European Journal of Clinical Nutrition (1999) 53, 360±366
ß 1999 Stockton Press. All rights reserved 0954±3007/99 $12.00
http://www.stockton-press.co.uk/ejcn
Bioavailability of starch in bread rich in amylose : metabolic
responses in healthy subjects and starch structure
C Hoebler1*, A Karinthi1, H Chiron2, M Champ1 and J-L Barry1
1
Centre de Recherche en Nutrition Humaine, CHU-HoÃtel-Dieu, 44093 Nantes Cedex 01, France and; 2Laboratoire de Technologie et de
Biochimie des Proteines, INRA, 44316 Nantes, France
Objective: This study investigated whether postprandial metabolic responses to bread could be lowered by
substituting high amylose maize starch for a part of the ¯our.
Design and subjects: Eight healthy subjects consumed test meals of equivalent nutritional composition based on
white wheat bread, bread rich in amylose (HAWB) and spaghetti as a breakfast meal. Blood samples were
collected to measure insulin and glucose concentration during two hours after consumption. The degree of starch
crystallinity was investigated by X-ray diffraction and DSC analysis.
Results: HAWB produced low glycaemic (60 18) and insulinaemic (57 20) indexes similar to those of
spaghetti (83 46, 61 16). In vitro amylase hydrolysis of the three foods showed that high amylose content in
HAWB signi®cantly lowered starch degradation in bread without affecting hydrolysis kinetics. Addition of
amylose in dough increased the resistant starch content of HAWB (14% of dry matter). The resistant starch
fraction was mainly composed of crystalline amylose (B-type X-ray diffraction pattern, melting temperature
105 C) attributable to native high amylose maize starch incompletely gelatinised during bread-cooking.
Conclusions: Bread produced by the substitution of high amylose maize starch for a part of wheat ¯our showed a
low glycaemic index. Resistant starch in HAWB corresponded to native crystalline amylose not gelatinised
during normal bread-processing conditions.
Sponsorship: Barilla, via Montova 166, 43100, Parma, Italy.
Descriptors: starch; amylose; bread; blood glucose; blood insulin
Introduction
Starchy foods have an intrinsic capacity to induce a wide
scattering in the glycaemic index (Fosterpowell & Miller,
1995). There is increasing evidence that a low glycaemic
index or a lente starch diet provides a potential bene®cial
effect by improving metabolic control of hyperlipidaemia
in diabetic patients (Jenkins et al, 1987; Jenkins et al, 1988)
as well as in healthy subjects (Jenkins et al, 1994). The
range of foods with a low glycaemic index is very limited.
Spaghetti is recognized as the main lente starchy diet, but
the range of foods with low glycaemic response needs to be
diversi®ed. Two strategies could be used to produce new
products with low metabolic responses. The ®rst involves a
modi®cation of the starch microenvironment to limit its
accessibility to amylase. The best example is ®rstly the
protein network which entraps starch in pasta, creating a
physical barrier that alters or delays the access of starch to
amylase (Fardet et al, 1998). Secondly a modi®cation of the
crystalline organisation of starch to generate resistant
starch.
Bread constitutes a large part of the diet, consequently
various studies have attempted to process bread in order to
limit starch degradation and=or glucose absorption and
therefore reduces the metabolic response following bread
ingestion. The two ways of processing bread have led to
*Correspondence: Dr C Hoebler, Institut National de la Recherche
Agronomique, Laboratoire des Fonctions Digestives et de Nutrition
Humaine, BP 71627, 44316, Nantes Cedex 03, France.
Received 19 September 1998; revised 13 November 1998;
accepted 17 November 1998
strategies for modifying the microstructure of bread, the
degree of starch crystallisation, or both. The content of
resistant starch in bread products is usually very low
(Englyst et al, 1992) and is related to recipe and baking
conditions. The modi®cation of bread microstructure was
achieved by incorporating the intact kernel or adding
viscous ®bre in paste, resulting in a ¯attening of postprandial glucose curves (Ellis et al, 1991; Liljeberg et al, 1992).
The incorporation of intact cereal kernels (Holm & BjoÈrck,
1992; Liljeberg et al, 1992; Holm & BjoÈrck, 1992) or
cracked kernels (Holt & Miller, 1994) introduces in bread
encapsulated starch. Viscous dietary ®bres such as guar
gum are also closely associated with starch, and their
addition at the bread matrix level acts as a physical barrier
to starch digestion (Brennan et al, 1996). Resistant starch in
crystalline form has been generated by raw starch granules,
retrograded amylose, or a mixed structure (Englyst et al,
1992). Some studies have attempted to capitalise on the
functional properties of amylose for the generation of
resistant starch by adding high amylose maize starch in
bakery products (Behall et al, 1989; Weststrate & van
Amelsvoort, 1993; Heijnen et al, 1995; Granfeldt et al,
1995). Moreover, inconsistent results are obtained on the
effect of high amylose content in various kinds of bread on
postprandial glucose and insulin (Weststrate & van Amelsvoort, 1993; Granfeldt et al, 1995; Heijnen et al, 1995). In
these studies (Behall et al, 1989; Weststrate & van Amelsvoort, 1993; Heijnen et al, 1995; Granfeldt et al, 1995), the
recipes and the sizes of wheat based-bread (baguette, roll or
bread) or maize based-bread (arepas), the baking conditions
and the cooling treatments are various. The cooking tem-
Bioavailability of starch in bread rich in amylose
C Hoebler et al
peratures used for processing bread are always high (175±
240 C) in order to achieve the gelatinisation of amylose,
however the degree of starch crystallinity and the crystalline structure in processed bread products have not always
been investigated. The ability to relate a low metabolic
response with the crystalline structure of starch truly
involved in the alteration of starch digestion has not been
completely explored.
The aim of this study was to modify starch bioavailability in order to produce a palatable bread eliciting low
postprandial metabolic responses. Experimental bread was
®rst produced by incorporating high amylose maize starch
into the dough (HAWB). In the latter case, the bread was
cooled at room temperature and freeze-dried to induce
amylose retrogradation. The glucose and insulin responses
of HAWB were then compared respectively to rapid (white
wheat bread: WWB) and lente (spaghetti) carbohydrates.
Finally, in vitro enzymatic hydrolysis, X-ray diffraction
analysis and DSC studies were performed to determine the
physico-chemical characterisation of starch with the three
substrates and elucidate the factors responsible for metabolic responses.
Subjects and methods
Cereal products
White wheat bread (WWB) and durum wheat spaghetti
were provided by Barilla (Parma, Italy). The bread was
processed with white wheat ¯our and frozen ( 7 20 C) in
slices for 6±12 months. Spaghetti (1001g) was cooked for
10 min in 3 l of boiling water (Evian , Danone, France)
containing 0.7% NaCl and then drained. Before analysis,
cooked spaghetti samples were frozen ( 7 20 C) and
freeze-dried. Test bread containing a high maize amylose
starch (HAWB) was prepared at INRA (Nantes, France).
HAWB was baked
from 900 g white wheat ¯our (type 55),
1
600 g Eurylon (high-amylose maize starch, 65±75% of
amylose: Roquette FreÁres, Lestrem, France), 65 ml glycerol
(Merck, Germany), 33 g sodium chloride, 37.5 g baker's
yeast and 1086 ml water. The dough was proofed for 45 min
at 27 C under 80% moisture, divided into 350 g pieces and
put into cans before a second proo®ng for 82 min at 27 C
(80% moisture). After undergoing baking at 250 C for
40 min, the bread was allowed to cool at room temperature
for about 4 h. The crust was then separated, and the bread
was sliced, frozen at 7 20 C and stored for 3±9 months.
Four hundred grams of bread crumbs were thawed before
ingestion. The volume of fresh bread was measured with a
loaf volunteer (Tripette et Renaud, France), and the density
expressed in g=cm3.
In vivo study
Eight healthy volunteers (6 men and 2 women; mean age
25 y, range 21±28: body weight 55±70 Kg) took part in this
experiment. This study was approved by the Ethics Committee of the University Hospital of Nantes, and written
informed consent was obtained from the volunteers. The
day before the experiment, subjects were asked to consume
a meal composed of bread (60 g), pasta (100 g), ham (60 g),
natural yoghurt and ripe banana. The subjects each consumed all three test foods (WWB, HAWB, spaghetti) at
different times (approximately 4 d apart) after an overnight
fast. At 9.15 am on the day of the experiment, a catheter
was inserted into an antecubital vein to ensure blood
sampling. On the day of the experiment, the meal composed of a single test food was served at 9.30 am and
1
ingested continuously over a 20 min period. Water (Evian ,
Danone, France), adapted to the dry matter of food (respectively 430 ml, 411 ml and 377 ml with WWB, HAWB and
spaghetti), was drunk with the meals. Volunteers ate a meal
equivalent to 100 ±104 g of equivalent glucose for bread
and 108 g for spaghetti (1740±2094 KJ=meal) (Table 1).
The potentially available starch ingested was respectively
92±95 g for WWB and spaghetti, and 74 g for HAWB.
Blood samples were collected 30 min before the meal,
every 15 min for 1 h and 45 min after the meal, and then
every 30 min for a further 2 h. Blood samples intended for
blood glucose analysis were collected with heparin, and
those for blood insulin analysis with EDTA. Samples were
immediately centrifuged (3000 g, 5 min), and the plasma
was removed and frozen ( 7 80 C) until analysis. Plasma
glucose was assayed using a glucose oxidase method
(Beckman Auto-analyser II, Fullerton, CA, USA). Plasma
insulin was tested by radioimmunoassay (Institut Pasteur,
Paris, France). Incremental plasma glucose and insulin
responses above baseline were the means of 7 15 and
0 min values. The glycaemic (GI) and insulin (II) indexes
were based on the area under the glycaemic response curve,
except for the area beneath the baseline beyond 170 min
(Wolever & Jenkins, 1986). The glycaemic and insulinaemic indexes of HAWB and spaghetti were calculated from
the incremental blood glucose or insulin area in relation to
the corresponding area obtained after WWB used as reference food.
Table 1 Composition of bread and spaghetti (dry basis) and the corresponding test mealsa
White wheat bread (WWB)
Products
Fat
Protein
Total starch
Resistant starch
Available starch
Carbohydrate (glucose equivalent)
Energy (KJ/meal)
a
g/kg dry matter
Corresponding
test meal
g
14
131
718
10
708
1.8
17
93
1.3
92
Each value is the mean of two analytical data.
104
1916
High amylose wheat bread (HAWB)
Spaghetti
g/kg dry matter
Corresponding
test meal
g
g/kg dry matter
Corresponding
test meal
g
22
76
782
143
639
2.5
9
89
16.5
74
24
152
702
18
684
3.3
20
91
2.5
95
100
1750
108
2108
361
Bioavailability of starch in bread rich in amylose
C Hoebler et al
362
In vitro starch hydrolysis
The freeze-dried bread samples were coarsely ground using
a mortar and pestle, and freshly cooked spaghetti was cut
into pieces before enzymatic treatment. Samples (2 g) were
incubated with porcine pancreatic a-amylase (30 000 IU;
Merck, Darmstadt, Germany) and 0.55 mg (166 IU) of
amylase=g dry starch in phosphate buffer 0.025 M (CaCl2
0.25 1073 M, NaCl 1.75 1073 M) at 37 C with continuous magnetic stirring (Bornet et al, 1990). Samples
(0.9 ml) withdrawn every 5 min for 15 min and then at 30,
60, 120 and 180 min were added to 4.5 ml of a glacial acetic
acid (1.5%) and ethanol (95%) mixture. After approximately 16 h at 4 C the samples were centrifuged (3000 g,
10 min), and the carbohydrate content in supernatants was
analysed using the sulphuric orcinol automatic method
(Tollier & Robin, 1979). The rate of hydrolysis was
expressed as the percentage of starch hydrolysed into
soluble oligosaccharides.
Physico-chemical analysis
A portion of each bread and cooked pasta was freeze-dried
prior to analysis. Moisture content was determined by
drying at 103 C overnight. All yield compositions were
expressed on a dry matter basis. The bread and spaghetti
samples were analysed for protein by the Kjeldahl method
using 5.7 as the conversion factor. Starch was determined
in bread and cooked and strained spaghetti by the method
of Karkalas (1985) adapted as follows: samples (100 mg)
were solubilised in 1 N
NaOH, then neutralised and hydro1
lysed with Termamyl (Novo, Denmark). After hydrolysis
by
amyloglucosidase, released glucose was measured by the
enzymatic NADP-ATP=hexokinase=glucose-6-phosphate
dehydrogenase system (Boehringer-Mannheim, Germany,
cat. No. 127825). Resistant starch (RS) was determined
according to the method previously described by Berry
(1986) and modi®ed by Champ (1992). All chemical
analyses were done in duplicate on each sample. Available
starch were calculated by difference between total starch
and resistant strach.
X-ray diffraction patterns were obtained using an INEL
spectrometer by a previously described procedure (BuleÂon
et al, 1982). Samples were ®nely ground and hydrated to
20±30% water content before analysis. DSC measurements
were performed using a Setaram DSC III instrument.
Freeze-dried ground samples (approximately 10 mg) were
weighed accurately in Setaram steel pans. About 120 ml of
distilled water were added, and a sample pan and a
reference pan (®lled with 120 ml of water) were sealed
and allowed to equilibrate overnight. The DSC run was
performed from 30±180 C at a heating rate of 3 C=min.
Statistical analysis
Results are expressed as mean values s.d. Statistical
evaluations by two-way analysis of variance (ANOVA)
were performed for the postprandial response at each starch
hydrolysis time by pancreatic amylase, at each point of
blood sample collection, on total and net areas under curves
and on glycaemic and insulin indexes. Statistical evaluations were performed with the Stat-ViewTM SE ‡ Graphics
(1987±1988 Abacus Concepts, Inc.) programme.
Results
Postprandial blood glucose and insulin response
Maximum plasma glucose level was reached at 45 min for
all three test foods. The lowest peak value occurred after
consumption of HAWB, but the differences in response
were not statistically signi®cant (Figure 1). The decreases
in curves were markedly different (Figure 1). WWB
showed a slow decline, reaching basal level at 170 min.
Following an HAWB meal, mean glycaemic response
decreased rapidly after 45 min, reaching basal level at
90 min (the plasma glucose concentration obtained at
120 min was not signi®cantly different from the fasting
value, P < 0.05). Plasma glucose was signi®cantly lower
with HAWB than with WWB (P < 0.05) at 90 and 105 min.
For spaghetti, mean plasma glucose response up to 180 min
was not signi®cantly different from that obtained with the
two types of bread. However, plasma glucose was signi®cantly higher at 200 and 230 min (P < 0.05) after a spaghetti meal than after HAWB and WWB consumption, and
the basal plasma glucose level was not reached by the end
of the experiment.
The three starchy foods elicited their highest blood
insulin responses between 30 and 60 min after which the
curves decreased slowly. At the end of the experiment
(230 min), the plasma insulin level of the two types of
bread was not signi®cantly different from the basal level
(P < 0.05). However, after a spaghetti diet the mean insulin
curve had not reached baseline by the end of the experiment (230 min) (Figure 1). The postprandial insulin curves
after consumption of the three foods were not signi®cantly
different within the 30±75 min period. Mean insulin
responses to HAWB were signi®cantly lower than for
WWB within the 90±200 min period. Plasma insulin
curves for HAWB and spaghetti differed signi®cantly
only at the 170 min time-point (Figure 1).
The glycaemic index (GI) obtained for HAWB (60 18)
was signi®cantly lower than the reference GI (Table 2).
Spaghetti showed a GI (83 46) between that of the two
types of bread but not signi®cantly different. The insulinaemic indexes (Table 2) obtained after ingestion of
spaghetti (61 16) and HAWB (57 20 were signi®cantly
lower than the reference index.
Starch availability in vitro
The rate of starch hydrolysis was estimated by the amount
of soluble oligosaccharides produced during in vitro hydrolysis by amylase (Figure 2). WWB starch was hydrolysed
more rapidly and intensively (about 70% after 120 min of
enzymatic treatment) than in the other two foods. In the 15±
90 min period, the extent of starch digestion was not
signi®cantly different for the two types of bread
(P < 0.05), whereas the ®nal rate of starch hydrolysis was
signi®cantly lower for HAWB than WWB (P < 0.05).
Starch digestion of HAWB was lowest after 180 min of
in vitro enzymatic hydrolysis (45% of the starch hydrolysis
obtained with the other two foods, P < 0.05). The kinetics
of enzymatic hydrolysis for spaghetti differed from that of
the two types of bread. During the ®rst period (30±
120 min), the hydrolysis of spaghetti starch was signi®cantly slower than that of WWB, whereas the ®nal rate
(180 min) was not signi®cantly different for these two
foods.
Bioavailability of starch in bread rich in amylose
C Hoebler et al
363
Figure 1 Incremental plasma glucose (a) and insulin (b) variations after intake of starchy foods: -j- White wheat bread (WWB); -m- bread containing
high amylose maize starch (HAWB); and -- spaghetti. Each meal contained 100 g of starch and lasted 20 min for all volunteers (n ˆ 8). For each timepoint, means not sharing the same letter are signi®cantly different. (P < 0.05). Incremental plasma insulin variations after intake of starchy foods: -jWhite wheat bread (WWB); -m- bread containing high amylose maize starch (HAWB); -- and spaghetti. Each meal contained 90 g of starch and lasted
20 min for all volunteers (n ˆ 8). For each time-point, means not sharing the same letter are signi®cantly different (a, b, c: P < 0.05).
Table 2 Glycaemic (GI) and insulinaemic (II) indexes of the two types of
bread (wheat white bread and high amylose white wheat bread) and spaghetti
GI
II
WWB
Spaghetti
HAWB
100
100
83 46
61 16*
60 18*
57 20*
x s.d. (n ˆ 8). Means followed by an asterisk are signi®cantly different
from reference indexes (100 : WWB) (* P < 0.05%).
Physico-chemical characteristics
On a dry basis, the bread and pasta products consisted
mainly of starch (70±78%) and proteins (7±15%)
(Table 1). Resistant starch was low in the commercial
bread (WWB) and spaghetti (respectively 1.4 and 2.6% of
total carbohydrate) but amounted to 18.3% of the total
starch present in HAWB. The two types of bread had
density values respectively of 0.22 for WWB and 0.33 for
HAWB.
Bioavailability of starch in bread rich in amylose
C Hoebler et al
364
Figure 2 In vitro starch hydrolysis by porcine pancreatic a-amylase
(166 IU=g of starch). The percentage of hydrolysis is expressed in terms of
glucose equivalence: -j- White wheat bread (WWB); -m- bread containing high amylose maize starch (HAWB); -- and spaghetti; the results
were the means of 3 experiments. Each point represents the mean
(x s.e.m) of 3 experiments. For each time-point, means not sharing the
same letter are signi®cantly different (P < 0.05).
In X-ray studies, the diffraction patterns for WWB and
cooked pasta displayed a weak, mixed A-type (characteristic peaks at 2y ˆ 9.9 , 11.2 , 15 , 17 , 23.3 ) and a V-type
(characteristic peaks at 2y ˆ 7.5 , 13±14 and 18.9±20
according to the degree of hydration) (Figure 3). A-type
is characteristic of native wheat starch granules, and V type
of the complexation between amylose and lipids. HAWB,
which was rich in amylose, exhibited a B-type pattern
(2y ˆ 5.6 , 17 , 22 and 24 ) (Figure 3) characteristic of
native or retrograded amylose. The DSC thermogram for
the two types of bread showed an endotherm at 55 C,
probably due to native or retrograded amylopectin (Figure
4). Another endotherm observed with WWB and spaghetti
at 96 C (Figure 4) was due to weak lipid-amylose complexes. Finally, the three test foods exhibited a slight
endotherm at 145 C corresponding to retrograded amylose
Figure 3 X-ray diffraction patterns of white wheat bread (WWB),
spaghetti (SP), and bread containing high amylose maize starch (HAWB).
Figure 4 Differential scanning calorimetry runs for white wheat bread
(WWB), spaghetti (SP), and bread containing high amylose maize starch
(HAWB).
crystallites (Eliansson & Krog, 1985). The endotherm at
105 C, which was only observed in HAWB, corresponded
to the melting of residual native high amylose maize starch.
After storage of HAWB at 4 C for 2 d, an endotherm
appeared at 45 C due to the retrogradation of amylopectin
(data not shown: Eliansson, 1985).
Discussion
The purpose of this study was to produce resistant starch
not digestible by intestinal amylase in order to process
bread products with low metabolic response. High-amylose
maize starch was added in the dough since it retrogrades
easily during bread cooling and storage and requires a
relatively high temperature to become fully gelatinised
(Colonna, et al, 1992; Kulp & Ponte, 1981). Two standard
foods were used in the same experiment to rate this new
bread according to the glycaemic index (GI): a commercial
white wheat bread used as a food reference for calculating
metabolic indexes and a commercial spaghetti representing
low-GI food. To determine the starch-related factors
responsible for the low index of HAWB, we studied the
molecular structure of the starch in this bread and compared
the in vitro enzymatic hydrolysis of HAWB to that of
WWB and spaghetti.
The substitution of high amylose maize starch for 40%
of white wheat ¯our maize starch produced a palatable
bread (HAWB) which induced low metabolic responses.
No signi®cant difference was noted between the glycaemic
and insulinaemic indexes between HAWB and spaghetti,
despite 24% lower available starch in HAWB. As previously reported (Granfeldt et al, 1995) a moderate amount
of additional available starch (25% higher) would not
modify the metabolic responses. The metabolic indexes
obtained for HAWB (GI: 60 18: II: 58 20) were in the
same range usually obtained for lente carbohydrate.
Unusually, postprandial metabolic responses decreased
slowly to baseline and were affected by a large discrepancy, and standard error was probably due to the portion
size of the test foods. The metabolic responses were
determined after ingestion of meal composed only with
Bioavailability of starch in bread rich in amylose
C Hoebler et al
test food and corresponded to the normal caloric load
(about 1700±2000 kJ). The starch amount ingested during
the test meal corresponded to about twice the usual amount
of carbohydrate (about 90 g instead of 50 g of starch) used
for measurement of glycaemic and insulinaemic indexes
(Wolever et al, 1991). It was previously shown that metabolic response depends on the amount of carbohydrate
ingested (Gannon et al, 1987; Wolever et al, 1991). Thus,
the large food portion probably induced the slow decrease
of plasma glucose and insulin responses curves and also, an
extension in the time required for glucose and insulin
concentrations to return to fasting baseline.
WWB produced the highest metabolic response since its
starch content was easily accessible to amylase, as shown
in the in vitro hydrolysis step. DSC analysis and the X-ray
diffraction pattern of WWB revealed the presence of low
starch retrogradation (DSC analysis), low native amylopectin not gelatinised during bread-cooking (Variano-Martson
et al, 1980) and an amylose-lipid complex (Russel, 1983).
However, these different crystalline states produced a low
amount of resistant starch (1.4% of total carbohydrate) and
had no signi®cant in¯uence on postprandial glycaemic and
insulinaemic responses. As previously determined, pasta
had lower metabolic responses than WWB (Granfeldt &
BjoÈrck, 1991), although the amount of potentially available
starch was similar. The starch structure of pasta studied by
DSC and X-ray diffraction analysis was almost completely
gelatinised, showing physical characteristics similar to
those of WWB. Moreover, spaghetti was ingested just
after cooking, which reduced the risk of amylose retrogradation (endotherm at 145 C), observed in freeze-dried
pasta used for analysis (results not shown), and leading to
the formation of resistant starch (2.6% of total carbohydrate). Therefore the low metabolic indexes and the
low in vitro bioavailability of starch in spaghetti were not
due to the physical structure of starch but to the protein
network surrounding starch (Granfeldt & BjoÈrck, 1991:
Fardet et al, 1998). The physical barrier created by the
protein network limits the accessibility of starch to amylase
and delays in vitro starch hydrolysis (Colonna et al, 1990;
Granfeldt & BjoÈrck, 1991). Physiological factors such as
the persistence of large particles after chewing (Hoebler et
al, 1998) and slow gastric emptying of pasta (Mourot et al,
1988) take part also in the low metabolic responses of
pasta. Such large differences between in vitro starch availability (Fardet et al, 1998) and metabolic responses were
also found with a meal based on bread or pasta baked with
exactly the same ingredients (white wheat ¯our or durum
wheat) (Granfeldt et al, 1991; Granfeldt & BjoÈrck, 1991),
con®rming that starch structure was not involved in the
differing bioavailability of bread and pasta. In our study,
we chose to use, as the reference foods, commercial bread
and pasta with well-known high and low metabolic indexes
in order to evaluate the in vivo and in vitro availability of
bread rich in amylose (HAWB).
Bread rich in amylose (HAWB) attenuated postprandial
glucose and insulin responses as compared to WWB and
pasta, though the difference was not signi®cant with respect
to spaghetti. The lowering of hyperglycaemic effects previously found with high amylose food was not always
consistent and may have depended on the nature and
amount of amylose added (Granfeldt et al, 1995), the
technological process applied to food and therefore the
physico-chemical characteristics of starch (Behall et al,
1989; van Amelsvoort & Weststrate, 1992; Granfeldt et al,
1995; Heijnen et al, 1995). In our study, the substitution of
amylose-rich maize starch for 40% of white wheat ¯our led
to an increase in the resistant starch content of HAWB (14%
of DM) which involved a decrease of the extent of bioavailability of starch. The amylose=amylopectin ratio (0.39) was
relatively high, although similar to those used in other
studies (Weststrate & Van Amelsvoort, 1993). The
X-ray diffraction pattern of HAWB showed the presence
of crystallised amylose from native or retrograded amylose.
The gelatinisation temperature for maize starch rich in
amylose ranges between 100 and 130 C (Colonna et al,
1992). Accordingly, many studies investigating the effect of
amylose content on metabolic response have used a high
cooking temperature (from 175±240 C) (Weststrate & Van
Amelsvoort, 1993; Granfeldt et al, 1995; Heijnen et al,
1995). During the bread-making process, the temperature
in crumbs should not rise above 98±99 C inside bread
(Zanoni et al, 1995), which means that amylose is probably
incompletely gelatinised in HAWB. Consequently, the fraction of resistant starch (14.3%) would be likely composed of
native amylose and would not result from amylose retrogradation just after the baking and cooling of fresh bread
(Kulp & Ponte, 1981; Berry et al, 1988). The thermogram
obtained by DSC showed the presence of native amylopectin
in WWB and HAWB, which is consistent with the A-type
X-ray diffraction pattern of WWB. The presence of native
amylopectin con®rmed that 30% of the starch was not
gelatinised during cooking because of hydrothermic conditions unfavourable to starch gelatinisation. In the present
study, HAWB was kept at 7 20 C for 3±12 months, which
probably caused amylopectin retrogradation (Kulp & Ponte,
1981; Miles et al, 1985). However, it has been determined
that amylopectin in crystalline form does not modify starch
digestibility (Rabe & Sievert, 1992). In the same way, lipids
in the two types of bread produced amylose-lipid complexes
that were revealed by endotherms (96 C, 107 C) and the
V-type X-ray diffraction pattern (Szczodrak & Pomeranz,
1991), which are apparently not involved in the low bioavailability of starch (Rabe & Sievert, 1992).
From all these results, HAWB starch was composed
with two fractions differing in their bioavailability and with
different physiological consequences: a resistant starch
fraction mainly composed of native amylose which produced low metabolic responses and decreased its in vitro
digestibility and a remaining fraction, easily hydrolysed.
This available starch fraction of HAWB produced a drop in
blood glucose concentration below the fasting blood level
in late phase and may have had an impact on subjective
feelings of satiety, hunger or appetite. However, the relationship of satiety to postprandial glycaemic response is not
clear (Barkeling et al, 1995), and the satiating effect seems
to be a complex process involving multiple factors (Holt et
al, 1994). Escaping digestion in the small intestine, resistant starch fraction present in HAWB become a substrate
for colonic fermentation with possible implication in cancer
(Hylla et al, 1998). Resistant starch was suspected to
produce a relative high production of butyrate (Martin et
al, 1998), considered as a bene®cial substrate against
colonic diseases.
Conclusions
These results indicate that it is possible to produce bread
with a low glycaemic index. This product, though more
dense than white wheat bread, has similar texture, is
365
Bioavailability of starch in bread rich in amylose
C Hoebler et al
366
palatable and could be introduced into the diet. Increased
amylose content did not lead to amylose retrogradation,
although physical analysis (X-ray diffraction and DSC)
revealed the presence of crystalline native amylose not
gelatinised during bread-processing. The ®ne characterisation of starch in food or in digesta recovered at the
digestion site should elucidate the potential changes in
starch likely to produce low glycaemic responses. Therefore, bread enriched in high amylose maize starch could
prove to be an important source of resistant starch for use in
diets intended to maintain good health or implied for cancer
prevention. The potential bene®ts of this special bread
intake can be controlled by adapting the amount of highamylose maize starch in the dough.
Acknowledgements ÐWe thank Mr A BuleÂon and Mr B Pontoire for their
technical assistance in X-ray diffraction and DSC analysis and their help in
the interpretation of X-ray and DSC results.
References
Barkeling B, Granfeldt Y, BjoÈrck I & Rossner S (1995): Effects of
carbohydrates in the form of pasta and bread on food intake and satiety
in man. Nutr. Res. 15, 467±476.
Behall KM, Scho®eld DJ, Yuhanaiak I & Canary J (1989): Diets containing high amylose vs amylopectin starch: effects on metabolic variables
in human subjects. Am. J. Clin. Nutr. 49, 337±344.
Berry CS (1986): Resistant starch; Formation and measurement of starch
that survives exhaustive digestion with amylolytic enzymes during the
determination of dietary ®bre. J. Cereal Sci. 4, 301±334.
Berry CS, Anson KI, Miles MJ, Morris VJ & Russell PL (1988): Physical
and chemical characterisation of resistant starch from wheat. J. Cereal
Sci. 8, 203±206.
Bornet F, Bizais Y, Bruley Des Varannes S, Pouliquen B, Delort-Laval J &
Galmiche J-P (1990): Alpha-amylase (E. C. 3.2.1.1.) susceptibility
rather than viscosity or gastric emptying rate controls plasma responses
to starch in healthy humans. Br. J. Nutr. 63, 207±220.
Brennan CS, Blake DE, Ellis PR, Scho®eld JD (1996): Effects of guar
galactomannan on wheat bread microstructure and on the in vitro and in
vivo digestibility of starch in bread. J. Cereal Sci. 24, 151±160.
BuleÂon A, Bizot H, Delage MM & Multon JL (1982): Evolution of
crystallinity and speci®c gracity of potato versus water ad-desorption.
Starch=StaÈrke 34, 361±366.
Champ M (1992): Determination of resistant starch in foods and food
products: interlaboratory study. Eur. J. Clin. Nutr. 46, (Suppl 2), S51±S62.
Colonna P, Barry J-L, Cloarec D, Bornet F, Gouilloud S & Galmiche J-P
(1990): Enzymic susceptibility of starch from pasta. J. Cereal Sci. 11,
59±70.
Colonna P, Leloup V & BuleÂon A (1992): Limiting factors of starch
hydrolysis. Eur. J. Clin. Nutr. 46, Suppl. 2, S17±S32.
Eliansson AC & Krog N (1985): Physical properties of amylose-monoglyceride complexes. J. Cereal Sci. 3, 239±248.
Eliansson AC (1985): Retrogradation of starch as measured by D.S.C. In
RD Hill & L Munck (eds) New Approaches to Research on Cereal
Carbohydrates, 93±98. Amsterdam (NLD): Elsevier SPBV.
Ellis PR, Dawoud FM & Morris ER (1991): Blood glucose, plasma insulin
and sensory responses to guar-containing wheat breadsÐ effects of
molecular weight and particle size of guar gum. Br. J. Nutr. 66, 363±379.
Englyst HN, Kingman SM & Cummings JH (1992): Classi®cation and
measurement of nutritionally important starch fractions. Eur. J. Clin.
Nutr. 46, Suppl 2, S33±S50.
Fardet A, Hoebler C, Baldwin PM, Bouchet B, Gallant DJ & Barry J-L
(1998): Involvement of the protein network in the in vitro degradation
of starch from spaghetti and lasagne: a microcopic and enzymatic study.
J. Cereal Sci. 27, 133±145.
Fosterpowell K & Miller JB (1995): International tables of glycemic index.
Am. J. Clin. Nutr. 62, S871±S890.
Gannon MC, Franck Q & Nutall MD (1987): Factors affecting interpretation
of postprandial glucose and insulin areas. Diabetes Care 10, 759±763.
Granfeldt Y & BjoÈrck I (1991): Glycemic response to starch in pasta: a study
of mechanisms of limited enzyme availability. J. Cereal Sci. 14, 47±61.
Granfeldt Y, BjoÈrck I & Hagander B (1991): On the importance of
processing conditions, product thickness and egg addition for the
glycaemic and hormonal responses to pasta: a comparison with bread
from `pasta ingredients'. Eur. J. Clin. Nutr. 45, 489±499.
Granfeldt Y, Drews A & BjoÈrck I (1995): Arepas made from high amylose
corn ¯our produce favorably low glucose and insulin responses in
healthy humans. J. Nut. 125, 459±465.
Heijnen MLA, Van Amelsvoort JMM & Westsrate JA (1995): Interaction
between physical structure and amylose:amylopectin ratio of foods on
postprandial glucose and insulin responses in healthy subjects. Eur. J.
Clin. Nutr. 49, 446±457.
Hoebler C, Karinthi A, Devaux M-F, Guillon F, Gallant DJG, Bouchet B
& Barry J-L (1998): Physical and chemical transformations of cereal
food during oral digestion in humans. Br. J. Nutr. 80, 429±436.
Holm J & BjoÈrck I (1992): Bioavailability of starch in various wheat-based
bread products: evaluation of metabolic responses in healthy subjects
and rate and extent of in vitro starch digestion. Am. J. Clin. Nutr. 55,
420±429.
Holt S, Miller JCB, Saveny C & Hansky J (1994): Glycemic index, satiety,
and the Cholecystokinin response. Am. J. Clin. Nutr. 59, S787.
Holt SHA & Miller JB (1994): Particle size, satiety and the glycaemic
response. Eur. J. Clin. Nutr. 48, 496±502.
Hylla S, Gostner A, Dusel G, Anger H, Bartram HP, Christl SU, Kasper H
& Scheppach W (1998): Effects of resistant starch on the colon in
healthy volunteers: possible implications for cancer prevention. Am. J.
Clin. Nutr. 67, 136±142.
Jenkins DJA, Wolever TMS, Kalumsky J, Guidici S, Gordiano C, Patten
R, Wong GS, Bird JN, Hall M, Buckley G, Csima A & Little JA (1987):
Low glycemic index diet in hyperlipidemia: use of traditional starchy
foods. Am. J. Clin. Nutr. 46, 516±520.
Jenkins DJA, Wolever TMS & Jenkins AL (1988): Starchy foods and
glycemic index. Diabetes care 11, 149±159.
Jenkins DJA, Jenkins AL, Wolever TMS, Vuksan V, Rao AV, Thompson
LU & Josse RG (1994): Low glycemic indexÐ lente carbohydrates and
physiological effects of altered food frequency. Am. J. Clin. Nutr. 59,
S706±S709.
Karkalas J (1985): An improved enzymatic method for the determination
of native and modi®ed starch. J. Agric. Food Agric. 36, 1017±1027.
Kulp K & Ponte JG (1981): Staling of white pan bread: fundamental
causes. CRC Crit. Rev. Food Sci. Nutr. 9, 1±48.
Liljeberg H, Granfeldt Y & BjoÈrck I (1992): Metabolic responses to starch
in bread containing intact kernels versus milled ¯our. Eur. J. Clin. Nutr.
46, 561±575.
Martin JM, Dumon HJW & Champ MMJ (1998): Production of shortchain fatty acids form resistant strach in a pig model. J. Sci. Food Agric.
77, 71±80.
Miles MJ, Morris VJ, Orford PD & Ring SG (1985): The roles of amylose
and amylopectin in the gelation and retrogradation of starch. Carbohydr. Res. 135, 271±281.
Mourot J, Thouvenot P, Couet C, Antoine JM, Krobicka A & Debry G
(1988): Relationship between the rate of gastric emptying and glucose
and insulin responses to starchy foods in young healthy adults. Am. J.
Clin. Nutr. 48, 1035±1040.
Rabe E & Sievert D (1992): Effects of baking, pasta production, and
extrusion cooking on formation of resistant starch. Eur. J. Clin. Nutr.
46, Suppl 2, S105±S107.
Russel PL (1983): A kinetic study of bread staling by differential scanning
calorimetry. The effect of painting loaves with ethanol. Stark=StaÈrke.
35, 277±281.
Szczodrak J & Pomeranz Y (1991): Starch and enzyme-resistant starch
from high-amylose barley. Cereal Chem. 68, 589±596.
Tollier M-T & Robin J-P (1979): Adaptation de la meÂthode aÁ l'orcinol
sulfurique au dosage automatique des glucides neutres totaux: conditions d'application aux extraits d'origine veÂgeÂtale. Ann. Technol. Agric.
28, 1±15.
van Amelsvoort JMM & Weststrate JA (1992): Amylose-amylopectin ratio
in a meal affects postprandial variables in male volunteers. Am. J. Clin.
Nutr. 55, 712±718.
Variano-Martson E, Huang G & Ponte J (1980): Comparison of methods to
determine starch gelatinization in bakery foods. Cereal Chem. 57, 242±
248.
Weststrate JA & van Amelsvoort JMM (1993): Effects of the amylose
content of breakfast and lunch on postprandial variables in male
volunteers. Am. J. Clin. Nutr. 58, 180±186.
Wolever TMS & Jenkins DJA (1986): The use of the glycemic index in
predicting the blood glucose response to mixed meals. Am. J. Clin. Nutr.
43, 137±172.
Wolever TMS, Jenkins DJA, Jenkins AL & Josse RG (1991): The
glycemic index: methodology and clinical implications. Am. J. Clin.
Nutr. 54, 846±854.
Zanoni B, Peri C & Bruno D (1995): Modelling of strach gelatinization
kinetics of bread crumb during baking. Lebensm Wiss u Technol. 28,
314±318.