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 Flint, A.
Right arrow Articles by Astrup, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Flint, A.
Right arrow Articles by Astrup, A.
Agricola
Right arrow Articles by Flint, A.
Right arrow Articles by Astrup, A.
American Journal of Clinical Nutrition, Vol. 84, No. 6, 1365-1373, December 2006
© 2006 American Society for Nutrition

ORIGINAL RESEARCH COMMUNICATION

Glycemic and insulinemic responses as determinants of appetite in humans1,2,3

Anne Flint, Bente K Møller, Anne Raben, Birgitte Sloth, Dorthe Pedersen, Inge Tetens, Jens J Holst and Arne Astrup

1 From the Department of Human Nutrition, The Centre for Advanced Food Studies, The Royal Veterinary and Agricultural University, Frederiksberg, Denmark (AF, BKM, AR, BS, DP, IT, and AA), and the Department of Medical Physiology, The Panum Institute, University of Copenhagen, Copenhagen, Denmark (JJ)

2 Supported by grant no. 42870 from the Danish Research Agency and by Kellogg's Europe.

3 Reprints not available. Address correspondence to B Sloth, Department of Human Nutrition, Centre for Advanced Food Studies, The Royal Veterinary and Agricultural University, 30 Rolighedsvej, DK-1958 Frederiksberg C, Denmark. E-mail: bsl{at}kvl.dk.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Background: The importance of the postprandial glycemic and insulinemic responses for appetite and energy intake (EI) is controversial.

Objective: The aim of the study was to test the hypothesis that postprandial appetite sensations and subsequent EI are determined by postprandial glycemic and insulinemic responses after the intake of a range of breakfast meals.

Design: The study was a randomized, crossover meal test including 28 healthy young men, each of whom tested 10 of 14 breakfast meals. Each meal contained 50 g carbohydrate with various glycemic index and energy and macronutrient contents. Blood samples were taken, and appetite sensations were measured 3 h after the meals. Subsequently, EI at lunch (EIlunch) was recorded.

Results: The glycemic response was unrelated to appetite sensations, whereas the insulinemic response was positively associated with postprandial fullness (R2 = 0.33, P < 0.05). In contrast, the insulinemic response was unrelated to the subsequent EIlunch, whereas the glycemic response was positively associated with EIlunch (R2 = 0.33, P < 0.05). Although no significant difference in EIlunch was observed between different breakfast conditions, a low breakfast EI was associated with a high EIlunch (R2 = 0.60, P < 0.001).

Conclusions: The current study does not support the contention that the postprandial glycemic response has an important effect on short-term appetite sensations, but a low–glycemic index meal may reduce subsequent EI. In contrast, postprandial insulin seems to affect short-term appetite sensations.

Key Words: Glucose • insulin • macronutrients • breakfast • satiety • hunger • glycemic index


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The current dietary recommendations of high-carbohydrate diets have been questioned in relation to overweight and obesity, and new alternative diets—eg, the Atkins (1) and South Beach (2) diets—and new dietary advice—eg, The Healthy Eating Pyramid (3)—have been put forward. The focus, therefore, is on carbohydrates, in terms of both quantity and of quality, and they are blamed for, among other things, the continuing rise in obesity in Western societies (3–6). Despite an intense, as yet undecided scientific debate about the quality of carbohydrates in relation to appetite and body weight regulation (4–8), some well-established scientists now advocate a low–glycemic index (low-GI) diet for the treatment and prevention of obesity (3, 4, 6, 8, 9). The claim that low-GI foods have a positive effect on obesity is partly based on the assumption that reduction of postprandial glucose and insulin responses will decrease carbohydrate oxidation and fat storage and increase fat oxidation (5, 6). Furthermore, lower and slower glucose and insulin responses are believed to promote satiety and suppress hunger, because large increases in blood glucose and insulin may subsequently induce hypoglycemia, which leads to increased hunger (5, 6). Pawlak et al (10) recently showed how the sequence of metabolic events after the consumption of a high-glycemic index (high-GI) diet in an animal model gave rise to greater body fat deposition.

The glucostatic theory of >50 y ago proposed a link between blood glucose concentrations and appetite sensations. According to this theory, high blood glucose concentrations signal satiety and termination of feeding, whereas low blood glucose concentrations trigger the onset of feeding (11). However, consistent evidence has still not been found that an increase in blood glucose, acute or sustained, is the primary determinant of satiety and food intake in humans (12–14).

The current study investigates the hypothesis that postprandial appetite and subsequent energy intake (EI) are determined by the source of carbohydrates in a range of typical European breakfast meals, according to the different glycemic and insulinemic responses to these meals. Prediction equations of postprandial responses of blood glucose, insulin, and appetite were developed by using the meal descriptors energy, energy density, macronutrient, and dietary fiber content.


   SUBJECTS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Subjects
Twenty-eight healthy men participated in the study. They were young (x ± SD age: 24.8 ± 0.5 y), normal-weight [body mass index (in kg/mg2): 22.5 ± 0.3), nonsmoking, and not elite athletes, and they had no history of metabolic disease. Flyers posted at several locations near the university, in supermarkets, and in libraries were used to recruit subjects.

Subjects gave written informed consent after the study was explained to them orally and in writing. The study was approved by the Municipal Ethics Committee of Copenhagen and Frederiksberg, with the approval code of (KF) 01–213/00.

Protocol
The study was a randomized, crossover test meal study carried out at the Department of Human Nutrition at the Royal Veterinary and Agricultural University. Each subject tested 9 of 13 test meals and a reference meal (consisting of white-wheat bread) on 10 different occasions, each separated by ≥7 d. The test meal selection was randomized by using a computerized randomization matrix. The order of meals was further randomized for each subject. This randomization matrix resulted in a protocol in which each breakfast meal was tested 18–21 times, and the reference meal was tested 28 times. The subjects were asked to refrain from physical activity and from alcohol consumption for 2 d before the test days. The subjects consumed a standardized dinner meal before 2100 on the evening before a test day. They were allowed to drink water until midnight.

After an overnight fast, the subjects arrived at 0800 at the department by the least strenuous means of transportation. After voiding, subjects were weighed (Lindell Tronic 8000; Lindell, Copenhagen, Denmark; precision to 0.01 kg) while wearing only underwear. On the first test day, the height of the subjects, without shoes, was measured. Body composition was measured by using bioelectrical impedance analysis on an Animeter (HTS Engineering Inc, Odense, Denmark). The subjects rested in a supine position for ≥10 min after which a catheter (Venflon; BOC Ohmeda AB, Helsingborg, Sweden) was inserted in an antecubital arm vein. Venous blood samples were drawn before (0 min) and 15, 30, 45, 60, 90, 120, 150, and 180 min after the test meal. Appetite was measured by using visual analogue scales before (0 min) and 15, 30, 60, 90, 120, 150, and 180 min after the test meal (15). The appetite sensations measured in this study were satiety, fullness, hunger, and prospective food intake.

The participants were instructed to lie down for ≥10 min before each blood sample, but otherwise they were allowed to read, watch videos, and walk around slowly during the test day. The test meal was given at 0830 and was to be eaten within 15 min. At 1130, after the last blood sample was drawn, the venflon catheter was removed and the subjects were given an ad libitum lunch, from which EI was measured. The subjects were allowed to drink as much water (<1 L) as they wished on the first test day, and they were required to drink the same amount of water with the lunch on the subsequent test days.

Test meals
The 13 test meals were designed to resemble typical breakfast meals in the United Kingdom, Germany, Italy, France, Denmark, Sweden, and Finland (Table 1Go). The meals included Finnish bread (God i kvadrat fuldkorn; Fazer, Helsinki, Finland), German bread (Bondebrød; ALDI, Essen, Germany), French bread (Boule Gilbert; Le Blé d'or, Copenhagen, Denmark), and white-wheat bread baked at the Department of Human Nutrition (also used as reference meal). The breads were served with butter (Lurpak; Arla Foods, Viby, Denmark), raspberry jam (Mine egne Hindbær; Irma, Rødover, Denmark), French jam (Confiture Bonne Maman; Andros, Biars sur Cére, France), and cow-milk cheese with 45% fat (Danbo; Arla Foods). The quantities of butter (22% by wt of bread), cheese (65% by wt of bread), and jam (50% by wt of bread) were means of the standard amounts for the different European countries. Cereal breakfast meals included cornflakes (Corn Flakes), sugar-coated cornflakes (Frosties), bran flakes (All-Bran), and fiber-rich cereal (All-Bran Plus; all cereals: Kellogg's, Glostrup, Denmark), which were served with 1.5%-fat milk (Letmælk; Arla Foods). The amount of milk given with the meals corresponded to the recommended amounts stated on the packages by the manufacturers. Other meals consisted of Italian cookies (Taralucci; Barilla, Parma, Italy), served with coffee and milk, and rolled oats (OTA Solgryn; Quaker Oats Scandinavia, Malmö, Sweden) served with milk and sugar or prepared as porridge served with applesauce (Æblemos; FDB, Copenhagen, Denmark).


View this table:
[in this window]
[in a new window]

  		TABLE 1 Composition of breakfast meals containing 50 g available carbohydrates1

 
All meals contained 50 g available carbohydrate, measured by using the manufacturer's procedure 3.2 as described in detail in the Englyst starch kit [procedure 3.2; Product no. 61–000; Englyst Carbohydrate Services Ltd, Chilworth, United Kingdom (16)]. Briefly, the sum of free glucose was measured after 30 min and, with the assumption that this free glucose was derived from sucrose, we estimated the total content of sucrose and added that to the amount of glucose released from starch after 120 min of in vitro digestion. In the case of milk products, the amount of available carbohydrate was taken from the Danish Food Composition Tables (17). The total amount of each breakfast meal was adjusted so that the meal contained 50 g available carbohydrate. The meals therefore differed in energy (range: 1134–2990 kJ), protein (5–28 g; 6–22% of energy), fat (3–42 g; 11–55% of energy), dietary fiber (1–24 g), and carbohydrate (30–75% of energy). Water (250 mL) was served with each test meal. The contents of fat, protein, dietary fiber, and digestible energy were calculated from the Danish Food Composition tables and from the information given by the manufacturer. The composition of the breakfast meals is shown in Table 1Go.

Standardization and ad libitum meals
The meal consumed by all participants on the evening before each test day consisted of a goulash with parboiled rice and a honey cake for dessert. The energy content of this meal was 35% of the daily estimated energy needs of each participant (18). The ad libitum lunch consisted of a pasta salad. The distribution of energy in both the evening meal and the ad libitum lunch was 50% from carbohydrate, 37% from fat, and 13% from protein.

Laboratory analyses
Blood samples were kept on ice and centrifuged at 2800 x g for 10 min at 4 °C. They were then separated into plasma or serum and kept at –20 °C until they were analyzed. The blood samples were analyzed for plasma glucose and serum insulin concentrations.

Plasma glucose concentrations were analyzed by using a COBAS MIRA Plus (Roche Diagnostic Systems, Hoffmann-La Roche LTP, Basel, Switzerland) and a glucose kit (HK 125; ABX Diagnostics, Montpellier, France) (19) with intraassay and interassay CVs <1.6% and <2.8%, respectively. Serum insulin concentrations were measured by using the principle of radioimmunoassay described by Albano et al (20) against a standard of human insulin. The tracer was human insulin that was monoiodinated in position A14 (Novo Nordisk A/S, Bagsvaerd, Denmark). The guinea pig antibody (code 2004) used crossreacted with proinsulin and split insulin. The intraassay CV was <5%, and the sensitivity was <5 pmol/L.

Statistical analysis
The incremental area under the curve (iAUC) was calculated by using the trapezoid rule with the negative values left out. The iAUC was used as a summary measure for the variables glucose, insulin, satiety, and fullness (iAUCglu, iAUCins, iAUCsat, and iAUCfull, respectively). For hunger and prospective food intake, the incremental area over the curve (iAOChun and iAOCpros, respectively) under the fasting level was used as a summary measure. The greater the iAOC values, the less the postprandial sensations of hunger and prospective food consumption. The iAUC and iAOC adjusted (by division) for the amount of energy in breakfasts were also analyzed. Data were log transformed when necessary to meet the assumptions of the different analyses. The effect of meals on glucose, insulin, appetite, and EIlunch and on total EI (breakfast plus lunch; EItotal) were analyzed by using analysis of covariance for the summary measures mentioned above, with subject and meal as class variables and with fasting value, total body weight, and percentage body fat of the subjects as covariates. Tukey-Kramer adjusted t tests were used as post hoc tests.

The associations between appetite, EI, insulin, and glucose were investigated by using correlation analysis of means. Multiple linear regression analyses were used to predict appetite, EI, insulin, and glucose with respect to the meal variables with stepwise exclusion of variables. The meal variables included were content of energy (kJ), energy density (kJ/g), weight of breakfast (g), protein (g and % of energy), fat (g and % of energy), carbohydrate (% of energy), and dietary fiber (g and g/MJ). The described statistical models are based on an assumption of linearity and are statistical attempts to approximate biology. The results should be interpreted with these limitations in mind.

All statistical analyses were conducted with the use of SAS software (version 8; SAS Institute Inc, Cary, NC), and the level of significance was P < 0.05. All results are given as means ± SEM.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Glucose and insulin
The iAUCglu and iAUCins for the 14 test meals are shown in Figure 1Go. The porridge with applesauce and reference bread meals produced significantly higher iAUCglu than did the German, Finnish, and reference breads with butter and cheese, Italian cookies with coffee and milk, reference bread with butter, or All-Bran Plus with milk (Figure 1Go). When values were expressed per energy unit in the meal, Corn Flakes with milk, porridge with applesauce, and reference bread produced significantly higher iAUCglu than did Finnish, German, or reference breads with butter and cheese, reference bread with butter, reference bread with butter and jam, Italian cookies with coffee and milk, or All-Bran Plus with milk.


Figure 1
View larger version (59K):
[in this window]
[in a new window]

  		FIGURE 1. Mean (±SEM) incremental area under the curve for plasma glucose (iAUCglu) and serum insulin response (iAUCins) after the consumption of 14 different breakfasts containing 50 g available carbohydrate. n = 18–28. The composition of the meals was reference bread (meal 1), reference bread with butter (meal 2), reference bread with butter and cheese (meal 3), Finnish bread with butter and cheese (meal 4), German bread with butter and cheese (meal 5), reference bread with butter and jam (meal 6), French bread with butter and jam (meal 7), All-Bran with milk (meal 8), All-Bran Plus with milk (meal 9), Corn Flakes with milk (meal 10), Frosties with milk (meal 11), Italian cookies with coffee and milk (meal 12), rolled oats with sugar and milk (meal 13), and rolled-oats porridge with water and applesauce (meal 14). All-Bran, All-Bran Plus, Corn Flakes, and Frosties are manufactured by Kellogg's (Glostrup, Denmark). Data were analyzed with ANCOVA with fasting value, total body weight, and percentage body fat as covariates (iAUCglu: P < 0.0001, iAUCins: P < 0.0001). Tukey-Kramer adjusted t tests were used as post hoc tests. Values with different letters are significantly different, P < 0.05.

 
The reference bread, All-Bran with milk, All-Bran Plus with milk, and reference and German breads with butter and cheese produced significantly higher iAUCins than did French bread with butter and jam (Figure 1Go). Corn Flakes, Frosties, and All-Bran with milk, reference bread, and porridge with applesauce had significantly higher iAUCins than did all other meals except rolled oats with sugar and milk when values were expressed per energy unit in the meal.

Appetite sensations
The EIs and the iAUC and iAOC for appetite sensations are shown in Figure 2Go. French bread with butter and jam induced a significantly lower iAUCsat than did Finnish and German breads with butter and cheese, Italian cookies with coffee and milk, All-Bran Plus with milk, or porridge with applesauce (Figure 2Go). When values were expressed per energy unit in the meal, porridge with applesauce induced significantly higher iAUCsat than did all other meals except Corn Flakes with milk.


Figure 2
View larger version (43K):
[in this window]
[in a new window]

  		FIGURE 2. Mean (±SEM) incremental area under or over the curve (iAUC and iAOC, respectively) for appetite response after the consumption of 14 different breakfast meals containing 50 g available carbohydrate. n = 18–28. The composition of the meals was reference bread (meal 1), reference bread with butter (meal 2), reference bread with butter and cheese (meal 3), Finnish bread with butter and cheese (meal 4), German bread with butter and cheese (meal 5), reference bread with butter and jam (meal 6), French bread with butter and jam (meal 7), All-Bran with milk (meal 8), All-Bran Plus with milk (meal 9), Corn Flakes with milk (meal 10), Frosties with milk (meal 11), Italian cookies with coffee and milk (meal 12), rolled oats with sugar and milk (meal 13), and rolled-oats porridge with water and applesauce (meal 14). All-Bran, All-Bran Plus, Corn Flakes, and Frosties are manufactured by Kellogg's (Glostrup, Denmark). Data were analyzed with ANCOVA with fasting value, total body weight, and percentage body fat as covariates [iAUC for satiety (iAUCsat), P < 0.0001; iAOC for hunger (iAOChun), P < 0.0001; iAUC for fullness (iAUCfull), P < 0.0001; iAOC for prospective fod consumption (iAOCpros), P < 0.0001]. Tukey-Kramer adjusted t tests were used as post hoc tests. Values with different letters are significantly different, P < 0.05.

 
The same pattern was seen for fullness; ie, French bread with butter and jam resulted in significantly lower iAUCfull than did Finnish, German, and reference breads with butter and cheese, All-Bran Plus with milk, or porridge with applesauce (Figure 2Go). When values were expressed per energy unit in the meal, porridge with applesauce induced significantly higher iAUCfull than did all other meals.

French bread with butter and jam, reference bread, and Frosties with milk resulted in significantly lower iAOChun (ie, more hunger) than did Finnish and German breads with butter and cheese, All-Bran Plus with milk, and porridge with applesauce (Figure 2Go). When values were expressed per energy unit in the meal, porridge with applesauce had a significantly higher iAOChun (ie, less hunger) than did all other meals except All-Bran Plus and Corn Flakes, both with milk.

The reference and French breads with butter and jam and Italian cookies with coffee and milk gave rise to significantly lower iAOCpros (ie, the subject could eat more) than did Finnish and German breads with butter and cheese or porridge with applesauce (Figure 2Go). When values were expressed per energy unit in the meal, porridge with applesauce produced a significantly higher iAOCpros (ie, the subject could eat less) than did all other meals.

Relations between the breakfast meal variables and glucose and insulin
Inverse correlations were seen between iAUCglu and energy content (kJ) (R = 0.94, P < 0.001), fat (g) (R = 0.91, P < 0.001), fat (% of energy) (R = 0.77, P < 0.001), and protein (g) (R = 0.75, P < 0.01) in the breakfast meals. A positive correlation was found between iAUCglu and carbohydrate (% of energy) in the breakfast meals (R = 0.85, P < 0.001; Table 2Go). Multiple linear regression analysis showed the association between iAUCglu and the meal variables, as given in the following equation:

Formula 1(1)
in which adjusted R2 = 0.88 (P < 0.001).


View this table:
[in this window]
[in a new window]

  		TABLE 2 Relations (R2) from multiple linear regression analysis between postprandial glucose and insulin and breakfast meal variables of 14 breakfasts containing 50 g available carbohydrate1

 
Positive correlations were observed between iAUCins and protein (g) (R = 0.73, P < 0.01) and protein (% of energy) (R = 0.62, P < 0.05) in the breakfast meals (Table 2Go). By multiple linear regression, iAUCins was best described by the meal components by using the following equation:

Formula 2(2)
in which adjusted R2 = 0.55 (P < 0.01).

Relations between breakfast meal variables and appetite sensations
Energy content and the amounts of fat (g) and protein (g) in the breakfasts were the variables positively related to the increase in satiety and fullness and to the suppression of hunger and desire to eat; each explained ≤37% of the variation of the appetite sensations (Table 3Go). By multiple linear regression analysis of the appetite sensations with the meal variables as explanatory variables, the models were best fitted as in the following equations:

Formula 3(3)
in which adjusted R2 = 0.72 (P < 0.001);

Formula 4(4)
in which adjusted R2 = 0.71 (P < 0.001);

Formula 5(5)
in which adjusted R2 = 0.66 (P < 0.001); and

Formula 6(6)
in which adjusted R2 = 0.51 (P < 0.001).


View this table:
[in this window]
[in a new window]

  		TABLE 3 Relations (R2) from multiple linear regression analysis between the breakfast meal variables and subjective appetite sensations and energy intake at ad libitum lunch (EIlunch)1

 
Relations between glucose, insulin, and appetite sensations
No significant associations were observed between iAUCglu and the appetite sensations (Table 4Go). IAUCins was positively related to iAUCfull (Figure 3Go), iAOChun, and iAOCpros (P < 0.05), which explains 28–35% of the variation in these appetite sensations (Table 4Go). A larger insulin response thus translates into sensations of greater fullness and less hunger and to less desire to eat. When the results were expressed in relation to the energy content of the breakfast, neither blood glucose nor insulin was associated with appetite.


View this table:
[in this window]
[in a new window]

  		TABLE 4 Relations (R2) from multiple linear regression analysis between measures of postprandial energy intakes, appetite, and glucose and insulin after intake of 14 breakfast meals containing 50 g available carbohydrate1

 

Figure 3
View larger version (10K):
[in this window]
[in a new window]

  		FIGURE 3. Correlation between the incremental area under the curve for mean insulin response (iAUCins) and mean fullness sensation (iAUCfull) over the course of 3 h after the consumption of 14 different breakfast meals containing 50 g available carbohydrate. n = 18–28.

 
Ad libitum energy intake
No significant differences between the breakfast meals were found in EIlunch, which ranged from 3269 to 3846 kJ (P = 0.16) (Table 5Go). The energy content of the breakfast meals was inversely related to EIlunch and explained most of the variation in EIlunch (R2 = 0.60, P < 0.001). Furthermore, EIlunch was positively associated with carbohydrate (% of energy) and negatively associated with energy content (kJ), fat and protein (g), and fat (% of energy) in the breakfast meals (Table 3Go). Using multiple regressions on EIlunch with the breakfast meal variables as explanatory variables, we found no significant models. Thus the best prediction for EIlunch is given in the following equation:

Formula 7(7)
in which R2 = 0.60 (P < 0.001).


View this table:
[in this window]
[in a new window]

  		TABLE 5 Energy intake at the ad libitum lunch (EIlunch) and total energy intake at breakfast and lunch (EItotal) after 14 different breakfasts each containing 50 g available carbohydrate1

 
We found positive correlations between the appetite measures and EIlunch (R2 = 0.22–0.29, P < 0.05; data not shown). A positive correlation was also found between iAUCglu and EIlunch (R2 = 0.33, P < 0.05) (Figure 4Go), whereas insulin was not related to EIlunch (Table 4Go). When multiple linear regression analysis was used with both blood glucose and insulin concentrations, we found no significant models that could predict EIlunch, and the best prediction of EIlunch is given in the following equation:

Formula 8(8
in which R2 = 0.33 (P < 0.05).


Figure 4
View larger version (10K):
[in this window]
[in a new window]

  		FIGURE 4. Correlation between the incremental area under the curve for mean blood glucose response (iAUCglu) over the course of 3 h after the consumption of 14 different breakfast meals containing 50 g available carbohydrate and the mean ad libitum energy intake at the lunch (EIlunch) following the consumption of 14 different breakfast meals. n = 18–28.

 
When values were expressed in relation to the energy content in the breakfast, the appetite sensations explained 57–58% of the variation in EIlunch (P < 0.01), whereas neither glucose nor insulin was related to EIlunch. EItotal (breakfast + lunch) ranged from 4691 to 6319 kJ. The EIs for the German, Finnish, and reference breads with butter and cheese were significantly higher than those for the rest of the meals except reference bread with butter (Table 5Go).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
In the current study, we found large variations in the incremental glycemic, insulinemic, and appetite responses after the intake of typical European breakfasts, including meals with breads, cookies, cereals, and porridge. The breakfast with French bread with butter and jam gave rise to the lowest sensation of satiety and fullness. In contrast, rolled-oat porridge with applesauce resulted in the second largest satiety response, and when the results were expressed in relation to the energy content of the breakfasts, that meal was the most satisfying. Rolled-oat porridge with applesauce was unusual among the breakfast meals in giving rise to a combination of high blood glucose and high satiety responses over the course of the 3-h postprandial period. However, when the whole range of meals was taken into account, no relation was found between the glycemic responses and the appetite responses. Thus, these results do not lend support to the hypothesized connection between glycemic response and sensations of hunger and satiety as measured by using visual analogue scales during the 3-h postprandial period.

The glucostatic theory (11) has been debated for many years. It has received support from studies showing that transient declines in blood glucose give rise to meal initiation (21, 22). Some studies showed that carbohydrate consumption and the resulting postprandial increase in blood glucose are positively associated with satiety (13, 23), and others found a positive correlation between the duration of a rise in blood glucose and the intermeal interval (24). However, the current study, together with several other studies (12, 25–27), does not support the glucostatic theory.

More recently, it was proposed that the type of carbohydrate, separated according to GI, could be responsible for the corresponding appetite response. According to this theory, foods with a high GI would subsequently result in a high insulin response that was followed by hypoglycemia and, consequently, greater hunger (4–6, 8, 9). It is worth noticing that all 14 meals in the current study resulted in blood glucose responses, which returned to or dropped below baseline within 60–90 min, and therefore appetite was measured during the period of postprandial hypoglycemia, which is hypothesized to mediate the greater hunger after high-GI meals than after low-GI meals. However, in relation to GI, a limitation in this study could be the use of venous rather than capillary blood for glucose measurements.

In 2 reviews that included 16 (4) and 20 (6) studies, most of the included studies found that low-GI foods have a short-term effect of suppressing hunger and EI, of increasing satiety, or both. However, this hypothesis is not supported by the appetite ratings in the current study, in which the GI of the breakfast varied from 26 to 116 (28), and yet the glycemic response does not seem to be of great importance to the subsequent appetite ratings. In contrast, our results are in agreement with a systematic review by Raben (7), which concluded that data from short-term human intervention studies do not provide convincing evidence that low-GI meals have a more positive effect on satiety and hunger than do high-GI meals. Raben's review included 31 test-meal studies comparing the ability of low- and high-GI meals to induce satiety. Of these 31 studies, 15 found that hunger was less or satiety was greater with low-GI meals than with high-GI meals, whereas 16 found either no difference or the opposite result.

Some studies that observed an association between blood glucose and satiety or EI also noticed that other factors (eg, insulin and incretin hormones) also vary in association with blood glucose (12, 13, 27). The role of insulin in the short-term regulation of appetite has received renewed interest over the past few years. A study of normal-weight and obese subjects given a carbohydrate-rich meal and a fat-rich meal evaluated the relation between insulin and EI (29). The normal-weight subjects were apparently able to compensate for the caloric overload in the fat-rich meal, and the relation between insulin response and subsequent food intake in that group was significant and inverse. The relation was absent in the obese group, in whom no such compensation was observed (29). Verdich et al (30) confirmed that insulin response was inversely correlated to subsequent ad libitum food intake in lean controls but not in obese persons. A study of lean subjects using group means likewise found an inverse relation between insulin response and subsequent food intake, although a relation between satiety and insulin was not evident (12), whereas another study of lean subjects, also using group means, found strong positive correlations between insulin and satiety (13). The current study found that insulin, in the short term and to some extent, plays a role in relation to postprandial appetite in nonobese humans. However, the amounts of food, energy, and fat seem to be more important in predicting the postprandial appetite responses (explaining 51–72% of the variation) than are the resulting glycemic and insulinemic responses (which explain < 37% of the variation). Furthermore, total energy will affect many hormonal and metabolic variables after a meal in addition to insulin: any combination of these, with or without the contribution of insulin, may account for the effects of total energy on appetite.

Differences in postbreakfast appetite were significantly related to subsequent EIs, and they explained 22–29% of the variation in the subsequent ad libitum EIlunch. These correlations indicate that our appetite ratings were valid and that they predicted eating behavior, and the magnitude of the relations is in accordance with findings of other studies (12, 15). After adjustment of the appetite ratings for EI at breakfast, appetite ratings explained close to 60% of the variation in EIlunch.

Despite a wide range of energy contents in the breakfast meals (1134–2990 kJ), no significant differences were found in EIlunch. However, the range of EIlunch (3270–3846 kJ) was >500 kJ, and linear regression analysis showed that 60% of the variation in EIlunch could be accounted for by differences in EI at breakfast.

The inverse relation between insulin and hunger responses during the postprandial phase did not translate into an inverse relation between insulin and subsequent EIlunch. In contrast, a positive relation was observed between glycemic response and subsequent EIlunch. Thus, these results support the theory that a high GI gives rise to a larger EI. However, because of the fixed absolute amount of carbohydrates in the current study, a greater amount of energy in the breakfast is reflected in a greater amount of fat, protein, or both, and both the fat and protein contents of the test meals were inversely related to the glycemic response. Thus, a low energy load in the breakfast was the best single predictor of a high glycemic response (R2 = 0.88), but it was also the main predictor of EIlunch (R2 = 0.60). It therefore seems that EI at breakfast is more important for EIlunch than is the glycemic response. As a consequence, the 3 breakfast meals with the highest energy content resulted in significantly larger EItotal over the test day.

In conclusion, the current study does not support the contention that the postprandial glycemic response is associated with appetite sensations in lean humans. However, a low GI was associated with reduced EI in a subsequent meal, but this finding was biased by the fact that low glycemic response after the breakfast was also correlated with higher energy content, which was the major determinant of subsequent EI. In contrast to glycemic response, the postprandial insulin response was associated with the development of short-term appetite sensations.


   ACKNOWLEDGMENTS
 
We thank I Skovgaard for advice on the statistical analysis and the statistical interpretation of data, I Timmermann and HR Christensen for technical assistance during data collection, and C Kostecki, KG Rossen, and Y Rasmussen for providing all of the meals from the metabolic kitchen.

AF, AR, IT, and AA designed and planned the data collection; AF coordinated and supervised the collection of data; BKM and DP were responsible for the conduct of the study; BKM conducted the data analysis; JJH analyzed the blood samples for insulin; IT conducted the in vitro analysis of available carbohydrate in the products; BKM, AF, and BS wrote the draft of the manuscript; and all authors contributed to the interpretation of the data and revision of the manuscript. Kellogg's Europe was involved in the selection of test meals but was not otherwise involved in the study or in the writing of the manuscript. AA is a medical advisor for Weight Watchers. The Department of Human Nutrition receives study funding from >50 food companies. None of the authors had a personal or financial conflict of interest.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Atkins RC. Dr. Atkins' new diet revolution. New York, NY: M Evans and Company, Inc, 2002.
  2. Agatston A. The South Beach diet. New York, NY: Rodale Press, 2003.
  3. Willett W. Eat, drink and be healthy: the Harvard Medical School guide to healthy eating. New York, NY: Simon & Schuster, 2002.
  4. Ludwig DS. Dietary glycemic index and obesity. J Nutr 2000;130:280S–3S.
  5. Ludwig DS. The glycemic index: physiological mechanisms relating to obesity, diabetes, and cardiovascular disease. JAMA 2002;287:2414–23.[Abstract/Free Full Text]
  6. Roberts SB. High-glycemic index foods, hunger, and obesity: is there a connection? Nutr Rev 2000;58:163–9.[Medline]
  7. Raben A. Should obese patients be counselled to follow a low-glycaemic index diet? No. Obes Rev 2002;3:245–56.[Medline]
  8. Pawlak DB, Ebbeling CB, Ludwig DS. Should obese patients be counselled to follow a low-glycaemic index diet? Yes. Obes Rev 2002;3:235–43.
  9. Brand-Miller JC, Holt SH, Pawlak DB, McMillan J. Glycemic index and obesity. Am J Clin Nutr 2002;76(suppl):281S–5S.[Abstract/Free Full Text]
  10. Pawlak DB, Kushner JA, Ludwig DS. Effects of dietary glycaemic index on adiposity, glucose homoeostasis, and plasma lipids in animals. Lancet 2004;364:778–85.[Medline]
  11. Mayer J. Glucostatic mechanism of regulation of food intake. N Engl J Med 1953;249:13–6.[Medline]
  12. Holt SH, Brand Miller JC, Petocz P. Interrelationships among postprandial satiety, glucose and insulin responses and changes in subsequent food intake. Eur J Clin Nutr 1996;50:788–97.[Medline]
  13. Raben A, Holst JJ, Christensen NJ, Astrup A. Determinants of postprandial appetite sensations: macronutrient intake and glucose metabolism. Int J Obes Relat Metab Disord 1996;20:161–9.[Medline]
  14. Anderson GH, Woodend D. Effect of glycemic carbohydrates on short-term satiety and food intake. Nutr Rev 2003;61:S17–S26.[Medline]
  15. Flint A, Raben A, Blundell JE, Astrup A. Reproducibility, power and validity of visual analogue scales in assessment of appetite sensations in single test meal studies. Int J Obes Relat Metab Disord 2000;24:38–48.[Medline]
  16. Englyst KN, Englyst HN, Hudson GJ, Cole TJ, Cummings JH. Rapidly available glucose in foods: an in vitro measurement that reflects the glycemic response. Am J Clin Nutr 1999;69:448–54.[Abstract/Free Full Text]
  17. Møller A, Saxholt E. The composition of foods. Mørkhøj, Denmark: Levnedsmiddelstyrelsen (The National Food Agency), 1996.
  18. Energy and protein requirements. Report of a Joint FAO/WHO/UNU Expert Consultation. World Health Organ Tech Rep Ser 1985;724:1–206.[Medline]
  19. Deeg R, Kraemer W, Ziegenhorn J. Kinetic determination of serum glucose by use of the hexokinase/glucose-6-phosphate dehydrogenase method. J Clin Chem Clin Biochem 1980;18:49–52.[Medline]
  20. Albano JD, Ekins RP, Maritz G, Turner RC. A sensitive, precise radioimmunoassay of serum insulin relying on charcoal separation of bound and free hormone moieties. Acta Endocrinol (Copenh) 1972;70:487–509.[Medline]
  21. Campfield LA, Smith FJ. Transient declines in blood glucose signal meal initiation. Int J Obes 1990;14(suppl):15–31.
  22. Smith FJ, Campfield LA. Meal initiation occurs after experimental induction of transient declines in blood glucose. Am J Physiol 1993;265:R1423–9.
  23. Anderson GH, Catherine NL, Woodend DM, Wolever TM. Inverse association between the effect of carbohydrates on blood glucose and subsequent short-term food intake in young men. Am J Clin Nutr 2002;76:1023–30.[Abstract/Free Full Text]
  24. Melanson KJ, Westerterp-Plantenga MS, Saris WH, Smith FJ, Campfield LA. Blood glucose patterns and appetite in time-blinded humans: carbohydrate versus fat. Am J Physiol 1999;277:R337–45.
  25. Holt S, Brand J, Soveny C, Hansky J. Relationship of satiety to postprandial glycaemic, insulin and cholecystokinin responses. Appetite 1992;18:129–41.[Medline]
  26. Granfeldt Y, Liljeberg H, Drews A, Newman R, Bjorck I. Glucose and insulin responses to barley products: influence of food structure and amylose-amylopectin ratio. Am J Clin Nutr 1994;59:1075–82.[Abstract/Free Full Text]
  27. Lavin JH, Wittert G, Sun WM, Horowitz M, Morley JE, Read NW. Appetite regulation by carbohydrate: role of blood glucose and gastrointestinal hormones. Am J Physiol 1996;271:E209–14.
  28. Flint A, Moller BK, Raben A, et al. The use of glycaemic index tables to predict glycaemic index of composite breakfast meals. Br J Nutr 2004;91:979–89.[Medline]
  29. Speechly DP, Buffenstein R. Appetite dysfunction in obese males: evidence for role of hyperinsulinaemia in passive overconsumption with a high fat diet. Eur J Clin Nutr 2000;54:225–33.[Medline]
  30. Verdich C, Toubro S, Buemann B, Lysgard MJ, Juul HJ, Astrup A. The role of postprandial releases of insulin and incretin hormones in meal-induced satiety–effect of obesity and weight reduction. Int J Obes Relat Metab Disord 2001;25:1206–14.[Medline]
Received for publication February 14, 2006. Accepted for publication July 7, 2006.




This article has been cited by other articles:


Home page
 JPEN J Parenter Enteral NutrHome page
A. Astrup
Dietary Management of Obesity
JPEN J Parenter Enteral Nutr, September 1, 2008; 32(5): 575 - 577.
[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 Flint, A.
Right arrow Articles by Astrup, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Flint, A.
Right arrow Articles by Astrup, A.
Agricola
Right arrow Articles by Flint, A.
Right arrow Articles by Astrup, A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS