner temporal profiles of nonprotein respiratory quotient (npRQ) in the basal session ({circ}), high-protein (HP) snack session ({triangleup}), and high-carbohydrate (HC) snack session ({square}); n = 8. Urinary nitrogen values (in g/min) were used to determine the npRQ (9). The open double arrow above the predinner curves indicates significant differences between sessions: a,bthe HC session was significantly different from the basal and HP sessions, P < 0.01 for both.

 
Analyses of the postsnack EE profile showed no effect of session type. However, there was a significant session-type effect in the predinner period (P < 0.05), when EE was significantly higher in the HP session than in the basal session (P < 0.01; Figure 10Go). Results of the ANOVAs for total substrate oxidized and EE are shown in Table 4Go; these findings are in complete agreement with the results for the macronutrient oxidation profiles.



View larger version (14K):
[in this window]
[in a new window]
  		FIGURE 10. Mean (± SEM) postsnack and predinner temporal profiles of energy expenditure in the basal session ({circ}), high-protein (HP) snack session ({triangleup}), and high-carbohydrate (HC) snack session ({square}); n = 8. The open double arrow above the predinner curves indicates a significant difference between sessions: athe HP session was significantly different from the basal session, P < 0.01.

 

View this table:
[in this window]
[in a new window]
  		TABLE 4 Total energy expenditure and total substrate oxidation during the postsnack and predinner periods of the basal, high-protein (HP) snack, and high-carbohydrate (HC) snack sessions1
 
Diet-induced thermogenesis (DIT) during the time period between snack consumption and the dinner request was calculated for each subject as the difference between EE in each snack session and EE in the basal session. We then analyzed the relation between DIT and the latency of the dinner request. The results showed that DIT and latency were significantly correlated (r = 0.499, P < 0.049). Differences in DIT between the HP snack and the HC snack and the differences in latency between these 2 snacks were highly correlated (r = 0.918, P < 0.001).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The results of this study show that HP or HC snacks consumed in a satiety state had poor satiating power. Compared with the basal session when no snack was provided, consumption of the 1-MJ HP snack did slightly prolong, by {approx}40 min, the satiety induced by lunch, whereas the HC snack did not do so. Note that this study was performed with only 8 subjects and thus it may lack statistical power. This could explain the failure to find significant changes as a result of the HC snack. However, our behavioral results were in complete agreement with the metabolic changes.

In the present study, subjects were time-blinded and the satiety effect was evaluated on the basis of hunger ratings and the duration of the interval between lunch and the spontaneous dinner request. Our behavioral observations are supported by several studies that determined the satiety effect of preloads consumed between 180 and 30 min before a meal (13–17). In these studies, the satiety effect was evaluated on the basis of hunger and fullness ratings or measurement of food intake at the next meal. Our results also confirm those of Poppitt et al (17), who evaluated the effects of a HP or HC preload consumed by lean women 90 min before lunch. These preloads contained similar amounts of fat and approximately the same amount of energy ({approx}2 MJ). Compared with the HC preload, after consuming the HP preload subjects reported significantly greater feelings of gastric fullness and less motivation to eat; they also consumed less food at the ad libitum lunch.

The results of our previous studies (3, 5) suggested that the duration of satiety depends on the disposal of metabolic fuels and particularly on glucose availability. Initiation of a meal in rats and hunger sensations in humans were shown to be preceded by a transient drop in blood glucose (of at least 5 min duration in humans), which is thought to reflect a sudden decrease in the supply of immediately available glucose (12, 18–21). After an eating episode, it would be expected that the intermeal interval (ie, the interval of time before glucose availability decreases) would depend on the carbohydrate content of the meal, the additional glucose provided by glycogenolysis and gluconeogenesis, glucose utilization (oxidation and storage), and glucose spared by fat oxidation. Note that our blood-collection protocol did not allow us to observe the preprandial decline in blood glucose. However, we observed in all sessions that glucose and insulin concentrations were low and were not significantly different between sessions at the time of the dinner request.

Can the substrate and hormonal profiles together with the substrate oxidation profiles suggest explanations for the behavioral findings? First, it is important to note that these snacks were ingested by subjects in a nonhungry state. Consequently, their stomachs were not empty and ingestion of the mandatory snacks simply altered the composition of the gastric contents. If we assume that the mean rate of gastric emptying is {approx}8.4 kJ/min (22, 23), we could estimate that the snacks were ingested near the half-emptying time. At that time, because fat is absorbed slowly, the gastric contents were fat-enriched with proportionally less protein and carbohydrate compared with the macronutrient composition of the lunch (16%, 26%, and 58% of energy from protein, fat, and carbohydrate, respectively). After the HP and HC snacks, the stomach contents had higher proportions of protein and glucose, respectively. Moreover, both snacks reduced the proportion of fat in the stomach contents.

Compared with the basal session, the HP snack did not alter the glucose and insulin profiles significantly but it was associated with lower fatty acid concentrations during the postsnack period. It is known that protein ingestion, and therefore amino acid absorption, increases glucose availability. On the one hand, hepatic glycogenolysis is induced by the release of glucagon and on the other hand, gluconeogenesis is the major pathway of amino acid degradation (24). Livesey and Elia (25) calculated that 1 g degraded protein forms 0.3 mmol (0.54 g) glucose. However, protein degradation and therefore gluconeogenesis are slow processes; Nuttall and Gannon (26) showed that plasma urea was still elevated 8 h after consumption of proteins. Under the conditions of the present study, degradation of protein consumed at the snack could still have been in progress at the time of the dinner request.

Proteins have a weak stimulatory effect on insulin release when ingested alone (27–29), but they potentiate glucose-induced insulin secretion when ingested simultaneously with carbohydrate (29, 30). However, the natural starch-protein association reduces the rate of carbohydrate absorption (31). In the present study, the HP snack contained a very small amount of carbohydrate, mainly provided by a natural association in the chicken breast muscle. Only circulating amino acids have a stimulatory effect on insulin release (32), and it was shown that absorption of amino acids depends on the digestibility of the ingested proteins (33). There are several possible reasons why the HP snack did not alter the insulin profile: 1) the low absorption rate for chicken protein (probably because of its texture) (34), 2) the slow rate of protein degradation, and 3) the reduced proportion of carbohydrate in the stomach because the snack was high in protein. These reasons could also account for the small but long-lasting contribution of the HP snack to the available glucose pool. Furthermore, the absence of a difference in protein oxidation between sessions could also be explained by the low rate of protein degradation.

It should be remembered that the thermogenesis induced by protein catabolism can be as great as 25% of the provided energy (35). In view of these features of protein absorption and degradation, it is not surprising that the higher EE (compared with the basal session) after the HP snack was accounted for by greater carbohydrate oxidation. Westerterp-Plantenga et al (36) observed that an increased DIT was associated with an increased feeling of satiety. Our results seem to confirm these findings, however in the present study, the induced DIT was evaluated over the interval of time between snack consumption and the dinner request; thus, it was not measured in its entirety.

In the HP session, the intermeal interval was prolonged by {approx}40 min compared with the basal session. In these young subjects, the basal EE was estimated to be {approx}5 kJ/min (200 kJ/40 min) and was slightly augmented by the thermogenesis from protein as noted above. The difference between energy intake (1 MJ) and EE could be explained by increased storage of glycogen and fat. Moreover, at the time of the dinner request, the protein load was still not entirely degraded. It is possible that substrate utilization in the postdinner period was modified by this ongoing protein catabolism.

Compared with the basal session, the HC snack was associated with higher plasma concentrations of glucose and insulin throughout the intermeal interval and lower fatty acid concentrations during the predinner period. These results and the findings on substrate oxidation are completely explained by the changes in the composition of the stomach contents induced by the snack. In the fed state, when a carbohydrate load is ingested there is an increase in the oxidative disposal of glucose (37–39), which spares fatty acid utilization (40, 41). After the HC snack, glucose was oxidized at nearly the maximal rate (ie, the npRQ was close to 1), but glucose was also stored, which would account for the absence of a satiety effect. We calculated that {approx}50% of the glucose content of the snack was oxidized and {approx}50% was stored. Therefore, glucose oxidation amounted to {approx}330 kJ, and because there was no change in EE relative to the basal session, an excess of {approx}660 kJ of fat was stored. Also, as noted for the HP snack, an effect on postdinner substrate utilization cannot be ruled out.

We showed that in subjects who usually ate 3 meals/d with no intermeal food intake, a snack consumed in a nonhungry state only slightly influenced the satiety effect of the preceding meal. Because we only studied the short-term effects of snack intake, these findings may not be generalizable to longer-term energy balance. For instance, Johnstone et al (42) recently showed that in subjects consuming a medium-fat diet ad libitum, 3 mandatory snacks/d over a 7-d period did not increase energy intake compared with a 7-d period in which the subjects had no snacks. It would be interesting to know whether their subjects tended to snack and also to have measures of their meal sizes and times over the 7-d period. Also, we studied normal-weight subjects and we cannot be sure that the same mechanisms occur in obese subjects. Blair (43) showed that obese subjects on restricted diets lost more weight when they stopped eating between meals.

In a recent paper, we proposed an endocrine and biological model of the intermeal interval (44). Our research showed that the hunger-triggered onset of a meal was preceded by a comprehensive endocrine and substrate profile characterized in particular by decreasing plasma glucose, insulin, and leptin concentrations. Currently, clear definitions of what constitutes a meal and what constitutes a snack are urgently needed. On the basis of metabolic and behavioral findings, we propose to define a meal as an eating episode (whatever its size or timing) motivated by hunger and therefore preceded by a particular metabolic pattern. In contrast, a snack would be defined as an eating episode not triggered by hunger but instead triggered by anything else. Because of differences between a meal and a snack in the metabolic state immediately before ingestion, the utilization of energy substrates provided by a meal compared with a snack would probably be quite different. This could account for the contradictory results from both experimental studies (45) and epidemiologic studies (46, 47) regarding the influence of feeding frequency on metabolism and anthropometry.


   ACKNOWLEDGMENTS
 
We thank Annie Jaeger, Jocelyne Brouillard, and Monsieur Poirot for their efficient technical assistance and Nicole Colas-Linhart and Anne Petiet, without whom the insulin measurements would not have been possible.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Basdevant A, Craplet C, Guy-Grand B. Snacking patterns in obese French women. Appetite 1993;21:17–23.[Medline]
  2. Takahashi E, Yoshida K, Sugimori H, et al. Influence factors on the development of obesity in 3-year-old children based on the Toyama study. Prev Med 1999;28:109–14.
  3. Marmonier C, Chapelot D, Louis-Sylvestre J. Metabolic and behavioral consequences of a snack consumed in a satiety state. Am J Clin Nutr 1999;70:854–66.[Abstract/Free Full Text]
  4. Marmonier C, Chapelot D, Louis-Sylvestre J. Effect of macronutrient content and energy density of snacks consumed in a satiety state on the onset of the next meal. Appetite 2000;34:161–8.[Medline]
  5. Himaya A, Fantino M, Antoine JM, Brondel L, Louis-Sylvestre J. Satiety power of dietary fat: a new appraisal. Am J Clin Nutr 1997;65:1410–8.[Abstract/Free Full Text]
  6. Garner D, Garfinkel P. The eating attitude test: an index of the symptoms of anorexia nervosa. Psychol Med 1979;9:273–9.[Medline]
  7. Stunkard AJ, Messick S. The three-factor eating questionnaire to measure dietary disinhibition and hunger. J Psychosom Res 1985;29:71–83.[Medline]
  8. Chabert MF, Verger P, Louis-Sylvestre J. A method for long-term and accurate measurement and recording of blood glucose level in man. Physiol Behav 1991;49:827–30.[Medline]
  9. Jéquier E. Measurement of energy expenditure in clinical nutritional assessment. JPEN J Parenter Enteral Nutr 1987;11:86–9.[Abstract]
  10. Jéquier E, Acheson K, Schutz Y. Assessment of energy expenditure and fuel utilization in man. Annu Rev Nutr 1987;7:187–208.[Medline]
  11. Ben-Porat M, Sideman S, Bursztein S. Energy metabolism rate equation for fasting and postabsorptive subjects. Am J Physiol 1983;244:R764–9.
  12. Melanson KJ, Westerterp-Platenga MS, Saris WHM, Smith FJ, Campfield LA. Blood glucose patterns and appetite in time-blinded humans: carbohydrate versus fat. Am J Physiol 1999;277:R337–45.[Abstract/Free Full Text]
  13. Booth DA. The physiology of appetite. Br Med Bull 1981;37:135–40.[Free Full Text]
  14. Booth DA. Mechanisms from models—actual effects from real life: the zero-calorie drink-break option. Appetite 1988;1(suppl):94–102.
  15. Rolls BJ, Kim S, McNelis AL, Fischman MW, Foltin RW, Moran TH. Time course of effects of preloads high in fat or carbohydrate on food intake and hunger in humans. Am J Physiol 1991;260:R756–63.[Abstract/Free Full Text]
  16. Porrini M, Santangelo A, Crovetti R, Riso P, Testolin G, Blundell JE. Weight, protein, fat, and timing of preloads affect food intake. Physiol Behav 1997;62:563–70.[Medline]
  17. Poppitt SD, McCormack D, Buffenstein R. Short-term effects of macronutrient preloads on appetite and energy intake in lean women. Physiol Behav 1998;64:279–85.[Medline]
  18. Louis-Sylvestre J, Le Magnen J. A fall in blood glucose levels precedes meal onset in free-feeding rats. Neurosci Biobehav Rev 1980;4(suppl):13–5.
  19. Smith FJ, Campfield LA. Meal initiation occurs after experimental induction of transient declines in blood glucose. Am J Physiol 1993;265:R1423–9.[Abstract/Free Full Text]
  20. Campfield LA, Smith FJ, Rosenbaum M, Hirsch J. Human eating: evidence for a physiological basis using a modified paradigm. Neurosci Biobehav Rev 1996;20:133–7.[Medline]
  21. Campfield LA, Smith FJ. Functional coupling between transient declines in blood glucose and feeding behavior: temporal relationships. Brain Res Bull 1986;17:427–33.[Medline]
  22. Carbonnel F, Lémann M, Rambaud JC, Mundler O, Jian R. Effect of the energy density of a solid-liquid meal on gastric emptying and satiety. Am J Clin Nutr 1994;60:307–11.[Abstract/Free Full Text]
  23. Collins PJ, Horowitz M, Cook DJ, Harding PE, Shearman DJC. Gastric emptying in normal subjects, a reproducible technique using a single scintillation camera and computer system. Gut 1983;24:1117–25.[Abstract/Free Full Text]
  24. Tappy L, Acheson K, Normand S, et al. Effects of infused amino acids on glucose production and utilization in healthy human subjects. Am J Physiol 1992;262:E826–33.[Abstract/Free Full Text]
  25. Livesey G, Elia M. Estimation of energy expenditure, net carbohydrate utilization, net fat oxidation and synthesis by indirect calorimetry: evaluation of errors with special reference to the detailed composition of fuels. Am J Clin Nutr 1988;47:608–28.[Abstract/Free Full Text]
  26. Nuttall FQ, Gannon MC. Metabolic response to egg white and cottage cheese protein in normal subjects. Metabolism 1990;39:749–55.[Medline]
  27. Krezowski PA, Nuttall FQ, Gannon MC, Bartosh NH. The effect of protein ingestion on the metabolic response to oral glucose in normal individuals. Am J Clin Nutr 1986;44:847–56.[Abstract/Free Full Text]
  28. Rabinowitz D, Maffezzoli R, Merimee TJ, Burgess JA. Patterns of hormonal release after glucose, protein, and glucose plus protein. Lancet 1966;2:454–6.[Medline]
  29. Westphal SA, Gannon MC, Nutall FQ. Metabolic response to glucose ingested with various amounts of protein. Am J Clin Nutr 1990;52:267–72.[Abstract/Free Full Text]
  30. Spiller GA, Jensen CD, Pattison TS, Chuck CS, Wittham JH, Scala J. Effect of protein dose on serum glucose and insulin response to sugars. Am J Clin Nutr 1987;46:474–80.[Abstract/Free Full Text]
  31. Jenkins DJA, Thorne MJ, Wolever TMS, Jenkins AL, Rao AV, Thompson LU. The effect of starch-protein interaction in wheat on the glycemic response and rate of in vitro digestion. Am J Clin Nutr 1987;45:946–51.[Abstract/Free Full Text]
  32. Schmid R, Schulte-Frohlinde E, Schusdziarra V, et al. Contribution of postprandial amino acid levels to stimulation of insulin, glucagon and pancreatic polypeptide in humans. Pancreas 1992;7:698–704.[Medline]
  33. Boirie Y, Dangin M, Gachon P, Vasson MP, Maubois JL, Beaufrère B. Slow and fast dietary proteins differently modulate postprandial protein accretion. Proc Natl Acad Sci U S A 1997;94:14930–5.[Abstract/Free Full Text]
  34. Uhe AM, Collier GR, O’Dea K. A comparison of the effects of beef, chicken and fish protein on satiety and amino acid profiles in lean male subjects. J Nutr 1992;122:467–72.
  35. Flatt JP. Energy costs of ATP synthesis. In: Kinney JM, Tucker HN, eds. Energy metabolism: tissue determinants and cellular corollaries. New York: Raven Press, 1992:319–43.
  36. Westerterp-Plantenga MS, Rolland V, Wilson SA, Westerterp KR. Satiety related to 24h diet-induced thermogenesis during high protein/carbohydrate vs high fat diets measured in a respiration chamber. Eur J Clin Nutr 1999;62:205–12.
  37. Abbott WGH, Howard BV, Christin L, et al. Short term energy balance: relationship with protein, carbohydrate and fat balances. Am J Physiol 1988;225:E332–7.
  38. Jéquier E. Carbohydrates as a source of energy. Am J Clin Nutr 1994;59(suppl):682S–5S.[Abstract/Free Full Text]
  39. Ritz P, Krempf M, Cloarec D, Champ M, Charbonnel B. Comparative continuous-indirect-calorimetry study of two carbohydrates with different glycemic indices. Am J Clin Nutr 1991;54:855–9.[Abstract/Free Full Text]
  40. Randle PJ. Metabolic fuel selection: general integration at the whole-body level. Proc Nutr Soc 1995;54:317–27.[Medline]
  41. Randle PJ, Garland PB, Hales CN, Newsholme EA. The glucose-fatty acid cycle: its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus. Lancet 1963;1:785–98.[Medline]
  42. Johnstone AM, Shannon E, Whybrow S, Reid CA, Stubbs RJ. Altering the temporal distribution of energy intake with isoenergetically dense foods given as snacks does not affect total daily energy intake in normal-weight men. Br J Nutr 2000;83:7–14.[Medline]
  43. Blair AJ. When are calories more fattening? Appetite 1991;17:161 (letter).[Medline]
  44. Chapelot D, Aubert R, Marmonier C, Chabert M, Louis-Sylvestre J. An endocrine and metabolic definition of the intermeal interval in humans: evidence for a role of leptin on the prandial pattern through fatty acid disposal. Am J Clin Nutr 2000;72:421–31.[Abstract/Free Full Text]
  45. Fabry P, Hejda S, Cerny K, Osancova K, Pechar J. Effect of meal frequency in schoolchildren. Changes in weight-height proportion and skinfold thickness. Am J Clin Nutr 1966;18:358–61.[Abstract]
  46. Fabry P, Fodor J, Hejl Z, Braun T, Zvolankova K. The frequency of meals: its relation to overweight, hypercholesterolemia and decreased glucose tolerance. Lancet 1964;2:614–5.[Medline]
  47. Edelstein SL, Barrett-Connor EL, Wingard DL, Cohn BA. Increased meal frequency associated with decreased cholesterol concentrations; Rancho Bernardo, CA, 1984–1987. Am J Clin Nutr 1992;55:664–9.[Abstract/Free Full Text]
Received for publication July 9, 2001. Accepted for publication September 26, 2001.




This article has been cited by other articles:


Home page
 Am. J. Physiol. Endocrinol. Metab.Home page
D. E. Cummings, R. S. Frayo, C. Marmonier, R. Aubert, and D. Chapelot
Plasma ghrelin levels and hunger scores in humans initiating meals voluntarily without time- and food-related cues
Am J Physiol Endocrinol Metab, August 1, 2004; 287(2): E297 - E304.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Marmonier, C.
Right arrow Articles by Louis-Sylvestre, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Marmonier, C.
Right arrow Articles by Louis-Sylvestre, J.
Agricola
Right arrow Articles by Marmonier, C.
Right arrow Articles by Louis-Sylvestre, J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS