Content uploaded by Andrew Patrick Macphail
Author content
All content in this area was uploaded by Andrew Patrick Macphail on Nov 20, 2014
Content may be subject to copyright.
Br.
J.
Nutr.
(1981),
45,
215
215
Factors affecting the absorption of iron from Fe(II1)EDTA
BY A.
P.
MAcPHAIL,
T.
H. BOTHWELL,*
J.
D. TORRANCE,
D. P.
DERMAN, W. R. BEZWODA
AND
R. W. CHARLTON
Joint UniversitylSouth African MRC Iron and Red Cell Metabolism Unit, Department of
Medicine, University of the Witwatersrand, Johannesburg, South Africa
AND
FATIMA MAYET
Department of Medicine, University of Natal, Durban, South Africa
(Received
20
June
1980
-
Accepted
24
July
1980)
1.
The modification of iron absorption from Fe(1II)EDTA by agents known to promote or inhibit absorption
was examined in
101
volunteer multiparous Indian women. Fe absorption from Fe(I1I)EDTA was compared with
absorption of intrinsic food Fe in a further twenty-eight subjects. Finally the urinary excretion of radio-Fe after
oral administration of 58Fe(III)EDTA was studied in twenty-four subjects and evidence of intraluminal exchange
of Fe was examined.
2.
Fe absorption from maize porridge fortified with Fe(1II)EDTA was more than twice that from porridge
fortified with FeSO, .7H,O.
3.
Although bran decreased Fe absorption from FeSO, .7H,O approximately
1
I-fold, it had no significant effect
on Fe absorption from Fe(I1I)EDTA. Nevertheless tea, which is a more potent inhibitor of Fe absorption,
decreased absorption from Fe(l1I)EDTA 7-fold.
4.
Fe absorption from Fe(lI1)EDTA given in water was only increased
40%
by addition
of
3
mol ascorbic
acid/mol Fe but by 7-fold when the relative proportions were increased to
6:
1.
This enhancing effect was blunted
when the Fe(ll1)EDTA was given with maize porridge.
In
these circumstances, an ascorbate:iron value of
3:
1
(which doubles absorption from FeSO, .7H,O) produced no significant increase in
Fe
absorption, while
a
value
of
6:
1
produced only
a
2.5-fold increase.
5.
Fe absorption from Fe(II1)EDTA was not altered by addition of maize porridge unless ascorbic acid was
present.
6.
Less
than
1
%
of j9Fe administered as 59Fe(III)EDTA was excreted in the urine and there
was
an inverse
relationship between Fe absorption and the amounts excreted
(r
0.58,
P
<
0.05).
7.
Isotope
exchange between "Fe(I1I)EDTA and 5BFeS0,. 7H,O was demonstrated by finding
a
similar relative
value for the two isotopes in urine and erythrocytes when the two labelled compounds were given together orally.
This finding was confirmed by in vitro studies, which showed enhanced 69Fe solubilization from "FeSO,. 7H,O
in maize porridge when unlabelled Fe(II1)EDTA was added.
8. Although Fe absorption from Fe(II1)EDTA was marginally higher it appeared
to
form
a
common
pool with
intrinsic food iron in most studies. It is postulated that the mechanism whereby Fe(1II)EDTA forms
a
common
pool with intrinsic
food
Fe differs from that
occurring
with simple
Fe
salts. When Fe is present in the chelated
form
it
remains in solution and is relatively well absorbed because it is protected from inhibitory ligands. Simple
Fe salts, however, are not similarly protected and are absorbed
as
poorly as the intrinsic food
Fe.
9.
It
is
concluded that Fe(I1I)EDTA may be a useful compound for food fortification of cereals because the
Fe is well absorbed and utilized
for
haemoglobin synthesis. The substances in cereals which inhibit absorption
of simple Fe salts do not appear to inhibit absorption of Fe from Fe(I1I)EDTA.
The high incidence
of
iron deficiency in populations subsisting on cereal-based diets is largely
due to the poor absorption
of
Fe from these foodstuffs. For example, the results
of
previous
studies indicated that the geometric mean absorption of Fe from rice and maize in Fe
deficient subjects was less than
5%
(Sayers
et
al.
1973; Sayers, Lynch, Charlton
&
Bothwell,
1974; Sayers, Lynch, Charlton, Bothwell
et al.
1974; Derman
et al.
1977). These results
underline the need for Fe-fortification programmes in populations subsisting on mainly
*
For reprints.
0007-1 145/81/3558-2407
$01.00
0
1981 The Nutrition Society
216
A.
P.
MACPHAIL
AND
OTHERS
cereal-based diets. Unfortunately, such programmes have not been successful in the past,
since the Fe salts used for fortification are as poorly absorbed as the intrinsic Fe present
in various cereals (Layrisse
et
al.
1973). This can be ascribed to the fact that the ligands
inhibiting the absorption of the food Fe, have the same effect on the added Fe. One way
of overcoming this problem has been to add ligands, such as ascorbic acid, which enhance
Fe absorption (Sayers
et al.
1973; Sayers, Lynch, Charlton
&
Bothwell, 1974; Sayers, Lynch,
Charlton, Bothwell
et
al.
1974; Disler, Lynch, Charlton, Bothwell
et
al.
1975). While
ascorbic acid has been shown to
be
effective, it has been difficult to find a suitable vehicle
to which to add it, since it is a highly-reactive compound, which is easily oxidized and which
sometimes causes unacceptable colour changes in food (Sayers, Lynch, Charlton, Bothwell
et
al.
1974). The results of recent work suggest that an alternative approach may be
promising. It has been shown that the Fe present in the Fe(II1)EDTA compound is better
absorbed in the presence of a meal than is the Fe in other salts, possibly due to the fact
that
it
is less susceptible to inhibitory ligands which form insoluble complexes with Fe
(Layrisse
&
Martinez-Torres, 1977; Martinez-Torres
et al.
1979).
In the present study further information was obtained on the absorption
of
Fe from
Fe(II1)EDTA and on the effects of various foods and of inhibitory and promoting ligands.
In addition, information was obtained on the intraluminal mixing of Fe and on the excretion
of Fe(1II)EDTA in the urine.
EXPERIMENTAL
Subjects
The subjects were 153 multiparous Indian housewives aged between 21 and 71 years (mean
36 years) living in a municipal housing complex at Chatsworth, near Durban. They belong
to a low socio-economic group in which Fe deficiency is common (Mayet
et
al.
1972).
Preparation and administriztion
of
the
meals
Sufficient maize meal
to
provide
50
g dry maize meal/subject was mixed with a small amount
of water to make a paste. This was added to four times its weight of boiling water and cooked
at 90° for 20-25 min to make a porridge. When intrinsically-labelled maize (Hussain
et
al.
1965; Layrisse
et
al.
1969) was used, sufficient to provide 3 pCi ssFe/subject was blended
with dry carrier maize to provide
50
g dry maize/subject. In some experiments the meal was
fortified with 3 or
5
mg Fe labelled with 3 pCi/subject by mixing the appropriate radio-Fe
compound into the paste. In other experiments the porridge was cooked without added Fe
and individual servings were subsequently fortified by adding
3
ml portions of solutions
containing 3 or
5
mg of Fe either as Fe(II1)EDTA in water or as FeSO,. 7H,O in
0.01 M-hydrochloric acid. In each instance the portion contained 3 pCi of the appropriate
radio-Fe labelled compound. The final mass of porridge eaten by each subject was 250 g.
Tea was made by adding
45
g leaves (Pot
0’
Gold;
O.K.
Bazaars Ltd, Johannesburg) to
1800 ml boiling water to provide 200 ml/subject. The tea was sweetened with cane sugar
fortified with the appropriate radio-Fe compound. The traditional Indian dhal (Sayers,
Lynch, Charlton
&
Bothwell, 1974) and the fortified sugar (Disler, Lynch, Charlton,
Bothwell
et
al.
1975) were prepared as described previously.
In each study the meal was consumed after an overnight fast and no food or drink was
allowed for
4
h after the test meal had been eaten. The same procedure was followed the
next day, but the meal was labelled with the other Fe isotope. After a period of 2 weeks,
blood samples were taken after an overnight fast for the determination of the concentrations
of 59Fe and 55Fe, haemoglobin, serum Fe, unsaturated Fe-binding capacity and serum
ferritin. Each subject then drank a ‘reference Fe salt’ consisting of
50
ml of a solution
Iron
absorption
from
Fe(III)EDTA
217
containing
30
mg ascorbic acid and
3
mg Fe as 5gFeS0,. 7H,O
(3
pCi 59Fe).
No
food or
drink was allowed for the following
4
h period. Samples of blood were again obtained a
further 2 weeks later and the 5gFe concentrations determined. The percentage absorption
of the Fe from the solution was calculated by difference and provided an index of each
individual's absorbing capacity.
Isotopic
and chemical methoh
Labelled Fe(1II)EDTA was prepared by mixing equimolar solutions of disodium EDTA
and carrier ferric chloride with a tracer amount of radio-labelled FeCl,. The pH of the
solution was then adjusted to pH
5
with sodium bicarbonate. The Fe(II1)EDTA complex
formed was precipitated by ethanol, the supernatant fraction discarded and the precipitate
redissolved in water. The complex was then precipitated again with ethanol and washed
three times with ethanol (Sawyer
&
McKinnie,
1960).
Duplicate blood samples (10 ml) and appropriate standards were prepared for differential
radioactivity determination by the method of Katz
et al.
(1964).
The quantities of 55Fe and
59Fe in the processed samples were determined using a liquid-scintillation system (Insta-Gel
;
Tri-Carb AM Spectrometer model
No.
3375;
Packard Instrument Co., Downers Grove,
Illinois,
USA).
The 59Fe activity in the
4
ml blood samples collected immediately before
the 'reference Fe salt' was administered, and
2
weeks later, was assessed against suitable
standards using a liquid-scintillation spectrometer (Auto-Gamma Tri-Carb Spectrometer
model
No.
3001
;
Packard Instrument Co.). All values for absorption
(%)
were calculated
on the assumption that 100% of the absorbed radioactivity was present in the haemoglobin
of circulating erythrocytes, and that the blood volume of each subject was
65
ml/kg.
Although the red cell utilization of radioiron varies according to iron status, it is normally
above
80%
and in an iron deficient group such as the one investigated in the present study,
most values would be expected to
be
close to 100%. Furthermore, any errors introduced
by
the assumption would not invalidate the findings, since the design of the absorption
experiments was such that each subject served as her own control, with comparisons being
made on the basis of paired data and not between different individuals.
When large volumes of urine were prepared for the determination of differential
radioactivity using the same method as for blood, a black compound was obtained which
was unsuitable for mixing in insta-gel. To overcome this problem, half the volume of urine
passed was evaporated to
1&15
ml by gentle boiling and prepared according to the method
used for blood samples. The resulting black compound was dissolved in
1
M-HCl and all
the Fe was converted to the ferric state by the addition of potassium permanganate solution.
The ferric-Fe was complexed with potassium thiocyanate and immediately extracted with
diethyl ether. The water phase was found to be free of radioactivity after four to five diethyl
ether extractions. The diethyl ether phase was evaporated to dryness and then redigested
and prepared for radioactivity determination as described previously.
Serum Fe concentrations were measured by the International Committee for Standard-
ization in Haematology
(1978)
method and the unsaturated Fe-binding capacity was
determined by the method of Herbert
et al.
(1967).
The Fe content of digested samples of
food was estimated by a modification (Bothwell
et al.
1979)
of the method of Lorber (1927).
The serum ferritin concentrations were measured by radioimmunoassay using the method
of Deppe
et al.
(1978).
The in vitro experiments were done'using a modification of a method described previously
(Bezwoda
et
al.
1978).
In place of the breadused in the original method, a soft maize
porridge was prepared using
70
g dry maize meal/l water.
This
was fortified with
3
mg Fe
in the form of the appropriate radio-Fe compound. Equal weighed portions of the porridge
were then incubated in duplicate for
30
min at room temperature in
5
ml HCl solutions
218
A.
P.
MACPHAIL
AND
OTHERS
covering the pH range
0.5-6.0.
Portions of the supernatant fraction were counted after
centrifugation at 3000
g
and the percentage Fe in solution was calculated.
Ethical considerations
Written consent was obtained from all subjects after the nature of the investigation had
been explained to them by an Indian social worker. Before starting the study approval was
obtained from the Committee for Research
on
Human Subjects of the Faculty of Medicine,
University of the Witwatersrand, Johannesburg. With regard to radiation exposure, it was
calculated that if the entire test dose of radioisotope was retained, the total radiation
averaged over 13 weeks would be approximately 20% and
0.2%
of the permissible
whole-body burden for continuous exposure to 69Fe and 55Fe respectively (International
Commission for Radiation Protection,
1960).
Statistical methodr
Serum ferritin and Fe absorption values in individual experiments showed considerable
variation. All values were therefore logarithmically transformed in order to correct for
positive skew. Since all results were expressed as geometric means and standard deviation
ranges, standard parametric statistical methods could be used.
RESULTS
Comparison of the absorption of Fe from maize-meal porridge fortiJied with FeSO,
.
7H20
or with Fe(IIl)EDTA
In order to establish whether the absorption of Fe from FeSO,
.
7H,O and Fe(II1)EDTA
differed, twelve subjects were given FeSO,
.
7H20-fortified maize porridge on one morning
and a similar Fe(I1T)EDTA-fortified porridge the following morning. Table
1
shows that the
geometric mean Fe absorption
PA)
from the Fe(II1)EDTA fortified meal (7.2) was
significantly greater than from the FeSO, .7H,O fortified meal (3.5)
(t
5.5,
P
<
0.001).
The effect of inhibitors
on
Fe absorption from Fe(IIl)EDTA
The 2-fold increase in Fe absorption found in the first study suggested that Fe in the
Fe(I1I)EDTA complex is protected from the inhibitors of Fe absorption present in maize
porridge. It was therefore decided to determine whether known inhibitors of Fe absorption,
such as bran and tea, had any effect on Fe absorption from Fe(II1)EDTA.
In order to assess the effect of bran on Fe absorption,
9
subjects drank 100 ml water
sweetened with sugar fortified with FeSO,. 7H,O on one morning while on the second
morning
10
g bran was mixed with the Fe-fortified water (Table
2).
Bran reduced the
geometric mean absorption 11-fold, from
(%)
16.5
to
1.5
(t
16.5,
P
<
0.001). In a second
study, which was carried out in ten subjects, FeSO, .7H,O was replaced by Fe(II1)EDTA.
There
was
no significant difference between the mean geometric absorption from Fe(II1)-
EDTA without bran (10.3) and with bran
(8.4)
(t
1.4,
P
>
0.1).
As
tea
is
known to be a potent inhibitor of Fe absorption (Disler, Lynch, Charlton,
Torrance
et al.
1975), a similar study was performed in eight subjects using black tea instead
of
bran. The geometric mean Fe absorption from Fe(II1)EDTA was reduced approximately
7-fold from 19.2 to 2.8, by the addition of tea
(t
10.35,
P
<
0-001).
The effect of increasing doses of ascorbic acid on the absorption of Fe from Fe(II2)EDTA
As the Fe in Fe(II1)EDTA seemed to be protected to some extent from inhibitory ligands,
the effect
of
a
known promoter
of
Fe absorption, namely ascorbic acid, was studied. Two
series
of
three experiments were done to find out whether increasing doses of ascorbic acid
improved Fe absorption from Fe(II1)EDTA. In the first set of experiments the Fe(II1)EDTA
Table
1.
Comparison of absorption
of
iron from FeSO,
.
IH,O
and Fe(IIl)EDTA fort@ed maize porridge
(3
mg be,
L3U
g porridge per
subject) given to female subjects on two successive mornings
(Mean values and standard deviations)
Fe absorption
(%)*
Transferrin Serum*
Haemoglobin saturation ferritin
(g/O
(%I
olg/l)
Referencet FeSO, .7H,O Fe(1II)EDT'A
salt fortified porridge fortified porridge
No.
of
Geometric Geometric Geometric Geometric
subjects Mean
SD
Mean
SD
mean
&
1
SD
mean
SD
mean k1
SD
mean
&
1
SD
12 127 26 26.2 11.9 15.3 (5.3-44.2) 35.3 (20.9-59.6) 3.5 (2.549) 7.2 (3.9-13.2)
2
3
*
Geometric means and
SD
ranges
used
because
values were positively skewed.
t
3 mg Fe as ferrous ascorbate given in the fasting state.
5-
$
3
Table
2.
The eflect
of
inhibitors
on
the absorption of iron
(5
mg) from FeSO,
,
IH,O
and
Fe(IIl)EDTA fortified sugar
(20
g) given in
100
ml water to female subjects
on
two successive mornings
s
(Mean values and standard deviations)
%
tb
n
Fe absorption
PA)*
3
Serum"
6
2
Haemoglobin Transfemn ferritin Reference? Without With
(g/U
saturation 01l3/1) salt inhibitor inhibitor
No.
of
(30-
GO-
Geo- Geo-
Fe sub- metric metric metric metric
Inhibitor compound jects Mean
SD
Mean
SD
mean
SD
mean
fl
SD
mean fl
SD
mean fl
SD
10
g
Bran FeSO, .7H,O 9 137 9 25.2
8.0
33.0 (13443.6) 22-5 (7.8-65'0)
16.5
(74-36.5)
1.5
(06-3.8)
10
g
Bran Fe(1II)EDTA
10
120 22 22.9
14.1
10.7 (2.6-43.8) 43.3 (19.1-97.7) 10.3 (56-19.0)
8.4
(3.2-22.0)
Teal Fe(II1)EDTA
8
129
14
28.0 18.8 21.1 (633-70.7) 49.2 (27.7-87.5) 19.2 (8945.9) 2.8 (1.342)
*
Geometric means and
SD
ranges
used because values were positively skewed.
1
45
g
tea leaves added
to
1800 ml boiling water to give 200 ml/subject.
3
mg
Fe
as ferrous ascorbate given in fasting state.
E
W
220
A.
P.
MACPHAIL
AND
OTHERS
was administered in water alone, while in the second set it was administered in maize
porridge. In each study the meal was given on one of the days with ascorbic acid and on
the other, without ascorbic acid. At the completion of each absorption study, a further one
was done in which the absorption of
a
reference dose of 3 mg Fe as ferrous ascorbate was
measured. In this
way
it was possible not only to express the actual results obtained but
also to standardize individual results
to
a 40% reference absorption, which has been used
as a
means of comparing the findings in individuals or groups with differing avidities for
Fe (Hallberg
er al.
1978).
When the unstandardized values were analysed (Table 3), it was clear that
25
mg ascorbic
acid had no effect on the absorption of Fe from Fe(1II)EDTA when administered in water
or porridge
(t
0-24,
P
>
0.1
;
t
1-03,
P
>
0-1 respectively). With the
50
mg dose there was
a modest, but significant increase
(t
2-84,
P
<
0.05)
in Fe absorption when Fe(II1)EDTA
was
given in water but none when it was given in porridge
(t
0.38,
P
>
0.1). A dose of 100 mg
ascorbic acid caused a more than 6-fold increase
(t
6.23,
P
-=
0.01) in Fe absorption when
the Fe(II1)EDTA was given in water and a 2-fold increase
(t
5.71,
P
<
0.01) when it was
given in porridge.
When the standardized absorptions of Fe from Fe(1II)EDTA in water were compared
with those obtained from Fe(1II)EDTA-fortified porridge (Table 3), both in the absence
of ascorbic acid, it was apparent that the addition
of
250 g maize-meal porridge had little
or no effect on the absorption from Fe(II1)EDTA
(F
1.3,
P
>
0.1). However, when 100 mg
ascorbic acid was present, the standardized absorption
(%)
was significantly reduced from
a geometric mean of
44
to 15.6 by the presence of maize porridge
(t
1-93,
P
<
0.05).
Comparison between the absorption
of
intrinsically-labelled cereals
and
59Fe(III)EDTA
Three different meals were used to compare the absorption of intrinsically labelled food
Fe and Fe(II1)EDTA. Two experiments were done using maize poridge as in the previous
studies. In the first of these experiments (Table 4) sugar fortified with 5gFe(III)EDTA was
sprinkled on the cooked maize porridge intrinsically labelled with 55Fe and the porridge
was eaten immediately. The geometric mean absorption from the extrinsic Fe(1II)EDTA
CL)
(12.9) was more than double that from the intrinsic label (4.9)
(t
3.74,
P
<
0.005).
This
suggested that the Fe in the form of Fe(II1)EDTA was protected from inhibitory ligands
in the maize, while the intrinsic Fe was not. However, since the two labels were not cooked
together, the study was repeated and the 59Fe(III)EDTA was added before cooking. When
this was done there was no significant difference
in
the absorption of the intrinsic
Fe
and
the extrinsic Fe administered as Fe(II1)EDTA
(t
2.05,
P
>
0.1). The observation was
confirmed in another similar study (Table 4) in which 59Fe(III)EDTA was added to
a
traditional intrinsically-labelled pea dhal before cooking. Although the absorption of the
Fe in 59Fe(III)EDTA was marginally greater in each subject there was no significant
difference between the geometric mean absorptions
(t
0.88,
P
>
Oal), and the ratio of the
geometric mean absorption from extrinsic Fe to corresponding value for intrinsic food Fe
was close to unity.
Excretion
of
Fe(III)EDTA in the urine
Since EDTA chelates are excreted by the kidney (Will
&
Vilter, 1954) a study was
undertaken to assess whether Fe given orally as Fe(II1)EDTA appeared
in
the urine.
Fourteen fasting subjects with normal renal function drank 100 ml water sweetened with
10 mg sugar fortified with
5
mg Fe as 59Fe(III)EDTA. The total urine output for the next
24 h was collected and the percentage of the dose of 59Fe appearing in the urine was
compared to that absorbed (Fig. 1). In
all
instances less than 1
%
of the administered dose
appeared in the urine (geometric mean
0.32%,
SD
range 0.18-0.60) and serial sampling
showed that all the radio-Fe was excreted in the first 24
h.
A significant inverse relationship
Table
3.
The effects of increasing amounts of ascorbic acid (AA)
on
the absorption of iron from Fe(IIl)EDTA
(5
mg Fe) given in either
100
ml water
(A)
or
250
g maize porridge
(B)
to fasting female subjects
(Mean values and standard deviations)
3
to
40%
Q
Fe absorption
(%)*
0
Standardized
3:
Serum* Actual absorption Reference7 reference
B
2
~ ~ ~
NoAA AA
$
AA Haemoglobin saturation
~
No.
of
in Fe:AA (g/l)
(%)
Geo- Geo- Geo- Geo- Geo- Geo-
sub-
meal Feeding molar metric metric metric metric metric metric
2
jects (mg) regimen ratio Mean
SD
Mean
SD
mean
*I
SD
mean
*I
SD
mean
SD
mean
*I
SD
mean mean
3
ferritin No
AA
AA
salt
absorption
Transferrin
@g/l)
II
25
A
1:1.5 134 20 24 14 25 (7-85) 6.4 (3.7-11.2) 6.6 (3'0-14.8) 37.4 (17.7-79.1) 7.2
8.0
10
50
A
1:3 133 12 36 23 31 (1G94) 7.7 (3.5-17.0)
11.0
(5.2-23.3) 26-8 (15.845.5) 9.8 16.8
9
100
A
1:6 134 26 25
15
14 (6-35) 7.0 (4.1-11.8) 47.7 (16'7-136-4) 41.5 (22'8-75.5) 6.8
44.0
12 25
B
111.5 142 12 26 14 19 (6-59) 6.4 (3.1-13.5) 6.9 (2.8-17.0) 25.4 (14~046~0) 8.5
8.9
I1
50
B
1:3 137 23 28
II
33 (11-103) 6.3 (2.9-13.4) 6.0 (2.5-14.7) 20.4 (6.465'0) 12.0 12.0
9
100
B
1:6 138 13 29 11
21
(6-71) 6.1 (2.3-16.1)
12.0
(2.8-13-5) 30.5 (15+62.1)
8.0
15.6
*
Geometric means and
SD
ranges
used
because
values were positively skewed.
t
3
mg Fe as ferrous ascorbate (molar ratio for Fe and AA of
1
:
3)
given in fasting state.
6
Table
4.
The absorption
of
Fe from intrinsically-labelled food
iron
(55Fe) and "Fe(ZZZ)EDTA
(3
mg Fe) given in the same meal to
fasting female subjects
(Mean values and standard deviations)
Fe absorption
PA)*
Serum*
Intrinsic
ferritin Reference? label 5BFe(III)EDTA
Transferrin
OldU
salt (56Fe) label
Haemoglobin saturation
Total (g/l)
(%)
Geo-
Geo- Geo- Geo- Ratio
NO.
Of
Fe metric metric metric metric 5BFe/
sub- (mg) Mean SD Mean
SD
mean
SD
mean
_+I
SD
mean
_+I
SD
mean
+1
SD 56Fe
jects
56Fe Maize pomdge (250 g)
10
6.5 131 13 24.1 13.7 19.7 (6.8-56.9) 23.8 (12.8-44-3) 4.9 (2'1-11.7) 12.9 (5'6-29.7) 2.6
58Fe(III)EDTA sugar (20 g)
65Fe Maize pomdge
(250
g) 9 7.3 148 12 25.3 8.7 25.3 (8.1-79-0)
-
-
5.5
(1.5-20.5) 8.1 (3.5-18.6) 1.2
Cooked
with
sBFe(III)EDTA
with 58Fe(lII)EDTA
55Fe
Dhal
cooked
(200
ml)
9 6.4 142
11
25.5 9.3 25.9 (16.540.6) 27.3 (12.1-61.4) 5.2 (2'3-12'0) 5.4 (2.1-13.9) 1.04
Geometric means and
SD
ranges used because values were positively skewed.
t
3 mg Fe as ferrous ascorbate given in fasting state.
z
U
0
4
21
m
w
v1
Iron
absorption
from Fe(1II)EDTA
0.8
0.6
I
0.4
U
8
-
(u
C
._
L
3
-
-
-
o.2
t
0
-.
0
'0
a
0
.
0
223
0.1
L
1
0
I
II
I
1
2
4
6
8
10
20
30
Absorption
(%)
Fig.
I.
The relationship between the excretion of 6BFe in the urine collected over
24
h
(%
dose) and the
absorption of 6BFe from
3
mg Fe as SBFe(III)EDTA given
in
water to fourteen subjects after an overnight
fast
(r
-0.58,
P
<
0.05).
was noted between the logarithm of the percentage
of
the dose absorbed and that excreted
in the urine
(r
-0.58,
P
<
0.05).
The appearance of radio-Fe in the urine provided a means whereby intraluminal exchange
of Fe could be studied. A similar study was therefore carried out in ten subjects using two
isotopes of Fe. Equimolar quantities of 55Fe(III)EDTA and 5eFeS04. 7H,O were mixed into
a 250 g maize porridge after cooking. The mean for absorption of 65Fe: absorption of 6BFe
from the meal was 1.22 (kO.11) which was similar to the value of 1.28
(kO.08)
found in
the urine
(t
2.19,
P
>
0.05).
Since these results suggested that there was some exchange
between the isotopes an in vitro experiment was performed using a modification
of
a method
described by Bezwoda
et
al.
(1978) in order to try and verify the point. The solubilization
of 59Fe from portions ofcooked maize meal, fortified with 5BFeS04. 7H,O or 5eFe(III)EDTA,
at various pH values is shown in Fig. 2. The percentage SeFe appearing in the supernatant
fraction at pH values greater than those normally found in the stomach was lowest when
5BFeS04. 7H,O was used alone and highest when 5eFe(III)EDTA was used alone. However,
if
an equimolar quantity of unlabelled Fe(II1)EDTA was added to the 5eFeS04.7H20
fortified maize the percentage 59Fe appearing in the supernatant fraction increased
markedly, shifting the curve to approximately midway between that produced by
59Fe(III)EDTA alone and 5BFeS04. 7H,O alone.
DISCUSSION
The prevalence of Fe deficiency has been shown to be approximately 26% in the population
from which the Indian women who volunteered for the absorption studies was drawn (Mayet
et
al.
1972). The current findings tend to confirm this in that in none of the groups did the
geometric mean of the serum ferritin concentration exceed the normal adult female value
of 35 pg/l (Jacbos
et
al.
1972). In addition, the over-all mean absorption
(%)
of
3 mg Fe
as ferrous ascorbate was 31.0 (14.3-67.6), which can
be
taken as further evidence that the
over-all group was avid for Fe (Kuhn
et
a[.
1968). This background of Fe deficiency gave
224
A.
P.
MACPHAIL
AND
OTHERS
1
2
3
4
5
6
PH
Fig.
2.
The percentage 5sFe appearing in the supernatant fraction after incubation
of
portions
of
maize
porridge fortified with Fe at various
pH
values.
(0-0)
sBFeSO,.
7H,O;
(A-A)
58Fe(III)EDTA;
(0-0)
S8FeS0,.
7H,O
and Fe(1II)EDTA. Results are the mean values with their standard errors
represented
by
vertical
bars
of
three experiments done in duplicate.
added relevance to the in vivo Fe absorption findings, since it is to target populations such
as this that Fe-fortification programmes should be directed. However, a major deterrent
to the mounting of such programmes is the fact that the fortification-Fe is as poorly
absorbed as is the intrinsic Fe present in cereal staples (Layrisse
et
al.
1973). In addition,
ionizable Fe salts can produce a number of undesirable colour changes in food and can
adversely affect the storage properties
of
certain foods (International Nutritional Anemia
Consultative Group, 1977). It is against this background that the current findings should
be
seen.
The chelator EDTA forms a very stable complex wi’h Fe and because of this the
Fe(II1)EDTA complex might be expected to be protected to some extent from the ligands
in cereals which inhibit Fe absorption. The results of the present investigation confirmed
that Fe(II1)EDTA was well absorbed and utilized for haemoglobin formation and that its
absorption was not inhibited by bran or maize porridge. Nevertheless more powerful
inhibitors of absorption such as tea were able to dclrease the absorption of Fe from
Fe(II1)EDTA. The absorption
of
Fe(II1)EDTA was alsrl altered by the addition of ascorbic
acid. In the absence of inhibitors (Table
3
A) ascorbic
’I
:id
did
promote
Fe
absorption from
Fe(II1)EDTA in a similar dose-dependent manner
ic;
that which has been previously
demonstrated with simple Fe salts (Sayers
et
al.
1973; ‘*a
.s,
Lynch, Charlton
&
Bothwell,
1974; Sayers, Lynch, Charlton, Bothwell
et
al.
1974;
Uj’
c-Rasmussen
&
Hallberg, 1974).
Thus, when present at sufficiently-high concentration
a
xorbate was able to compete
successfully for the Fe in Fe(II1)EDTA. There were, however, quantitative differences
between the behaviour of Fe(1II)EDTA and simple Fe salts. For example, in one study
(Table 3) the absorption
(%)
of Fe from Fe(1II)EDTA alone was 7.7 compared with
26.8
from the ferrous ascorbate reference salt. When ascorbate was added to give a molar ratio
of 3:
1
for ascorbate and Fe (which is the same as that in the reference salt) absorption
increased to only 11. This result suggests that perhaps one-sixth
of
the
Fe
was transferred
to ascorbate in the presence of a 3-fold molar excess of ascorbic acid. Fe absorption from
Fe(1II)EDTA was increased to the same level as that from the reference salt only when the
molar ratio for ascorbate and Fe was
6:
1. Under such circumstances all the Fe was
presumably in the form of ascorbate. In the presence of maize there was no improvement
Iron
absorption
from
Fe(IIl)EDTA
225
in the absorption of Fe from Fe(II1)EDTA until a molar ratio for ascorbate and Fe of
6:
1
was present and even then the increase was a modest one, from a standardized absorption
(%)
of approximately 8-1
5.6.
These results suggest that although ascorbate Fe is very well
absorbed its absorption was reduced considerably (from
44
to
15)
by the presence of maize
porridge. On the other hand, the standardized absorption of EDTA-complexed Fe,
although only 6-10, was not significantly affected by the presence of the maize porridge.
In this context, the unchanged absorption of Fe from Fe(I1I)EDTA in maize porridge when
the molar ratio for ascorbate and Fe was
1
:
3
is not surprising if it is accepted that only
one-sixth
of
the Fe was present
as
the ascorbate complex and that only approximately 12%
of this was absorbed (Table
3).
These results suggest that competitive binding
of
Fe occurs
in the gastrointestinal tract and that the amount of Fe absorbed depends
on
the relative
affinities for Fe of the various ligands which either promote or inhibit Fe absorption.
Most Fe salts which are added as fortificants to cereals have been shown to enter
a
‘common non-haem-Fe pool’ and are absorbed to the same extent as the intrinsic food Fe
(Layrisse
&
Martinez-Torres, 1971
;
Cook
et
ul.
1972; Bjorn-Rasmussen
&
Hallberg, 1972).
Although Fe(II1)EDTA appears to form a common pool with food Fe in some situaiions
it
is preferentially absorbed in others. Thus when Fe(II1)EDTA was cooked with either
maize porridge or a @a-based dhal, absorption of extrinsic Fe(II1)EDTA: absorption
of
intrinsic food Fe was close to unity. However, absorption
of
Fe from Fe(II1)EDTA added
to pre-cooked maize porridge was
3-fold
higher than that of the intrinsic maize Fe, which
was presumably due to incomplete mixing. It should be noted that even when the relative
absorption value was close to unity the absorption of the radio-Fe added as Fe(II1)EDTA
was always marginally greater. A similar deduction can be made from results obtained by
Layrisse and co-workers (Layrisse
&
Martinez-Torres, 1977; Martinez-Torres
et
ul.
1979).
Statistical analysis of their values using Student’s
t
test for paired values indicates the
absorption from Fe(II1)EDTA
(5
mg and
50
mg Fe) given in a maize-soyabean meal was
significantly greater than from the intrinsic Fe
(t
3.54,
P
<
0.005;
t
5.04,
P
<
0.005
respectively). This may indicate that the Fe in the Fe(II1)EDTA complex may frequently
not exchange completely with the intrinsic Fe or indeed with added Fe salts. Evidence that
there is indeed an exchange between the Fe in the EDTA complex and other Fe salts was
provided by the demonstration of both 65Fe and 58Fe in the urine of subjects given porridge
fortified with 55Fe(III)EDTA and 59FeS0,.7H,0. Since none of the Fe absorbed from
68FeS0,. 7H,O is excreted in the urine, the appearance
of
almost equal quantities of the
two isotopes in the urine implies that almost complete exchange must have occurred in the
lumen of the gut. This exchange process was confirmed in vitro by the enhanced
solubilization of 5sFe from 68FeS0,
.
7H2O-fortified maize porridge by the addition of
unlabelled Fe(II1)EDTA. The relatively good absorptions seen when Fe(II1)EDTA was used
to fortify cereal-based meals can probably be ascribed to the fact that Fe in the form of
the Fe(II1)EDTA complex remains in solution and relatively unaffected by inhibitory
ligands in food even at pH
5.
Thus in the region of the duodenum and jejunum, where most
Fe absorption takes
place,
the Fe is in
a
highly-soluble form and formation of insoluble,
poorly-absorbed Fe complexes is prevented. There is one final point concerning the
exchange of Fe between Fe(II1)EDTA and intrinsic food Fe which merits comment. In one
experiment FeSO, was given with porridge on one morning and Fe(I1I)EDTA was fed on
the other. Since both Fe compounds form a common pool with food Fe it would have been
expected that Fe absorption on the
2
days would have been the same. It was, however, twice
as great when the fortificant was Fe(II1)EDTA. This suggests that the mechanism whereby
Fe(1II)EDTA forms a common pool with intrinsic Fe differs from simple Fe salts. The
apparent enhancement of intrinsic food Fe absorption by Fe(III)EDTA, which was also
previously noted by Layrisse
&
Martinez-Torres (1977), could
be
due to an exchange of
A.
P.
MACPHAIL
AND
OTHERS
isotope between intrinsic 55Fe and 59Fe in the EDTA chelate. The intrinsic 55Fe which
complexes with EDTA remains in solution and is well absorbed. On the other hand, in the
instance of simple
Fe
salts no ‘solubilizing’ chelate is present and the Fe is bound to
inhibitory ligands in the food and is as poorly absorbed as the intrinsic Fe.
The fact that small amounts of radio-Fe appeared in the urine after ingestion of labelled
Fe(I1I)EDTA indicates that some of the Fe was probably absorbed intact in the form of
the Fe(II1)EDTA complex, since Fe and other EDTA chelates have been shown to be
rapidly excreted by the kidney (Will
&
Vilter,
1954).
In the present study all radio-Fe was
excreted in
18-24
h and in no subject did the amount exceed
1
%
of the administered dose.
An inverse relationship was noted between the Fe absorbed and the Fe appearing in the
urine (Fig.
1)
which suggests that less Fe was absorbed as the chelate when the mucosa was
avid for Fe. At the same time, it was noteworthy that most of the Fe was probably absorbed
in a non-chelated form, even by individuals who were not Fe-deficient. Whether the
Fe
dissociates from the complex before entering the mucosa or within the mucosa is not clear.
The results of the present study confirm the potential usefulness of Fe(II1)EDTA as an
Fe
fortificant (Layrisse
&
Martinez-Torres,
1977;
Martinez-Torres
et
af.
1979).
It
can
be
added to food with little or no colour change and is well absorbed even in the presence
of known inhibitors. If
it
can
be
shown that the long-term administration of Fe(1II)EDTA
does not adversely affect trace mineral metabolism, then
it
may well fulfil an important role
in Fe-fortification programmes.
This work was supported by grants from the South African Atomic Energy Board and
the South African Sugar Association. The authors are indebted
to
Mrs Uta Schmidt, Mrs
Devi Moodly, Miss Fawzia Khan and Mrs Jane Lamprey for their invaluable assistance.
REFERENCES
Bezwoda, W., Charlton, R., Bothwell, T., Torrance,
J.
&
Mayet, F.
(1978).
J.
Lab. din.
Med.
92,
108.
Bjorn-Rasrnussen,
E.
&
Hallberg, L.
(1972).
Am.
J.
clin. Nutr.
25,
317.
Bjorn-Rasrnussen,
E.
&
Hallberg, L.
(1974).
Nutr. Metab.
16,
94.
Bothwell, T.
H.,
Charlton, R.
W.,
Cook,
J.
D.
&
Finch, C.
A.
(1979).
In
Iron Metabolism
in
Man,
p.
443.
Oxford:
Cook,
J.
D., Layrisse, M., Martinez-Torres. C., Walker, R.
B.,
Monsen,
E.
&
Finch. C.
A.
(1972).
J.
clin. Invest.
Deppe,
W.
M., Joubert,
S.
M.
&
Naidoo,
P.
(1978).
J.
din.
Path.
31,
145.
Derman, D., Sayers, M., Lynch,
S.
R., Charlton, R. W.
&
Bothwell,
T.
H.
(1977).
Br.
J.
Nutr.
38,
261.
Disler,
P.
B., Lynch,
S.
R., Charlton, R. W., Bothwell,
T.
H., Walker, R. B.
&
Mayet, F.
(1975).
Br.
J.
Nutr.
34,
Disler, P.
B.,
Lynch,
S.
R., Charlton,
R.
W., Torrance,
J.
D.,
Bothwell. T.
H.,
Walker, R. B.
&
Mayet, F.
(1975).
Hallberg, L., Bjom-Rasmussen,
E.,
Garby,
L.,
Pleehachinda, R.
&
Suwanik, R.
(1978).
Am.
J.
din.
Nutr.
31,
1403.
Herbert,
V.,
Gottlieb, C. W.
&
Lau, K.
S.
(1967).
J.
nucl.
Med.
8,
529.
Hussain, R., Walker, R. B., Laynsse, M., Clark,
P.
&
Finch, C.
A.
(1965).
Am.
J.
clin. Nutr.
16,
464.
International Commission for Radiation Protection
(1960).
Report
of
Committee
I I
on
Permissible
Dose
of
Internal
International Committee for Standardisation in Haematology
(1978).
Br.
J.
Huemat.
38,
291.
International Nutritional Anemia Consultative Group
(1977).
Guidelines
for
the
Eradication
of
Iron
Deficiency
Jacobs,
A.,
Miller,
F.,
Wonvood, M., Beamish, M. R.
&
Wardrop, C.
A.
(1972).
Br.
med.
J.
iv,
206.
Katz,
J.
H., Zoukis, M., Hart,
W.
L.
&
Dern, R.
J.
(1964).
J.
Lab. din. Med.
63,
885.
Kuhn,
I.
N., Monsen,
E.
R., Cook,
J.
D.
&
Finch, C.
A.
(1968).
J.
Lab.
din.
Med.
71,
715.
Layrisse, M., Cook,
J.
D., Martinez-Torres. C., Roche, M., Kuhn,
I.
N., Walker, R. B.
&
Finch,
C.
A.
(1969).
Layrisse, M.
&
Martinez-Torres,
C.
(1971).
Prog. Haemat.
7,
137.
Layrisse, M.
&
Martinez-Torres, C.
(1977).
Am.
J.
clin. Nutr.
30,
1166.
Layrisse, M., Martinez-Torres, C., Cook,
J.
D.,
Walker, R.
&
Finch, C.
A.
(1973).
Blood41,
333.
Lorber, L.
(1927).
Biochem.
Z.
181,
391.
Martinez-Torres, C., Romano,
E.
L.,
Renzi,
M.
&
Layrisse,
M.
(1979).
Am.
J.
din.
Nutr.
32,
809.
Blackwell.
51,
805.
141.
Gut
16,
193.
Radiation 1959. ICRP Publication
no.
2
Oxford: Pergamon Press.
Anaemia,
p.
16.
New
York:
Nutrition Foundation.
Blood
33,
430.
Iron
absorption
from
Fe(11I)EDTA
227
Mayet,
F.
G.
H.,
Adams,
E.
B., Moodley,
T.,
Kleber,
E.
E.
&
Cooper,
S.
K.
(1972).
S.
Ajr.
med.
J.
46,
1427.
Sawyer,
D.
T.
&
McKinnie,
J.
M. (1960).
J.
Am.
Chem.
SOC.
82,4191.
Sayers,
M.
H.,
Lynch,
S.
R.,
Charlton,
R.
W.
&
Bothwell,
T.
H.
(1974).
Br.
J.
Nutr.
31,
367.
Sayers, M.
H.,
Lynch,
S.
R.,
Charlton,
R.
W.,
Bothwell,
T.
H.,
Walker,
R.
B.
&
Mayet,
F.
(1974).
Br.
J.
Huemat.
Sayers, M.
H.,
Lynch,
S.
R.,
Jacobs,
P.,
Charlton,
R.
W., Bothwell,
T.
H.,
Walker,
R.
B.
&
Mayet,
F.
(1973).
Will,
J.
J.
&
Vilter,
R.
W.
(1954).
J.
Lab. din.
Med.
44,
499.
28, 483.
Br.
J.
Haemat.
24,
209.
Printed in Great Britain