Human and Clinical Nutrition

Iron Absorption from Bread in Humans: Inhibiting

Effects of Cereal Fiber, Phytate and Inositol
Phosphates with Different Numbers of
Phosphate Groups1
MATS BRUNE, LENA ROSSANDER-HLTN,LEIF HALLBERG/
ANN GLEERPAND ANN-SOFIE SANDBERG*
Department of Medicine II and Clinical Nutrition, University of Gteborg,
Sahlgrenska Hospital, and *Department of Food Science,
Chalmers University of Technology, S-413 45 Gteborg,Sweden

INDEXING KEY WORDS:

iron absorption phytate

Inositol phosphates fiber
humans

Supported by grants from the Swedish Medical Research

Council Project, B88 - 19X-04721-13B, the Swedish Council for
Forestry and Agriculture Research, 864/86, 31:2 and 599-89L-135:2
and the Swedish Agency for Research Co-operation with De
veloping Countries 9.49/SAREC 85/46:2.
^To whom correspondence and reprint requests should be ad
dressed.

The primary dietary factors determining the

amount of iron absorbed in humans are the contents
of heme and nonheme iron and the balance between
factors influencing the bioavailability of nonheme
iron in particular (1). Some dietary factors enhance
0022-3166/92

$3.00 1992 American Institute of Nutrition. Received 19 November 1990. Accepted 22 August 1991.

442

Downloaded from jn.nutrition.org by guest on August 26, 2015

iron absorption in humans, e.g., ascorbic acid and

meat or fish, whereas others inhibit the absorption of
nonheme iron, e.g., iron-binding phenolic compounds
(2), calcium (3) and, notably, phytate (4).
Bread is a staple food in many countries and is an
important source of both iron and the inhibiting phy
tate. The chemical composition of flour depends on
the proportion of the cereal grain removed by the
milling process. Low extraction white flour mainly
originates from the endosperm, whereas flours with
higher extraction also contain increasing amounts of
bran. White flour has a low native iron content and is
therefore enriched with iron in several countries. The
higher the extraction of the flour, the higher the
content of iron and phytate that originates from the
bran.
Iron absorption from bread baked from high ex
traction flour is often poor in spite of its higher iron
content. The lower absorption has been ascribed to its
content of bran. The inhibitory effect of bran on iron
absorption in humans has been clearly demonstrated
using both the chemical balance method (5) and the
extrinsic radioiron technique (6). However, the
negative effect of bran on iron absorption is not seen
in all species. Results from a careful comparative
study (7) on humans and rats explain why the inhib
itory effect of bran is not observed in a rat model (8,
9).

ABSTRACT Iron absorption was measured from five

kinds of bread made from various types of flour and
fermented in different ways in order to obtain a wide
variation in the content of fiber, phytate (inositol hexaphosphate) and its degradation products, inorganic
phosphate and inositol phosphates with fewer numbers
of phosphate groups (inositol pentaphosphate through
monophosphate). Each experiment had 9-10 subjects
and, in each subject, iron absorption was measured from
control rolls made from low extraction wheat flour and
one kind of test roll using two different radioiron tracers:
55pe and ^Fe. 7he inhibition of iron absorption was
closely related to the content of phytate-phosphorous as
determined using the AOAC method, and to the sum of
the tri- through hexaphosphate groups as determined
using the HPLC method. As an example, prolonged fer
mentation of whole-rye bread reduced total inositol
phosphates to the same amount as in the control rolls
and increased fractional iron absorption to the same
high level, in spite of a fiber content five times as great.
The results strongly suggest that the inhibitory effect of
bran on iron absorption is due to its content of phytate
and other inositol phosphates present after fermenta
tion, rather than to its content of fiber or other constitu
ents. Thus, effective fermentation will increase the bioavailability of iron in whole-meal bread. J. Nutr. 122:
442-449, 1992.

BREAD AND IRON ABSORPTION

adjusting the pH to 4.5 with lactic acid.

In vitro studies suggest that inositol phosphates
with fewer numbers of phosphate groups (the trithrough pentaphosphate groups) may induce a smaller
inhibition of iron absorption than inositol hexaphosphate (16), and that complete reduction of
phytate by soaking, germination and fermentation
leads to a marked increase in the bioavailability of
iron (14, 18). It is not known, however, to what extent
different inositol phosphates inhibit the absorption of
iron in humans and hence it is not known to what
extent variations in bread-making techniques may
influence iron absorption in humans.

The main purpose of the present study was to

determine the amount of different inositol phosphates
in bread prepared in different ways and to relate these
amounts to the bioavailability of iron in humans. The
use of two readily measurable radioiron isotopes
makes it possible to compare iron absorption from
differently prepared breads by relating the absorption
in each subject to a standard white wheat bread with
almost no remaining inositol phosphates. Another
purpose of the study was to compare the amounts of
phytate in bread and flour as determined by the
anion-exchange resin method (AOAC method) (19)
and an HPLC method (16, 17) determining individual
inositol phosphates in the triphosphate to hexaphosphate range. Moreover, we examined the possi
bility of reducing the inhibiting effect of inositol
phosphates on the absorption of iron from bread by
more effective fermentation.

MATERIALS AND METHODS

Experimental design. Each experiment had 9-10
subjects and iron absorption was measured in each
subject using two different radioiron tracers, from one
control meal and one kind of test meal. Thus, each
subject served as his or her own control. The control
meal (two rolls), served in each experiment, consisted
of white wheat bread, butter and water. The test
meals in the five experiments also consisted of two
rolls, butter and water. The test rolls, however, were
made from different kinds of flours, or flour mixtures,
and fermented in different ways in order to obtain
different contents of inositol phosphates. The control
rolls were prepared in a standardized way from low
extraction (55%) wheat flour to produce rolls with a
very low content of inositol phosphate.
Test rolls and control rolls (a and b) were served on
alternate mornings after an overnight fast on four
consecutive days in the order abba or baab. The rolls
were labeled with two different radioiron isotopes,
55Fe and 59Fe. A blood sample was drawn 2 wk after
serving the last roll to determine the content of 55Fe
and 59Fe.The total retention of 59Fewas measured by
whole-body counting at the same time. The total
retention of 55Fewas calculated from the ratio of 55Fe
to 59Fein red cells. An oral reference dose (see below)
was then given to the fasting subject and a second
dose was given on the following morning. The ab
sorption of the reference doses were then measured by
whole-body counting 2 wk later (see the expression of
results below). For a review of the extrinsic tag
method, see reference (1).
Subjects. Forty-nine subjects, 20 men and 29
women, participated in the five experiments. All sub
jects were healthy volunteers, aged 19-47 y, and each
group of 9-10 subjects included both men and
women. Some of the subjects in each group were

Downloaded from jn.nutrition.org by guest on August 26, 2015

It was suggested early that the inhibitory effect of

bran on iron absorption in humans was due to the
phytate content of bran (5). This was not confirmed in
a study comparing the effect on iron absorption of
native bran and dephytinized bran (10). The
dephvtinization, however, was not complete in that
study. Further studies on the effect of removing
phytate from bran (11) and studies showing that even
small amounts of phytate have a marked inhibitory
effect on iron absorption (4) may well explain why
seemingly contradictory results were obtained. There
is thus much evidence that phytate is one of the key
nutritional factors influencing the bioavailability of
dietary nonheme iron.
The high content of fiber in bran has also been
suggested as an explanation of its inhibitory effect on
the absorption of iron. This effect, however, has never
been established.
The content of phytate-P in flour can vary from a
few milligrams to several hundred milligrams per 100
g of flour (12). In unprocessed flours and bran the
inositol phosphates occur mainly in the hexaphosphate form (phytate) and to a small extent in the
pentaphosphate form (13, 14). During bread fer
mentation these inositol phosphates are partly or
completely degraded to inorganic phosphates and to
inositol phosphates with fewer numbers of phosphate
groups, mono- through pentaphosphate (14, 15). The
contents of the different inositol phosphates re
maining in bread after fermentation and baking
depend on 1} the choice of cereal (e.g., wheat, rye,
oats); 2) the extent of extraction during milling of the
grain; 3) the freshness of the flour (influencing the
phytase activity); and 4} the techniques employed
during fermentation and baking (e.g., leavening time,
temperature, amounts and properties of the yeast
used).
It currently is possible to determine the inositol
phosphates remaining in bread after fermentation and
baking by employing a recently described HPLC
method (16, 17). Using this method, Larsson and
Sandberg (14) demonstrated that it is possible to
reduce the phytate content of bread containing rye
bran by 97% by means of rye sourdough fermentation
or, in rye bread, by scalding with 70Cwater and

443

BRUNE ET AL.

444

TABLE 1
ronabsorption front different breads1

valuesExperimentalMeals2Control

Absorption
mean

ReferenceSubjects310[1M;9F(2)]10[7M(2);3F]9[1M;8F(3)]10[8M(2);2F(1)]85%
iron

of
individual
absorption
testrolls:
controlrollsn

rollsWhole
5.223.6

0.130.59

ns..
OA + n
WControlrye -f
0.140.68

5.627.2

rollsWhole
2.423.2

0.060.58

nuW.
70
\J.\JOn-a
17
i+ n
WControlwheat +
0.060.57

2.422.8

rollsWheat
0.070.23

2.89.2

(itU.
-u
n
U.UAn
07
I
extr.Control
85%
0.050.64

2.025.6

rollsWhole
0.100.23

4.09.2

fisU.OZi
29 X+ n
U.UOO-o
WheatControl
rye +
extraction10[3M(2);7F(2)]content4mg4.0(3.7)3.7(2.0)4.0(3.7)3.7(2.3)4.0(3.7)3.4(2.0)4.0(3.7)3.4(1.0)4.0(3.7)3.0Meal%18.818.219.8
0.070.55

2.822.0

rollsWheat
2.43.2

0.060.08

mi^ m
-L n\i.\>~
bran + WNonheme
0.02ratios5%24.4 0.8Mean
U.Uv)n
7H

corresponding to a 40% reference dose absorption.

regular blood donors (Table 1), which provided a

reasonable range of intersubject variation in iron ab
sorption. The volunteers were given oral and written
information about the aims and procedures of the
study. The project was approved by the Ethical Com
mittee of the Medical Faculty of the University of
Gteborg.
Meals, bread-making and labeling of meals. Each
control or test meal consisted of two rolls made from
80 g of unfortified flour or flour mixture. In all experi
ments the rolls were served with 20 g of butter and
150 mL of water. The calculated energy content of the
meals was 1.8 MJ. Iron as ferrous sulfate was added to
the dough of all rolls except the bran rolls, to obtain a
similar iron content in all rolls (Table 1).
Control rolls were made from 40 g of unfortified
white wheat flour (55% extraction), yeast, sugar, table
salt and water. The amounts of yeast, sugar and salt
per kg of flour were 65 g, 35 g and 10 g, respectively.
The dough was fermented for l h at 23'C. It was then
kneaded again and weighed amounts transferred to
small aluminum forms, which were left standing for
20 min for further fermentation. The bread was then
baked at 250'C for 15 min.
In Experiment 1, test rolls were made from whole
rye flour and low extraction (55%) wheat flour.
Whole-meal rye (500 g), yeast (25 g) and water (750
mL; 40"C) were mixed and left at 23'C for 48 h to
ferment. The sourdough formed was then mixed with
yeast (40 g; dissolved in a small amount of water),
white wheat flour (500 g), sugar (35 g) and table salt

(10 g). The dough was kneaded and fermented for 1 h.

The rolls were then made in exactly the same way as
the control rolls.
In Experiment 2 test rolls were made from whole
wheat flour and low extraction (55%) wheat flour.
Whole wheat flour (500 g), yeast (25 g) and water (750
mL; 40C)were mixed and fermented for 48 h at
23 C.The dough was then treated exactly as the rye
rolls in Experiment 1.
In Experiment 3 test rolls were made from wheat
flour (85% extraction). The amounts and proportions
of flour, yeast, sugar, table salt and water were the
same as in the control rolls. The dough was fer
mented for l h at 23'C. The rolls were then made
exactly as the control rolls.
In Experiment 4 test rolls were made from whole
rye flour and 85% extraction wheat flour. Whole rye
flour (500 g), yeast (25 g) and water (750 mL: 40C)
were mixed and fermented at 23 Cfor 48 h. The
sourdough formed was then mixed with yeast (40 g;
dissolved in a small amount of water), and 85% ex
traction wheat flour (500 g), sugar (35 g) and table salt
(10 g). The dough was kneaded and then fermented for
an additional 3 h at 23C.The rolls were then made
exactly as the control rolls.
In Experiment 5 test rolls were made from wheat
bran and low extraction (55%) wheat flour. Wheat
flour (675 g) was mixed with wheat bran (325 g).
Water, yeast, salt and sugar were added in the same
proportions as in the other experiments. The dough
was fermented for 1 h at 23 C.The rolls were then

Downloaded from jn.nutrition.org by guest on August 26, 2015

Values are means SEM,n - 9-10.

2W - white wheat flour.
^M - male and F - female subjects. Number in parentheses designates blood donors.
*The amount of fortification iron in parentheses.
"The mean value of the individual absorption ratios (meal:reference dose) were multiplied by 40 to obtain the percentage absorption of iron

445

BREAD AND IRON ABSORPTION

ference in iron bioavailability between the two meals.

In the present experiments, a standardized control
meal was served on alternate days with the test meal.
In the presentation of the results, the mean of the
individual absorption ratios is considered to be the
most accurate basis for comparisons between the dif
ferent experiments.
The ratio of the absorption of nonheme iron from a
meal and from a reference dose is an expression of the
bioavailability of nonheme iron in the meal. In each
experiment the mean value and the SEMof the indi
vidual absorption ratios (meal:reference dose) were
multiplied by 40 to obtain the percentage absorption
of iron that corresponds to a 40% reference dose
absorption. Absorption values adjusted to a 40% ab
sorption from reference doses were chosen because
they roughly correspond to the absorption expected in
subjects who are borderline iron-deficient (25). This
adjusted absorption value is the basis for comparisons
between the amounts of iron absorbed from different
meals, served to different groups.
Statistical methods. All statistical analyses were
made using a Statview n computer program (Abacus
Concepts, Berkeley, CA). For statistical comparisons,
the mean and SEMof the individual absorption ratios
(test meal:control meal) in each experiment were
used. The possible statistical significance of the dif
ference between the mean absorption and 1 was in
each experiment examined by a t test.

RESULTS
Iron absorption expernents.In Experiment 1, iron
absorption from the rolls baked from whole rye flour
and low extraction wheat flour was not statistically
different from the absorption from the control rolls.
The mean of the individual iron absorption ratios
(test rolls:control rolls) was 0.94, (not significant)
(Table 1). In Experiment 2, iron absorption from the
whole wheat rolls was lower than from the control
rolls. The difference was statistically significant (ab
sorption ratio 0.79; P < 0.05). In Experiments 3-5 the
differences between iron absorption from the test
rolls and the control rolls were all statistically signifi
cant. The mean individual iron absorption ratios (test
rollsrcontrol rolls) were 0.39 (P < 0.01), 0.32 (P < 0.01)
and 0.13 (P < 0.01), respectively.
Analysis of phytate and inositol-phosphates.
Values for the content of phytate (AOAC method) and
inositol phosphates (sum of the tri-, tetra-, penta- and
hexaphosphate groups; HPLC method) in the different
flours and rolls are given in Table 2. Using the HPLC
method only inositol hexaphosphate and small
amounts of inositol pentaphosphate could be detected
in the flours and bran. In the rolls, after fermentation
and baking, there were different amounts of various
inositol phosphates (the tri-, tetra-, penta- and hexa
phosphate groups). The greatest reduction of inositol

Downloaded from jn.nutrition.org by guest on August 26, 2015

made exactly as the control rolls. The bran used was a

Swedish commercial wheat-bran product containing
20% whole grain and a dietary fiber content of 50%.
In all studies, each meal was labeled with 46.2 kBq
of 59Fe or 55Fe. High specific activity radioiron
(Amersham International, Amersham, Buckingham
shire, U.K.) was added to the dough fluid as ferric
chloride in 0.01 mol/L hydrochloric acid. In Experi
ments 1-4 the test rolls were labeled with 55Fe and
the control rolls with 59Fe, whereas in Experiment 5
the control rolls were labeled with 55Fe and the test
rolls with 59Fe.
Oral reference doses. A solution of 10 mL of 0.01
mol/L hydrochloric acid containing 3 mg of Fe as
FeS04 and 30 mg of ascorbic acid labeled with 59Fe
was used as a reference in all studies. The 10-mL vials
containing the Fe solution were rinsed twice with
water and this was also consumed. Each subject
received two reference doses on two consecutive
mornings after an overnight fast. No food or drink
was allowed for 3 h after the reference dose. Each
subject received a total of 55.5 kBq of 59Fe.
Chemical measurements. Aliquots of rolls were
freeze-dried and ground to a powder in a porcelain
mortar. Weighed amounts of this powder were ana
lyzed for total iron (20) and phytic acid phosphorus.
The latter analyses were made in duplicate using the
AOAC method (19). In rolls with a low content of
phytate, 6 mL of the original HC1 extract was used
instead of 1 mL (original method), and the amount of
NaOH-EDTA was increased proportionally. This
modification markedly increased the lower limit of
detection of phytate in the extract. The precision of
the method and the recovery of added sodium phytate
were unchanged. Samples of flour and freeze-dried
rolls were also analyzed by an HPLC method to
identify and quantify inositol tri-, tetra-, penta- and
hexaphosphates (16, 17). Several complete analyses
were made from most samples (Table 2).
The AOAC and the HPLC methods were carefully
compared and proved to give the same results for pure
sodium phytate standards.
Total fiber content in the rolls was analyzed in
duplicate (21, 22).
iron absorption measurements. Relative ab
sorption of 55Fe and 59Fe was calculated from
analyses of blood samples. Absolute absorption of the
two tracers was calculated from whole-body counting
of 59Fe and the relative absorption of the two tracers.
Analyses of 55Fe and 59Fe in blood were made by a
modification of the method described by Eakins and
Brown (23), using a liquid scintillation spectrometer
(Tri-Carb, Packard Instruments, San Antonio, TX). All
procedures and methods of calculation have been
described previously (24).
Expressing results of iron absorption measure
ments. The mean of the individual absorption ratios,
test meal:control meal, is an expression of the dif

BRUNE ET AL.

446

TABLE

Amount

of phytate-P

(AOAC method) or inositol phosphates

(HPLC method) in rolls prepared

fermented in different wavs1-'

from different flours and

rollsLow

SeriesType
+low
rye
offlourFermentationtimeAOAC-methodFlour,
extractionwheat
flour2

+low wheat
extractionwheat
flour2

extraction1

+85%
rye
extractionwheat
flour2

+low bran
extractionwheat
flour1

extractionwheat
(range)1

d+20
+ 1h
d+20
+ 1 h
h + 20
min955.789.3483.4 min1018.292.8492.9min871572266.3

d+20
+ 3 h
min15734.1122.94137.1
min3792271524334.4
h + 20
min14.2-12.35.5-6.55.8-8.087.0h + 20

(2.77)6.0
(0.24)89.430.21

(2.9)5.5
(0.15)142.662.3

flowRolls,
mg P/80 g
P/iollDifference,
mg
ragHPLC-methodFlour,

samplesIP6,
mgIPS,
mgn>4,
mgIP3,
mgS
mgDifference
IP,
(flour-rolls)E
mgFiber
IP,
g/rollAbsorption31Whole
content,

(2.4)4.6
(0.11)97.540.17

(2.2)3.4
(0.12)69.733.95

(7.29)20.8
(0.56)355.22156.6

(0.06)0.04
(0.04)0.14
(0.22)0.65
(0.36)2.6
(1.55)8.5
(0.01)0.23
(0.01)0.26
(0.19)14.9
(0.19)14.0
(0.05)0.87
(0.04)1.64
(0.83)10.9
(0.08)0.82
(0.02)3.17
(0.32)24.4
(0.19)1.3088.16.20.94(0.21)2.2195.35.20.79
(0.40)8.6461.03.80.39
(1.2)203.5151.517.80.13
(0.55)30.7111.99.90.32

0.052Whole

0.033Wheat85%0.024Whole

'Values are means (so).

Abbreviations used: IP, inositol phosphate; IP3, inositol triphosphate;

0.035Wheat

IP4, inositol tetraphosphate;

0.02Control

IPS, inositol pentaphosphate;

IP6,

inositol hexaphosphate.
3Values are means SEM,n - 9-10.

phosphate content in the rolls compared with the

flour (99%) was in Experiment 1.
The agreement between the results obtained with
the two methods was good. However, the values for
phytate-P as determined with the AOAC method
were consistently slightly higher than the sum of the
inositol phosphates determined using the HPLC
method. This was noted for both flours and rolls. The
control rolls contained on average 6 mg of phytate-P
per portion (80 g of flour), using the AOAC method,
in contrast to the finding of at most 1.6 mg of re
maining inositol phosphates using the HPLC method.
The difference (6 - 1.6 = 4.4 mg) observed was almost
exactly the same in the control rolls and the white
wheat flour used. Moreover, the differences were of
the same order of magnitude (range 3.4-6.4 mg
phosphorus/portion) as in the test rolls used in Ex
periments 1-4. In Experiment 5 (bran rolls), however,
the difference was greater (24 mg phosphorus/
portion).
Iron absorption and phytate content. Figure 1
shows the absorption of iron from the five different
test rolls in relation to the iron absorption from the

control rolls, graphed against the content of re

maining inositol phosphates (sum of tri-, tetra-, pentaand hexaphosphates; HPLC method) and phytate
(AOAC method) in the rolls. The results from both
analytical methods are expressed as milligrams of
phosphorous per portion. The solid Unes in the graph
describe results from a previous study (4) illustrating
the effect of adding known amounts of sodium
phytate to the same kind of control rolls as used in
the present study. It should be noted that in Figure 1
(but not in Table 2) the AOAC values for phytate
content in the experimental rolls have, as a correc
tion, been reduced by 4.4 mg phosphorus.
Analysis of fiber content. The fiber (nonstarch
polysaccharides including lignin) contents of the rolls
are shown in Table 2. There was a wide range in fiber
content from 1.1 g per portion in the control rolls to
17.8 g per portion in the bran rolls.

DISCUSSION
In the present study, iron absorption from five
kinds of rye or wheat bread, fermented in different

Downloaded from jn.nutrition.org by guest on August 26, 2015

samplesIP6,
gIPS,mg/80
gZ mg/80
gRolls,
IP, mg/80

BREAD AND IRON ABSORPTION

1001
While wheat bread wilh added sodium phyiale
o Remaining phytale in difieren) bread (AOAC method)
A Remaining IP6-IP3 in dilferenl bread (HPLC method)
S 80 H
c
o

o.
o
w

60

40

20

10

Total amount

100

of Inositol

Phosphates

1000

( mg P )

ways, was compared with iron absorption from white

wheat bread made from flour of very low extraction
(55%). The aim of the different baking procedures and
choice of flours was to obtain a wide variation in the
content of inositol-phosphates and fiber in the rolls
(Table 2). Several unexpected observations were made
in the experiments.
The first observation was the relationship between
the content of inositol phosphates in the different test
rolls and the mean absorption ratio in the five experi
ments. In spite of the different fiber content and the
different kinds of flours used, there was a close rela
tionship between the total content of those inositol
phosphates that could be determined using the HPLC
method (inositol tri-, tetra-, penta- and hexaphosphates) and the inhibition of iron absorption as
graphed in Figure 1. The form of this graph with its
clear breakpoint has been discussed in a previous
paper (4). It is evident from the figure that there was
also a close relationship between the content of
phytate-P as determined using the AOAC method and
the inhibition of iron absorption.
It is especially notable that iron absorption from
the whole rye rolls (Experiment 1) was the same as
from the white wheat control rolls (absorption ratio
0.94 0.05). Evidently, it was possible, by sourdough
fermentation, to degrade the phytate almost com
pletely in a bread baked from whole rye flour. The
initial phytate content was great, but after fer
mentation the amounts of remaining inositol phos
phates, and the effect on iron absorption, were not
different from those of the control rolls.

The iron absorption from the whole wheat bread

(Experiment 2), fermented in the same way as the
whole rye bread, was also high (absorption ratio 0.79),
but slightly lower than from the whole rye bread. As
shown in Table 2, the content of inositol phosphate,
as analyzed with both analytical methods used,
differed between the two kind of breads,- this fact may
explain the observed difference in iron absorption.
The cause of the difference in phytate content be
tween the rye and the wheat bread may be the usually
higher phytase activity in rye flour (26).
The inhibition of iron absorption by adding sodium
phytate to wheat rolls observed in a previous study is
graphed in Figure 1 (4). As seen in the same graph, the
extent of inhibition was about the same per milligram
of phosphorus from the mixtures of various inositol
phosphates in the presently used rolls. It should be
emphasized that inositol triphosphate and inositol
tetraphosphate constituted a considerable fraction of
the total inositol phosphates in all breads (19-84%).
The relationship between the inhibition of iron ab
sorption and the content of inositol phosphate would
have been weak, if for example, only inositol hexaphosphate and possibly inositol pentaphosphate act
as inhibitors. An illustration of this is the bread baked
from whole rye flour and 85% extraction wheat flour
(Experiment 4). The iron absorption from this bread
was reduced by -70%, in spite of the fact that there
was only 4.5 mg of inositol hexaphosphate and in
ositol pentaphosphate remaining after fermentation
and baking. Our findings thus strongly suggest that all
inositol phosphates, at least the inositol hexaphosphates to inositol triphosphates analyzed, in re
lation to their phosphorus content, inhibit iron ab
sorption to about the same extent.
The fiber content was quite high (6.2 g) in the
whole rye rolls. The control rolls, however, had a very
low fiber content (1.1 g). In spite of this difference, the
fractional, as well as the absolute iron absorption, was
almost identical. This finding indicates that the rye
fibers per se had no inhibitory effect on iron absorp
tion. In fact, in the present study there seems to be no
relationship between the fiber content in rye or wheat
rolls and the degree of inhibition of iron absorption
(Table 2). This result is in accordance with earlier
studies in which completely dephytinized bran (11) or
cellulose (27) have been found to have no effect on
iron absorption.
Complex chemical processes take place during
sourdough fermentation and some of those may affect
iron absorption. Still, the most likely explanation for
the increased bioavailability of iron observed is the
degradation of phytate, because there is a close in
verse relationship between iron absorption and re
maining inositol phosphates. It has previously been
shown that sourdough fermentation influenced
neither the content of dietary fiber nor the distri
bution of soluble and insoluble fiber components (28).

Downloaded from jn.nutrition.org by guest on August 26, 2015

FIGURE 1 Iron absorption in relation to content of inositol phosphate in different types of bread. The relative
iron absorption is the ratio between test rolls and control
rolls x 100. The solid Unes represent the relationship ob
served in another study (4) in which sodium phytate was
added to the same kind of control rolls. The AOAC values
plotted have been reduced by 4.4 mg.

447

448

BRUNE ET AL.

containing fraction exists in the control rolls that is

determined as phytate-P but is actually not inositol
phosphate. As was pointed out in Results, this
fraction seems to be of the same order of magnitude
in all test rolls except the bran rolls, hi Figure 1, the
AOAC values for phytate-P content of all rolls have
been arbitrarily reduced by 4.4 mg to correct for this
source of error.
One conclusion of the present comparisons be
tween the two methods is that it is only at very low
inositol phosphate levels that the overestimation of
the phytate values, using the AOAC method, is of any
practical significance. Knowledge of this source of
error, however, is important considering the signif
icant inhibitory effect on iron absorption even of very
small amounts of phytate (4). Moreover, it should be
emphasized that very low phytate-P contents cannot
be detected using the original AOAC method, but
only by the presently used modification of the
method (see Materials and Methods).
The present findings have important nutritional
implications. One is the demonstration that not only
phytate inositol hexaphosphate, but also its inositol
phosphate-degradation products, seem to be potent
inhibitors of iron absorption. An even more practical
implication is the finding that effective fermentation
of whole-meal bread to very low inositol phosphate
levels will markedly improve the bioavailability of
iron and that the effect of fiber per se on iron ab
sorption can be disregarded. Consequently, the
amounts of iron absorbed from bread baked from
whole wheat flour, with its high content of iron,
would be greater than from white bread, provided the
fermentation procedures are optimized.
It is well known that phytate interferes with the
absorption of other minerals, e.g., zinc and calcium. It
is possible that many of the findings and implications
of the present study may also be valid for these ele
ments.

LITERATURE CITED
1. Hallberg, L. (1981) Bioavailability of dietary iron in man.
Annu. Rev. Nutr. 1: 123-147.
2. Brune, M., Rossander, L. &. Hallberg, L. (1989) Iron absorption
and phenolic compounds. Importance of different phenolic
structures. Eur. J. Clin. Nutr. 43: 547-558.
3. Hallberg, L., Brune, M., Erlandsson, M., Sandberg, A-S. &.
Rossander, L. (1991) Calcium: effect of different amounts on
non-heme iron and heme-iron absorption in humans. Am. J.
Clin. Nutr. 53: 112-119.
4. Hallberg, L., Brune, M. &. Rossander, L. (1989) Iron absorption
in man: ascorbic acid and dose-dependent inhibition by phy
tate. Am. J. Clin. Nutr. 49: 140-144.
5. Widdowson, E. M. &. McCance, R. A. (1942) Iron exchanges of
adults on white and brown bread. Lancet 1: 588-591.
6. Bjrn-Rasmussen, E. (1974) Iron absorption from wheat bread.
Influence of various amounts of bran. Nutr. Metab. 16:
101-110.

Downloaded from jn.nutrition.org by guest on August 26, 2015

Lactic acid is the main organic acid formed during

sourdough fermentation. No effect on iron absorption
was seen when very large amounts of this acid were
added to different meals (29, 30). Thus, the reduction
of the inhibition of iron absorption observed in Ex
periments 1, 2 and 4 is not likely due to changes in
the fiber content or the formation of lactic acid during
prolonged fermentation.
From Table 2 it can be calculated that during the
degradation of phytate in the test rolls, rather large
amounts of inorganic phosphates were formed (about
60-150 mg phosphorus/portion). These inorganic
phosphates had no detectable inhibitory effect on the
absorption of iron, even in this sensitive experimental
model.
Taken together, the findings clearly indicate that
the inhibition of iron absorption by wheat and rye
bran could be ascribed to its content of phytate, or
after baking to its degradation products (inositol phos
phates), and not to fiber components or other constit
uents of bran.
The AOAC method for phytate determination is
based on the separation of inorganic phosphates and
phytate by means of an ion-exchange resin column.
As can be seen in Table 2, the agreement between the
values for phytate-P (AOAC method) and the sum of
the tri-, tetra-, penta- and hexaphosphates (HPLC
method) were good. This indicates that the AOAC
method determines not only phytate but also, at least,
inositol tri-, tetra- and pentaphosphates.
However, slightly higher values for phytate-P were
consistently found using the AOAC method com
pared with the sum of the tri-, tetra-, penta- and
hexaphosphates obtained by the HPLC method (Table
2). This difference is probably due to several factors.
In bread, it is possible that some of the phosphorus
determined using the AOAC method might actually
be inositol mono- and diphosphates, which cannot be
determined by the HPLC method. This contribution
to the AOAC values is probably negligible in flour
because inositol phosphates in flour occur almost
exclusively as inositol pentaphosphates and inositol
hexaphosphates (13, 14, 31).
A source of overestimation of the true content of
inositol phosphates using the AOAC method is that,
besides inositol phosphates, certain other phosphates
will also be measured. This is the case with
nucleotide-phosphorous
compounds
(31) and
adenosine-triphosphate
(unpublished observations).
The consistent finding of small amounts of phytate-P
(AOAC method) in the control rolls (Table 2) is
probably mainly due to such a systematic error. This
conclusion is supported by the finding of negligible
amounts of inositol phosphate in the control rolls (up
to 1.6 mg) using the HPLC method, and by the fact
that even after prolonged and repeated fermentation
the content of phytate-P in the control rolls remained
unchanged (values not shown). Thus a phosphate-

BREAD AND IRON ABSORPTION

tology. (Cook, J. D., d.),

pp. 116-133. Churchill, London, U.K.
21. Asp, N-G., Johansson, C-G., Siljestrom, M. &. Hallmer, H.
(1983). Rapid enzymatic assay of insoluble and soluble dietary
fibre. J. Agrie. Food Chem. 31: 476-482.
22. Prosky, L., Asp, N-G., Schweizer, T. F., deVries, J. W. &. Furda,
I. (1988) Determination of total dietary fiber in foods and food
products: collaborative study. J. Assoc. Off. Anal. Chem. 68:
677-679.
23. Eakins, J. D. &. Brown, D. A. (1966) An improved method for
the simultaneous
determination
of iron-55 and iron-59 in
blood by liquid scintilltion counting. Int. J. Appi. Radit.Isot.
17: 391-397.
24. Bjm-Rasmussen, E., Hallberg, L., Magnusson, B., Rossander,
L., Svanberg, B. & Arvidsson, B. (1976) Measurements of iron
absorption from composite meals. Am. J. Clin. ute. 29:
772-778.
25. Magnusson, B., Bjrn-Rasmussen, E., Hallberg, L. & Rossander,
L. (1981) Iron absorption in relation to iron status. Model
proposed to express results of food iron absorption measure
ments. Scand. J. Haematol. 27: 201-208.
26. Bartnik, M. & Szafranska, I. (1987) Changes in phytate content
and phytase activity during the germination of some cereals. J.
Cereal Sci. 5: 23-28.
27. Gillooly, M., Bothwell, T. H., Challn, R. W., Torrance, J. D.,
Bezwoda, W. R., MacPhail, A. P., Derman, D. P., Novelli, L.,
Morral, P. & Mayet, F. (1984) Factors affecting the absorption
of iron from cereals. Br. J. Nute. 51: 37-46.
28. Siljestrom, M., Bjrck,I., Lonner, A. C., Nyman, M. & Asp, NG. (1988) Effects on polysaccharides during baking and storing
of breadin vitto and in vivo studies. Cereal Chem. 65: 1-8.
29. Hallberg, L. &. Rossander, L. (1984) Improvement of iron
nutrition in developing countries. Comparison of adding meat,
soy protein, ascorbic acid, citric acid and ferrous sulphate on
iron absorption from a simple Latin-American type of meal.
Am. J. Clin. Nutr. 39: 577-583.
30. Baynes, R. D., Macfarlane, B. J., Bothwell, T. H., Siegenberg,
D., Bezwoda, W. R., Schmidt, U., Lamparelli, R. D., Mayet, F.
& MacPhail, A. P. (1990) The promotive effect of soy sauce on
iron absorption in human subjects. Eur. J. Clin. ute. 44:
419-424.
31. Phillippy, B. Q., Johnston, M. R., Tao, S-H. &. Fox, M.R.S.
(1988) Inositol phosphates in processed foods. J. Food Sci. 53:
496^499.

Downloaded from jn.nutrition.org by guest on August 26, 2015

7. Reddy, M. B. & Cook, J. D. (1991) Assessment of dietary

determinants of non-heme iron absorption in humans and rats.
Am. J. Clin. ute. 54: 723-728.
8. Morris, E. R. &. Ellis, R. (1980) Bioavailability to rats of iron
and zinc in wheat bran: response to low-phytate bran and
effect of the phytate:zinc molar ratio. J. Nutr. 110: 2000-2010.
9. Morris, E. R. & Ellis, R. (1976) Isolation of monoferric phytate
from wheat bran and its biological value as an iron source to
the rat. J. Nutr. 106: 753-760.
10. Simpson, K. M., Morris, E. R. &. Cook, J. D. (1981) The
inhibitory effect of bran on iron absorption in man. Am. J.
Clin. Nutr. 34: 1469-1478.
11. Hallberg, L., Rossander, L. & Sknberg,A-B. (1987) Phytates
and the inhibitory effect of bran on iron absorption in man.
Am. J. Clin. ute. 45: 988-996.
12. Harland, B. F. & Oberleas, D. (1987) Phytate in foods. In: World
Review of Nutrition and Dietetics, vol. 52, pp. 235-259.
Karger, Basel, Switzerland.
13. Sandberg, A-S. &. Svanberg, U. (1991) Phytate hydrolysis by
phytase in cereals, effects on in vitro estimation of iron availa
bility. J. Food Sci. 56: 1330-1333.
14. Larsson, M. & Sandberg, A-S. (1991) Phytate reduction in bread
containing oat flour, oat bran or rye bran. J. Cereal Sci. 14:
141-149.
15. Nayini, N. R. & Markakis, P. (1983) Effect of fermentation
time on the inositol phosphates of bread. J. Food Sci. 48:
262-263.
16. Sandberg, A-S., Carlsson, N-G. &. Svanberg, U. (1989) Effects of
inositol tri-, tetra-, penta-, and hexaphosphates on in vitro
estimation of iron availability. J. Food Sci. 54: 159-162.
17. Sandberg, A-S. & Adherirme, R. (1986) HPLC method for deter
mination of inositol tri-, tetra-, penta-, and hexaphosphates in
foods and intestinal contents. J. Food Sci. 51: 547-550.
18. Svanberg, U. &. Sandberg, A-S. (1988) Improved iron availa
bility in weaning foods using germination and fermentation.
In: International Workshop on Household Level Technologies
for Improving Young Child Feeding in Eastern and Southern
Africa. (Alnwick, D., Moses, S. & Schmidt, O. G., eds.), pp.
366-373. EDRC, Ottawa, Canada.
19. Harland, B. F. & Oberleas, D. (1986) Anion-exchange method
for determination of phytate in foods: collaborative study. J.
Assoc. Off. Anal. Chem. 69: 667-670.
20. Hallberg, L. (1980) Food iron absorption. In: Methods in Hema-

449