Download Breakfast glycemic index affects subsequent daily

Breakfast glycemic index affects subsequent daily energy intake in
free-living healthy children1⫺3
Anette E Buyken, Karin Trauner, Anke LB Günther, Anja Kroke, and Thomas Remer
KEY WORDS
Glycemic index, glycemic load, children, satiety, energy intake, breakfast
SUBJECTS AND METHODS
INTRODUCTION
The regular consumption of carbohydrates that elicits higher
relative postprandial blood glucose and insulin concentrations,
ie, high– glycemic index (GI) foods, was proposed to contribute
to the development of overweight, impaired insulin sensitivity,
and other components of the metabolic syndrome (1, 2). This
notion builds on the assumption that the downstream effects of
early postprandial peaks in insulin and glucose concentrations
persist into the middle postprandial phase, ie 2– 4 h after nutrient
absorption has declined, subsequently provoking reactive hypoglycemia in the late postprandial phase, which may in turn induce
hunger and higher voluntary energy intakes (2).
For adolescents, 3 experimental studies lend support to this
hypothesis, reporting remarkably higher subsequent voluntary
980
energy intakes or shorter periods of satiety after the consumption
of high-GI meals when compared with isocaloric low-GI meals
(3–5). However, the satiating effects of lower GI foods may
require 쏜3 h to emerge (6). In fact, studies in healthy adults
assessing the impact of the dietary GI on voluntary energy intake
in the early postprandial phase suggest that high-GI foods may
suppress short-term voluntary energy intake more effectively
than do low-GI foods (7–11). Such opposing effects of a high-GI
meal in the early and late postprandial phase can potentially be
ascribed to a satiating effect of blood glucose spikes in the early
postprandial phase (6), which ceases once glycemia drops to
concentrations below baseline in the later postprandial phase (2).
However, it remains to be clarified whether such opposing early
and late postprandial effects of a high-GI meal on subsequent
energy intake can also be observed in healthy children under
“real life conditions” and, if so, whether such influences extend
to the entire subsequent daily energy intake. These insights
would also be of interest for early childhood and the time of the
adiposity rebound, because these periods are proposed as being
critical for the development of later overweight (12, 13).
Detailed dietary assessment conducted as part of the
DOrtmund Nutritional and Anthropometrical Longitudinally
Designed (DONALD) Study allowed us to determine the dietary
GI and the carbohydrate content of meals, as well as the time
between meals. From these repeatedly assessed data we first
analyzed the association between the dietary GI at breakfast and
the timing of the subsequent meal and second examined whether
the timing of the next meal after breakfast modifies the effect of
the breakfast GI (GIbr) on the subsequent daytime energy intake
of 381 healthy free-living children at ages 2, 4 –5, and 7 y.
DONALD Study
The DONALD Study is a longitudinal (open cohort) study
collecting detailed data on diet, growth, development, and
1
From the Research Institute of Child Nutrition, Rheinische FriedrichWilhelms-Universität Bonn, Dortmund, Germany (AEB, KT, AK, ALBG,
and TR), and the Fulda University of Applied Science, Fulda, Germany (AK).
2
Supported by the Ministry of Science and Research of North Rhine
Westphalia, Germany.
3
Reprints not available. Address correspondence to AE Buyken, Nutrition
and Health Unit, Research Institute of Child Nutrition, Heinstüeck 11, 44225
Dortmund, Germany. E-mail: [email protected].
Received March 19, 2007.
Accepted for publication June 1, 2007.
Am J Clin Nutr 2007;86:980 –7. Printed in USA. © 2007 American Society for Nutrition
Downloaded from ajcn.nutrition.org at PENNSYLVANIA STATE UNIV PATERNO LIBRARY on May 13, 2016
ABSTRACT
Background: Experimental studies have reported that the effect of
a meal’s glycemic index (GI) on subsequent energy intake depends
on the timing of the subsequent meal.
Objective: We examined whether the timing of the next meal after
breakfast modifies the effect of the breakfast GI (GIbr) on subsequent
daytime energy intake of healthy free-living children.
Design: Analyses included 381 participants of the DOrtmund Nutritional and Anthropometrical Longitudinally Designed (DONALD)
Study who had provided weighed dietary records at ages 2, 4 –5, and 7 y.
Results: At all ages, among children who consumed their next meal
in the early postprandial phase (after 3– 4 h), children with a lower
GIbr consumed more calories throughout the remainder of the day
than did children with a higher GIbr, independent of major dietary
confounders. For the age groups 2, 4 –5, and 7 y, energy intakes in
tertiles 1 and 3 were 785 kcal (95% CI: 743– 830 kcal) and 717 kcal
(678 –758 kcal), P for trend ҃ 0.2; 993 kcal (941–1047 kcal) and 949
kcal (900 –1000 kcal), P for trend ҃ 0.05; 1255 (1171–1344) and
1166 (1090 –1247 kcal), P for trend ҃ 0.03, respectively. Conversely, among children consuming their next meal in the late postprandial phase (쏜3– 4 h), subsequent daytime energy intake was not
associated with GIbr.
Conclusion: This study confirms differential early and late postprandial effects of the GIbr on subsequent daytime energy intake for
free-living children at different ages. Interestingly, the apparent
short-term satiating effect of a higher GIbr, in particular, persisted
throughout the day, if a second breakfast was consumed
midmorning.
Am J Clin Nutr 2007;86:980 –7.
GLYCEMIC INDEX AND ENERGY INTAKE IN CHILDREN
Nutrition assessment
Food consumption in the DONALD Study was assessed with
the use of weighed 3-d dietary records as described previously
(15). Parents of the children weighed and recorded all foods and
beverages consumed, as well as leftovers, with the use of electronic food scales (앐Ҁ1 g) on 3 consecutive days. Recipes for
meals prepared at home were recorded. The packaging of commercial food products was kept. Semiquantitative recording (eg,
number of spoons, scoops) was allowed if weighing was not
possible. At the end of the 3-d record period, a dietitian visited the
family and checked the record for completeness and accuracy.
Energy and nutrient intakes were calculated with the use of the
in-house nutrient database LEBTAB (16), which is continuously
updated to include all recorded food items. LEBTAB is based on
the German standard food tables (17) with complementary data
from other national food tables (18 –20) and data obtained from
commercial food products. Because the labels of commercial
products do not usually provide information on all the nutrients
included in LEBTAB, the composition of these products was
simulated. This simulation procedure included all ingredients
declared on the label. The respective amount of each ingredient
was then estimated to match the simulated nutrient data with the
available declared nutrient data (16). From this simulation, data
for all nutrients in LEBTAB could then be calculated. For all
complex foods not included in the food tables, food composition
was calculated with the use of the recipes provided by the study
participants or with local recipes.
Glycemic index and glycemic load
Dietary GI is defined as the incremental area under the glucose
response curve after the intake of 50 g carbohydrate from a test
food compared with the area under the glucose curve induced by
the same amount of carbohydrates from ingested glucose (21).
The glycemic load (GL) has been proposed as an indicator of the
absolute glucose response induced by a serving of a food (22) and
corresponds to the amount of each carbohydrate multiplied by its
respective GI.
For the present analysis each carbohydrate-containing food
recorded in the dietary record was assigned a dietary GI according to a standardized procedure (23). In brief, foods were either
assigned 1) a published GI (24 –26) or 2) the GI of a close match,
or 3) the GI was calculated from the GI values of the food’s
ingredients with the recipes available from the in-house database. The carbohydrate content of the food was the principal
consideration when matching a particular food with one listed in
the tables. Foods providing mainly fat or protein with a carbohydrate content 쏝5 g/100 g were assigned a GI of 0 (eg, cold
meats). The mean GI and GL of each subject’s breakfast (GIbr
and GLbr) was determined by multiplying the carbohydrate content (in g) of each food consumed by the food’s GI and dividing
the sum of these products (which corresponds to the GLbr) by the
total carbohydrates consumed at breakfast.
Study population
In the DONALD population a total of 395 children had provided 3-d dietary records at each of the following ages: 2, 4 –5,
and 7 y. To be eligible for the present analysis, children had to
have at least 1 recording day, fulfilling the following criteria: 1)
an 8-h overnight fasting period before breakfast (ie, only dietary
records kept on the second and third day were eligible because
information on fasting during the preceding night was not available for the first recording day), and 2) a minimum carbohydrate
intake of 3, 4, or 5 g at breakfast at ages 2, 4 –5, and 7 y, respectively, corresponding to 앒1% of total recommended energy at
the respective age.
Meals that were consumed 쏝30 min apart were considered as
1 meal. Overall, 381 children had at least one 1-d record, meeting
these criteria at each of the following ages: 2, 4 –5, and 7 y
(number of eligible records: 721, 1451, and 748, respectively).
Among these, one 1-d record was randomly selected for each
child at each age.
Statistical analysis
Values of GIbr and GLbr were energy-adjusted with the use of
the residual method (27) and then grouped into tertiles. Differences in general characteristics, nutritional intake, or carbohydrate intake from food groups between tertiles were analyzed
with the Kruskal-Wallis tests for continuous data and the chisquare tests for categorical data. All data were presented separately for ages 2, 4 –5, and 7 y because the relations of carbohydrate (percent of energy), fat (percent of energy), protein (percent
of energy), and fiber (g/1000 kcal) intakes with tertiles of GIbr
were found to differ between the age groups (P 쏝 0.0001 to P ҃
0.0004, test for interaction in a full model).
Separate sets of multiple regression models were run to examine the association of the energy-adjusted residuals of GIbr or
GLbr with both time to second meal and subsequent energy intake
as outcome variables. Subsequent energy intakes (daily energy
intake minus energy intake at breakfast) were log-transformed to
improve homoscedasticity. Analyses, including energy-adjusted
residuals of GIbr or GLbr as continuous variables, were performed
to test for a continuous trend, whereas analyses with energyadjusted tertiles of GIbr or GLbr yielded mean times to second
meal and mean subsequent daily energy intakes for each tertile
adjusted for potential confounders. The adjusted means were the
values predicted by the model when the other variables were held
at their mean value. Potential confounders considered in multiple
Downloaded from ajcn.nutrition.org at PENNSYLVANIA STATE UNIV PATERNO LIBRARY on May 13, 2016
metabolism between infancy and adulthood. Details on the subject selection procedure and the study protocol are described
elsewhere (14). Briefly, the study was initiated in 1985 with
children and adolescents of different ages participating in ongoing anthropometric studies at the Research Institute of Child
Nutrition in Dortmund, Germany. Since then, each year 앒40 –50
healthy infants were recruited from the city of Dortmund and
surrounding communities by personal contacts, maternity wards,
or pediatric practices. The study, which is exclusively observational and noninvasive until age 18 y, was approved by the Ethics
Committee of the University of Bonn, and all examinations were
performed with parental consent.
Annual visits include a medical examination, anthropometric
measurements, and the completion of a weighed 3-d dietary
record. Standing height was measured to the nearest 0.1 cm with
the use of a digital stadiometer. Weight was measured to the
nearest 0.1 kg with the use of an electronic scale. Skinfold thickness measurements were taken on the right side of the body at the
biceps, triceps, subscapular, and suprailiac sites to the nearest 0.1
mm with the use of a Holtain caliper (Holtain Ltd, Crymych,
United Kingdom). Parents were interviewed, and their anthropometric measurements were obtained by the study nurses on
their child’s admission to the study.
981
982
BUYKEN ET AL
TABLE 1
General characteristics of the sample by tertile (T) of glycemic index at breakfast (GIbr) (total n ҃ 381)1
Age 2 y
T1 (n ҃ 127)
Girls (%)
Weight (kg)
BMI (kg/m2)
Body fatness (%)4
Overweight (%)5
Maternal school education 욷 12 y (%)
Maternal overweight (%)6
Smoking in household (%)
T3 (n ҃ 127)
Age 4 –5 y
T1 (n ҃ 127)
43
47
43
13.0 (12.0–14.0)2 12.6 (11.7–13.8) 18.1 (16.7–20.2)
16.5 (15.4–17.1) 16.1 (15.3–16.9)3 15.5 (14.7–16.4)
18.9 (17.0–21.5) 18.6 (16.6–20.6) 16.8 (15.3–19.2)
9.5
7.9
7.9
65.4
62.2
59.8
24.4
25.2
29.9
26.8
29.9
21.3
Age 7 y
T3 (n ҃ 127)
T1 (n ҃ 127)
T3 (n ҃ 127)
47
18.2 (16.3–20.0)
15.5 (14.7–16.5)
17.5 (15.7–20.3)
5.5
62.2
18.93
28.4
46
25.2 (22.8–27.9)
15.8 (14.9–17.1)
16.7 (14.0–20.4)
11.0
60.6
25.2
34
49
24.5 (22.5–28.2)
15.6 (14.8–17.2)
16.5 (14.4–20.5)
10.2
70.1
25.2
19.73
Tertiles of GIbr were based on GIbr values adjusted for energy consumed at breakfast with the use of the residual method (27).
Median; 25th–75th percentile in parentheses (all such values).
3
Significant difference between tertiles, P 쏝 0.05 (Kruskal-Wallis test for continuous variables and Mantel-Haenszel chi-square test for categorical
variables including data from all 3 tertiles).
4
Determined according to Deurenberg et al (29).
5
Defined as exceeding the 90th percentile of German reference data (30).
6
Defined as BMI 욷 25.
2
regression analyses were energy-adjusted residuals of carbohydrate, fiber, protein, fat, and sugar at breakfast, as well as the
quantity of caloric beverages (milk, cocoa, fruit juices, and soft
drinks) at breakfast. Initial analyses also considered nondietary
factors (sex, maternal overweight, maternal education, smoking
in the household, current body mass index, body fat, and body
weight), which did not, however, affect any estimate. All models
for age 4 –5 y included age as a covariate.
To test the assumption that the association between GIbr or
GLbr and the subsequent energy intake may be modified by the
time to second meal, we performed tests for interactions in nonadjusted models, categorizing time to the second meal as 울1,
쏜1–2, 쏜2–3, 쏜3– 4, or 쏜4 h. In case of a significant interaction,
separate sets of linear regression models were run for each time
category to distinguish time categories with a negative estimate
from those with a positive estimate for GIbr or GLbr. Subsequently, multiple linear regression analyses stratified for the 2
resulting time strata (early compared with late consumption of a
second meal) were performed with strata-specific energyadjusted tertiles. Finally, a concluding test for interaction with
age groups was conducted with the data from all age groups. In
the full models testing for a 3-factor interaction (eg, age group ҂
GIbr҂ time to second meal), all lower order 2-factor interactions
and main effects were kept in the full model.
Because analyses indicated no interactions between sex and
the relations of the GIbr or GLbr to time to the second meal or to
subsequent energy intake, data from girls and boys were pooled
for analyses. All statistical analyses were performed with the
SAS program, version 8.2 (SAS Institute, Cary, NC). A P value
쏝 0.05 was considered to indicate statistical significance.
Power considerations
The data from 381 children were sufficient to detect a difference of 0.5 h in the time to second meal between the lowest and
the highest tertiles of GIbr at the 0.05 level, with 80% power. For
the effect on subsequent daytime energy intake we estimated the
power for age 7 y as an example, assuming that two-thirds of the
children will, as in a pilot analysis (28), report a second meal in
the early postprandial phase. Should this assumption hold, the
numbers of children consuming their meal either in the early
postprandial phase (n 앒 250) or in the later postprandial phase
(n 앒 130) would be sufficient to detect differences of 75 kcal and
105 kcal, respectively, in subsequent energy intake between the
lowest and the highest tertiles of GIbr at the 0.05 level, with 80%
power.
RESULTS
The general characteristics (Table 1), nutritional intakes
(Table 2), and carbohydrate intakes (Table 3) from different
food groups of the 381 children at ages 2, 4 –5, and 7 y are shown
by the lowest and highest tertiles of energy-adjusted GIbr. For the
general characteristics, few differences were observed between
the tertiles (Table 1). When compared with their respective counterparts reporting the highest tertile of GIbr, at age 2 y children in
tertile 1 had a slightly higher body mass index, at age 4 –5 y
frequency of maternal overweight was higher in children reporting the lowest tertile of GIbr, and at age 7 y smoking in the
household was more frequent among children in tertile 1.
As expected, at all 3 ages, nutritional intakes of children reporting the highest tertiles of GIbr were characterized by a significantly higher GL (Table 2), resulting from both significantly
higher GIbr values and higher carbohydrate intakes in children in
tertile 3 than children in tertile 1. Furthermore, at all ages, a
higher GIbr was associated with a higher relative intake of carbohydrates and sugar and a lower relative intake of fat and protein. At age 2 y, a lower GIbr reflected a high-fat, high-protein
breakfast providing no fiber. By age 7 y, a lower GIbr reflected a
breakfast with a relatively high-carbohydrate content accompanied by a higher fiber intake and a lower sugar intake. Overall,
compared with age 2 y, GIbr values were higher at ages 4 –5 and
7 y, with the difference between tertile 1 and tertile 3 becoming
smaller.
At age 2, milk and dairy products constituted most of the
carbohydrates consumed at breakfast by children reporting the
lowest tertile of GIbr, whereas bread or rolls with sweet spread
provided most of the carbohydrates among children reporting the
highest tertile of GIbr (Table 3). By age 7 y, the difference in GIbr
was mostly attributable to a difference in the consumption of
milk and dairy products. Other foods such as fruit, fruit juices, or
Downloaded from ajcn.nutrition.org at PENNSYLVANIA STATE UNIV PATERNO LIBRARY on May 13, 2016
1
983
GLYCEMIC INDEX AND ENERGY INTAKE IN CHILDREN
TABLE 2
Nutritional intake at breakfast by tertiles (T) of glycemic index at breakfast (GIbr) (total n ҃ 381)1
Age 2 y
T1 (n ҃ 127)
T3 (n ҃ 127)
2
T1 (n ҃ 127)
Age 7 y
T3 (n ҃ 127)
2
T1 (n ҃ 127)
T3 (n ҃ 127)
30 (27–34)
3.4 (2.6–5.3)
11.8 (9.0–15.4)
31.1 (29.4–41.3)
56 (53–61)
12.8 (8.6–16.7)2
22.1 (14.8–29.5)2
55.3 (46.9–64.9)2
44 (35–47)
10.7 (5.1–17.0)
24.7 (15.1–37.2)
43.5 (36.7–51.5)
60 (58–63)
18.9 (13.1–26.6)2
32.1 (21.2–43.6)2
57.0 (48.0–67.0)2
47 (44–49)
14.8 (10.0–22.2)
32.2 (22.1–46.5)
48.2 (40.8–58.6)
62 (60–65)2
23.2 (16.2–30.3)2
35.6 (25.7–48.1)2
58.9 (47.7–66.1)2
0 (0–12.1)
0 (0–2.4)
50.2 (40.5–50.2)
20.5 (17.0–20.5)
146 (106–179)
9.1 (1.5–17.4)2
9.9 (5.7–15.8)2
32.3 (25.7–40.5)2
11.5 (8.0–14.5)2
178 (118–224)2
8.6 (0–14.9)
5.6 (2.2–9.0)
41.3 (30.8–46.9)
15.7 (12.7–19.5)
243 (158–319)
11.6 (3.3–17.7)2
9.7 (6.6–14.7)2
30.7 (22.9–39.6)2
11.6 (8.2–14.7)2
234 (153–319)2
10.4 (3.5–18.1)
7.8 (4.7–12.2)
35.9 (27.1–43.3)
14.1 (11.1–17.7)
283 (197–354)
13.4 (47.7–66.1)
7.4 (5.9–10.8)2
29.5 (21.5–38.8)2
11.6 (9.3–13.9)2
253 (167–351)2
1
Values are medians; 25th–75th percentiles in parentheses. Tertiles of GIbr were based on GIbr values adjusted for energy consumed at breakfast with the
use of the residual method (27). GL, glycemic load.
2
Significant difference between tertiles, P 쏝 0.05 (Kruskal-Wallis test, including data from all 3 tertiles).
breakfast cereals made only minor or no contributions to carbohydrate intake at breakfast at all ages.
At all ages, lower values of GIbr were independently associated with a shorter time to the second meal (P for continuous
trends ҃ 0.01– 0.05, adjusted for intake of carbohydrate, protein,
energy, and caloric beverages at breakfast) (Figure 1); however,
the strength of this association differed between the age groups
(P ҃ 0.03 for interaction of age group ҂ GIbr tertiles in the full
model). Further adjustment for intakes of fiber, sugar, or fat at
breakfast did not affect the results (data not shown).
In nonadjusted models higher values of GIbr were related to
lower subsequent energy intake (P ҃ 0.02, P ҃ 0.01, and P ҃
0.06 at ages 2, 4 –5, and 7 y, respectively). As hypothesized, at all
ages, categories of time to second meal showed a significant
interaction with GIbr in its association with subsequent energy
intake (P for interaction ҃ 0.004, P 쏝 0.0001, and P ҃ 0.0001 at
ages 2, 4 –5, and 7 y, respectively). Time categories with a negative estimate for the association between GIbr and subsequent
energy intakes were distinguished from those with a positive
estimate, yielding 2 strata of consumption of the next meal in the
early and late postprandial phase, ie, 울4 h compared with 쏜4 h
at ages 2 and 4 –5 y and 울3 h compared with 쏜3 h at age 7 y.
At all ages, among children consuming their next meal earlier
(after 3 or 4 h), children reporting the lowest GIbr tertile consumed more calories throughout the remainder of the day than
did their counterparts in the higher GIbr tertiles; however,
the continuous trend was only significant at ages 4 –5 and 7 y
(Figure 2). Conversely, among children consuming their next
meal 쏜3 or 4 h after breakfast, subsequent daily energy intake
was not associated with GIbr, despite a nonsignificant suggestion
of a higher energy intake among children in the highest GIbr
tertile at age 2 y. In the full model with data from all ages, the
absence of a 3-factor interaction (P ҃ 0.8 for age group ҂ GIbr
tertiles ҂ time to second meal strata) and the presence of a
2-factor interaction for GIbr tertiles ҂ time to second meal strata
(P ҃ 0.02) underpin that the time to second meal influences the
association between GIbr and subsequent energy intake in a similar way at all 3 ages. Our analysis adjusts for potential differences between the GIbr tertiles in carbohydrate, fiber, energy, and
caloric beverages at breakfast. Further adjustments for intakes of
protein, sugar, or fat at breakfast did not affect the results (data
not shown).
When analyses were repeated with GLbr in place of GIbr, GLbr
was neither related to the time to the second meal (P ҃ 0.3, P ҃
0.5, and P ҃ 0.2 at ages 2, 4 –5, and 7 y, respectively, adjusted for
fat, energy, and caloric beverages at breakfast) nor to subsequent
daytime energy intake (P ҃ 1.0, P ҃ 0.7, and P ҃ 0.3, respectively, adjusted for fat, energy, and caloric beverages at breakfast). In addition, no interaction between GLbr and time to next
meal was observed (P ҃ 0.9, P ҃ 0.2, and P ҃ 0.7 at ages 2, 4 –5,
and 7 y, respectively, unadjusted) in its association with subsequent daytime energy intake.
TABLE 3
Carbohydrate intake from different food groups at breakfast by tertile (T) of glycemic index at breakfast (GIbr) (total n ҃ 381)1
Age 2 y
T1 (n ҃ 127)
Milk and dairy products (g)
Breakfast cereals (g)
Bread and rolls (g)
Fruit and fruit juices (g)
Sugary foods (g)3
10.0 (7.1–13.0)
0 (0–0)
0 (0–0)
0 (0–0)
0 (0–0)
Age 4 –5 y
T3 (n ҃ 127)
2
2.2 (0–5.9)
0 (0–0)2
8.9 (0–13.2)2
0 (0–3.3)2
2.5 (0–6.1)2
T1 (n ҃ 127)
11.3 (7.6–15.9)
0 (0–0)
5.2 (0–13.0)
0 (0–3.6)
0 (0–5.0)
Age 7 y
T3 (n ҃ 127)
2
2.0 (0–5.9)
0 (0–6.6)2
12.5 (0–22.0)2
0 (0–3.9)
4.9 (0–10.4)2
T1 (n ҃ 127)
T3 (n ҃ 127)
10.4 (5.4–16.7)
0 (0–0)2
9.6 (0–15.9)
0 (0–7.7)
0 (0–7.5)
3.8 (0–7.2)2
0 (0–22.5)2
11.7 (0–25.9)2
0 (0–0)2
0 (0–9.6)2
1
Values are medians; 25th–75th percentiles in parentheses. Tertiles of GIbr were based on GIbr values adjusted for energy consumed at breakfast with the
use of the residual method (27).
2
Significant difference between tertiles, P 쏝 0.05 (Kruskal-Wallis test, including data from all 3 tertiles).
3
Soft drinks, cakes, cookies, and sweet spreads.
Downloaded from ajcn.nutrition.org at PENNSYLVANIA STATE UNIV PATERNO LIBRARY on May 13, 2016
GI
GL (g)
Carbohydrate
Carbohydrate (% of
energy)
Sugar (% of energy)
Fiber (g/1000 kcal)
Fat (% of energy)
Protein (% of energy)
Energy (kcal)
Age 4 –5 y
984
BUYKEN ET AL
DISCUSSION
3.5
P for continuous trend = 0.0496
3.0
2.5
2.0
0
=
Low GI
(T1)
Medium
GI (T2)
High GI
(T3)
Time to second meal (h)
Tertiles of breakfast GI
3.5
P for continuous trend = 0.01
3.0
2.5
2.0
=
0
Low GI
(T1)
Medium
GI (T2)
High GI
(T3)
Tertiles of breakfast GI
Age 7 y
Time to second meal (h)
3.5
P for continuous trend = 0.03
3.0
2.5
2.0
=
0
Low GI
(T1)
Medium
GI (T2)
High GI
(T3)
Tertiles of breakfast GI
FIGURE 1. Time to second meal after breakfast by tertile (T) of glycemic
index (GI) at breakfast in 381 children at age 2 y, age 4 –5 y, and age 7 y. Data
are presented as mean and 95% CI, adjusted for energy, energy-adjusted
residuals of carbohydrate and protein, and consumption of caloric beverages
(in g) at breakfast. Models at age 4 –5 y also consider the age of the participants. P for continuous trend refers to the P values obtained in linear regression models with the GI at breakfast as a continuous variable.
This study also confirms differential early and late postprandial effects of the GIbr on subsequent daytime caloric intake for
free-living children. Among free-living children who consumed
a second meal at midmorning, a higher dietary GI at breakfast
yielded significantly lower subsequent daily caloric intakes.
Conversely, GIbr was not related to subsequent daytime energy
intake among children who consumed a second meal in the late
postprandial phase. These differential associations were consistently observed at ages 2, 4 –5, and 7 y, despite different carbohydrate sources being responsible for a lower or higher GIbr, and
were independent of major dietary factors known to influence
satiety, ie, consumption of protein, fat, sugar, fiber, or caloric
beverages at breakfast (31, 32).
Suppressed energy intakes 1–2 h after the consumption of
high-GI carbohydrates were also reported by most (7–11) but not
all (33, 34) short-term experimental studies in adults. Interestingly, in our study the potential short-term satiating effect of a
high-GI meal appears to persist throughout the remainder of the
day, an aspect not analyzed in the short-term experimental studies. Elevated blood glucose concentrations after consumption of
a high-GI meal are generally accompanied by peaks in insulin
concentrations (2), which could be involved in daytime satiety
signaling (35). Appropriate insulin signaling is, in addition to
peripheral and hypothalamic neuropeptides, required for effective anorectic action of the satiety signal leptin. Because high-GI
meals may elicit both postprandial insulin peaks (36) and increases in leptin concentrations (37, 38), it could thus be speculated that it is the combination of these effects after the ingestion
of a high-GI breakfast that results in a persistent reduction of
energy intake throughout the remainder of the day, if a second
meal is consumed midmorning. It remains to be further explained
whether a habitual increase in insulin secretion on a morning-tomorning basis could contribute to the established involvement of
insulin in the long-term regulation of energy balance (35).
The GI of a meal is only one characteristic of carbohydrates
that influences satiety signaling. Carbohydrates also affect the
stimulation of several hormones involved in food regulation
(35). Our finding that subsequent energy intake was clearly related to GIbr (ie, the potential relative increase in glycemia) but
not to GLbr (ie, the potential absolute increase in glycemia) is
nonetheless in accordance with a review concluding that absolute
glucose concentrations have no straightforward relation to appetite, whereas short-term dynamic changes in blood glucose
concentration are strongly related to satiety (35), as originally
proposed in the glucostatic theory (39).
In the present observational study, children who consumed
their next meal in the later postprandial phase did not report
higher energy intakes after a breakfast with a higher GI, as suggested by experimental studies in normal weight and obese adolescents (3, 4). This discrepancy may lie in the nature of our
observational study in that the GI in the highest GI tertiles [GI
ranging from 56 to 62, values considered a medium GI (24)] may
not have been high enough for the described hunger-inducing
drop in glycemia to values below baseline to occur. In addition,
at least at age 2 y, the nonsignificance of the suggested direct
association between GIbr and subsequent energy intake among
children who consumed their next meal 쏜4 h after breakfast may
be due to lack of power in this group of children (n ҃ 33).
Downloaded from ajcn.nutrition.org at PENNSYLVANIA STATE UNIV PATERNO LIBRARY on May 13, 2016
Time to second meal (h)
Age 2 y
985
GLYCEMIC INDEX AND ENERGY INTAKE IN CHILDREN
Subsequent energ y intake
(kcal/d)
Children consuming their
next meal < 4 h later (n = 33)
1000
1000
900
900
P for continuous trend = 0.1
P for continuous trend = 0.2
800
800
700
700
600
600
500
500
=
400
0
=
0
400
Low GI
(T1)
Medium
GI (T2)
High GI
(T3)
Low GI
(T1)
Medium
GI (T2)
High GI
(T3)
Tertiles of GI at breakfast
Tertiles of GI at breakfast
Children consuming their
next meal < 4 h later (n = 331)
Children consuming their
next meal < 4 h later (n = 50)
1100
1100
1000
1000
P for continuous trend = 0.9
900
900
P for continuous trend = 0.0497
800
800
700
700
600
0
=
0
600
Low GI
(T1)
Medium
GI (T2)
=
Low GI
(T1)
High GI
(T3)
Medium
GI (T2)
High GI
(T3)
Tertiles of breakfast GI
Tertiles of breakfast GI
Subsequent energ y intake
(kcal/d)
Age 7 y
Children consuming their
next meal < 3 h later (n = 251)
1400
P for continuous trend = 0.03
1300
Children consuming their
next meal < 3 h later (n = 130)
1400
1300
1200
1200
1100
1100
1000
1000
=
900
0
P for continuous trend = 0.2
=
0
900
Low GI
(T1)
Medium
GI (T2)
High GI
(T3)
Tertiles of breakfast GI
Low GI
(T1)
Medium
GI (T2)
High GI
(T3)
Tertiles of breakfast GI
FIGURE 2. Subsequent daily energy intake by tertile (T) of glycemic index (GI) at breakfast in 381 children at age 2 y, age 4 –5 y, and age 7 y by strata
of time to the second meal. Data are presented as mean and 95% CI, adjusted for energy, energy-adjusted residuals of carbohydrate and fiber, and consumption
of caloric beverages (in g) at breakfast. Models at age 4 –5 y also consider the age of the participants. P for continuous trend refers to the P values obtained in
linear regression models with the GI at breakfast as a continuous variable.
Downloaded from ajcn.nutrition.org at PENNSYLVANIA STATE UNIV PATERNO LIBRARY on May 13, 2016
Subsequent energ y intake
(kcal/d)
Age 2 y
Children consuming their
next meal < 4 h later (n = 348)
986
BUYKEN ET AL
management of overweight in childhood or adolescence. Our
data would at most suggest that children who habitually consume
a second breakfast at midmorning are unlikely to benefit from a
low-GI breakfast with respect to their satiety.
In conclusion, this study confirms differential early and late
postprandial effects of the GIbr on subsequent daytime energy
intake for free-living children at different ages. Interestingly, the
apparent short-term satiating effect of a higher GIbr, in particular,
persisted throughout the day, if a second breakfast was consumed
midmorning.
We thank Christa Chahda, Ruth Schäfer, and Wolfgang Sichert-Hellert for
collecting and processing the nutritional data. We also thank Guo Cheng for
assigning further glycemic indexes and Nadina Karaolis-Danckert for proof
reading the text.
The author’s responsibilities were as follows—AEB and KT: conceived
the project, performed the data analyses, and drafted the manuscript; ALBG
and AK (lead epidemiologist on the DONALD Study): provided critical input
on the data analyses and on the early versions of the manuscript; AEB and TR:
supervised the study. All authors contributed to interpretation of the data and
revision of the manuscript. None of the authors had any personal or financial
conflicts of interest.
REFERENCES
1. Brand-Miller JC, Holt SH, Pawlak DB, McMillan J. Glycemic index and
obesity. Am J Clin Nutr 2002;76(suppl):281S–5S.
2. Ludwig DS. The glycemic index: physiological mechanisms relating to
obesity, diabetes, and cardiovascular disease. JAMA 2002;287:2414 –
23.
3. Ludwig DS, Majzoub JA, Al-Zahrani A, Dallal GE, Blanco I, Roberts
SB. High glycemic index foods, overeating, and obesity. Pediatrics
1999;103:E26.
4. Warren JM, Henry CJ, Simonite V. Low glycemic index breakfasts and
reduced food intake in preadolescent children. Pediatrics 2003;
112:e414.
5. Ball SD, Keller KR, Moyer-Mileur LJ, Ding YW, Donaldson D, Jackson
WD. Prolongation of satiety after low versus moderately high glycemic
index meals in obese adolescents. Pediatrics 2003;111:488 –94.
6. Anderson GH, Woodend D. Effect of glycemic carbohydrates on shortterm satiety and food intake. Nutr Rev 2003;61(suppl):S17–26.
7. 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.
8. 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.
9. Woodend DM, Anderson GH. Effect of sucrose and safflower oil preloads on short term appetite and food intake of young men. Appetite
2001;37:185–95.
10. Anderson GH, Woodend D. Consumption of sugars and the regulation of
short-term satiety and food intake. Am J Clin Nutr 2003;78(suppl):
843S–9S.
11. Rogers PJ, Carlyle JA, Hill AJ, Blundell JE. Uncoupling sweet taste and
calories: comparison of the effects of glucose and three intense sweeteners on hunger and food intake. Physiol Behav 1988;43:547–52.
12. Dietz WH. Periods of risk in childhood for the development of adult
obesity–what do we need to learn? J Nutr 1997;127(suppl):1884S– 6S.
13. Yanovski JA. Rapid weight gain during infancy as a predictor of adult
obesity. Am J Clin Nutr 2003;77:1350 –1.
14. Kroke A, Manz F, Kersting M, et al. The DONALD Study: history,
current status and future perspectives. J Nutr 2004;43:45–54.
15. Kersting M, Sichert-Hellert W, Lausen B, Alexy U, Manz F, Schoch G.
Energy intake of 1 to 18 year old German children and adolescents. Z
Ernahrungswiss 1998;37:47–55.
16. Sichert-Hellert W, Kersting M, Chahda C, Schaefer R, Kroke A. German
food composition database for dietary evaluations in children and adolescents. J Food Comp Anal 2007;20:63–70.
17. Souci SW, Fachmann W, Kraut H. Food composition and nutrition
Downloaded from ajcn.nutrition.org at PENNSYLVANIA STATE UNIV PATERNO LIBRARY on May 13, 2016
Whether this effect exists, is, however, of little practical relevance, because only a few 2-y-old children appear to consume
their next meal 쏜4 h after breakfast. At age 7 y, we would have
been sufficiently powered to detect a difference well below the
20 – 80% caloric difference suggested by experimental studies
that compared subsequent voluntary energy intakes after the
consumption of high-GI meals and isocaloric low-GI meals (3,
4). Alternatively, the experimental studies may have been confounded by concomitant differences in protein content between
the low- and high-GI diets (3, 4) because protein is known to have
notable satiating properties (31). In addition, the hungerinducing drop in glycemia several hours after consumption of a
high-GI breakfast may, if present, have affected only the caloric
intake at the subsequent meal but not the overall subsequent
daytime energy intake.
In the present study, children who consumed a breakfast with
a lower GI consumed their second meal earlier than children who
consumed a breakfast with higher GI. Although this finding
contradicts results reported from experimental studies in children and adolescents (3, 4), it is in accordance with the previously
discussed short-term satiating properties of a high-GI meal,
which were more relevant in this cohort, because most children
consumed their second meal within 3– 4 h after breakfast.
The main advantages of this observational study include its
relatively large sample size with dietary data assessed at ages 2,
4 –5, and 7 y with a prospective weighed 3-d dietary record,
which allows the determination of the dietary GI of meals, as well
as our ability to control for major nutritional factors known to
influence satiety (31, 32). In addition, because the same children
were analyzed at the 3 ages, potential bias otherwise introduced
by differences in socioeconomic variables was controlled for.
However, 2 limitations require specific attention when interpreting our findings. First, we had to calculate GI values from their
ingredients for a large percentage of the carbohydrate-containing
foods, a procedure which was controversially discussed (40 –
44). However, several investigative groups have shown that the
GI of a whole diet or mixed meal can be accurately estimated
from the GI value of the individual foods it contains (40, 42, 45).
In the absence of GI values for all foods and meals recorded in
observational studies, GI calculation thus appears the only valid
and feasible approach for epidemiologic studies. Second, we
were not able to adjust for physical activity. If a lower GI at
breakfast would reflect a healthier lifestyle, a lack of adjustment
for physical activity could have confounded our results in that
children in the lowest GIbr tertile may have had higher energy
intakes as a result of higher levels of physical activity.
Our findings may have implications for strategies to prevent
overweight in children: among children consuming a second
meal within 3 h after the first, a high GI of the first meal may
result in a lower subsequent daily energy intake. We can, however, not preclude adverse influences of a high-GI breakfast
among persons who do not ingest a second breakfast or consume
breakfasts with much higher GI values than in our study, which
may be more frequently the case among adolescents (46). As a
corollary, the differential effects of a meals’ GI on subsequent
energy intake, depending on the timing of the subsequent meal,
among healthy children may explain why strong associations
between the dietary GI and body composition are rarely reported
for children and adolescents (47– 49). Clearly, our observational
data in healthy children cannot be interpreted as a general argument against the usefulness of diets low in GI, GL, or both in the
GLYCEMIC INDEX AND ENERGY INTAKE IN CHILDREN
18.
19.
20.
21.
22.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
between the glycaemic response to breakfast cereals and subjective
appetite and food intake. Nutr Res 1997;17:1249 – 60.
Kaplan RJ, Greenwood CE. Influence of dietary carbohydrates and glycaemic response on subjective appetite and food intake in healthy elderly
persons. Int J Food Sci Nutr 2002;53:305–16.
de Graaf C, Blom WA, Smeets PA, Stafleu A, Hendriks HF. Biomarkers
of satiation and satiety. Am J Clin Nutr 2004;79:946 – 61.
Galgani J, Aguirre C, Diaz E. Acute effect of meal glycemic index and
glycemic load on blood glucose and insulin responses in humans. Nutr J
2006;5:22.
Herrmann TS, Bean ML, Black TM, Wang P, Coleman RA. High glycemic index carbohydrate diet alters the diurnal rhythm of leptin but not
insulin concentrations. Exp Biol Med (Maywood) 2001;226:1037– 44.
Pawlak DB, Bryson JM, Denyer GS, Brand-Miller JC. High glycemic
index starch promotes hypersecretion of insulin and higher body fat in
rats without affecting insulin sensitivity. J Nutr 2001;131:99 –104.
Mayer J. Glucostatic mechanism of regulation of food intake. N Engl
J Med 1953;249:13– 6.
Wolever TM, Yang M, Zeng XY, Atkinson F, Brand-Miller JC. Food
glycemic index, as given in glycemic index tables, is a significant determinant of glycemic responses elicited by composite breakfast meals.
Am J Clin Nutr 2006;83:1306 –12.
Wolever TM, Jenkins DJ. The use of the glycemic index in predicting the
blood glucose response to mixed meals. Am J Clin Nutr 1986;43:167–
72.
Bornet FR, Costagliola D, Rizkalla SW, et al. Insulinemic and glycemic
indexes of six starch-rich foods taken alone and in a mixed meal by type
2 diabetics. Am J Clin Nutr 1987;45:588 –95.
Hollenbeck CB, Coulston AM. The clinical utility of the glycemic index
and its application to mixed meals. Can J Physiol Pharmacol 1991;69:
100 –7.
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.
Chew I, Brand JC, Thorburn AW, Truswell AS. Application of glycemic
index to mixed meals. Am J Clin Nutr 1988;47:53– 6.
Sjoberg A, Hallberg L, Hoglund D, Hulthen L. Meal pattern, food
choice, nutrient intake and lifestyle factors in The Goteborg Adolescence
Study. Eur J Clin Nutr 2003;57:1569 –78.
Nielsen BM, Bjornsbo KS, Tetens I, Heitmann BL. Dietary glycaemic
index and glycaemic load in Danish children in relation to body fatness.
Br J Nutr 2005;94:992–7.
Scaglioni S, Stival G, Giovannini M. Dietary glycemic load, overall
glycemic index, and serum insulin concentrations in healthy schoolchildren. Am J Clin Nutr 2004;79:339 – 40.
Hui LL, Nelson EA. Meal glycaemic load of normal-weight and overweight Hong Kong children. Eur J Clin Nutr 2006;60:220 –7.
Downloaded from ajcn.nutrition.org at PENNSYLVANIA STATE UNIV PATERNO LIBRARY on May 13, 2016
23.
tables. 5th ed. Stuttgart, Germany: Medpharm Scientific Publishing,
1994.
Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgat DAT.
Mc-Cance and Widdowson’s the composition of foods. 5th and extended
edition ed. Cambridge, United Kingdom: Royal Society of Chemistry,
Ministry of Agriculture, Fisheries, and Food, 1992.
USDA. Internet: www.nal.usda.gov/fnic/foodcomp/search (accessed 8
August 2007).
Voorlichtingsbureau_voor_de_Voeding_(ed). Nevo Tabel. Den Haag,
Netherlands: Nederlands voedingsstoffenbestand, 1990.
Jenkins DJ, Wolever TM, Taylor RH, et al. Glycemic index of foods: a
physiological basis for carbohydrate exchange. Am J Clin Nutr 1981;
34:362– 6.
Salmeron J, Ascherio A, Rimm EB, et al. Dietary fiber, glycemic load,
and risk of NIDDM in men. Diabetes Care 1997;20:545–50.
Buyken AE, Dettmann W, Kersting M, Kroke A. Glycaemic index and
glycaemic load in the diet of healthy schoolchildren: trends from 1990 to
2002, contribution of different carbohydrate sources and relations to
dietary quality. Br J Nutr 2005;94:796 – 803.
Foster-Powell K, Holt SH, Brand-Miller JC. International table of glycemic index and glycemic load values: 2002. Am J Clin Nutr 2002;76:
5–56.
Chantelau E. The glycaemic index of carbohydrate foods: an update
from a diabetologist’s perspective. Aktuel Ernaehr Med 2000;25:176 –
85.
Henry CJ, Lightowler HJ, Strik CM, Renton H, Hails S. Glycaemic index
and glycaemic load values of commercially available products in the
UK. Br J Nutr 2005;94:922–30.
Hu FB, Stampfer MJ, Rimm E, et al. Dietary fat and coronary heart
disease: a comparison of approaches for adjusting for total energy intake
and modeling repeated dietary measurements. Am J Epidemiol 1999;
149:531– 40.
Buyken AE, Trauner K, Remer T, Kroke A. Breakfast glycemic index
affects subsequent daily caloric intake in healthy children. Results from
the DONALD Study. Obes Rev 2006;7:94 (abstr).
Deurenberg P, Weststrate JA, Seidell JC. Body mass index as a measure
of body fatness: age- and sex-specific prediction formulas. Br J Nutr
1991;65:105–14.
Kromeyer-Hauschild K, Wabitsch M, Kunze D, et al. Perzentile für den
Body-Mass-Index für das Kindes- und Jugendalter unter Heranziehung
verschiedener deutscher Stichproben. [Percentiles of body mass index in
children and adolescents evaluated from regional German studies.]
Monatsschr Kinderheilkd 2001;149:807–18 (in German).
Astrup A. The satiating power of protein–a key to obesity prevention?
Am J Clin Nutr 2005;82:1–2.
Anderson GH. Sugars-containing beverages and post-prandial satiety
and food intake. Int J Obes (Lond) 2006;30(suppl):S52–9.
Stewart SL, Black RM, Wolever TM, Anderson GH. The relationship
987