796. Sodium iron EDTA (WHO Food Additives Series 32)

SODIUM IRON EDTA

First draft prepared by
Dr I.C. Munro
CanTox Inc., Mississauga, Ontario
Canada

1 EXPLANATION

The Committee was asked to comment on the safety of sodium iron
(III) ethylenediaminetetraacetate (sodium iron EDTA, sodium iron
edetate, NaFeEDTA, NaFe(III)EDTA) as a dietary supplement for use in
supervised food fortification programmes in populations in which
iron-deficiency anaemia is endemic. The Committee was informed that
use of iron in this form would be restricted to this specific
application and would be supervised.

Sodium iron EDTA has not previously been evaluated by the Joint
FAO/WHO Expert Committee on Food Additives. However, disodium and
calcium disodium EDTA were evaluated at the seventeenth meeting
(Annex 1 reference 32). An ADI of 2.5 mg EDTA CaNa₂EDTA/kg body
weight/day was established. Sodium iron EDTA was placed on the
agenda to provide an assessment of its safety for use in supervised
food fortification programmes in populations in which iron
deficiency anaemia is endemic.

With respect to iron, a provisional maximum tolerable daily
intake of 0.8 mg/kg/bw was established by the Committee at the
twenty-seventh meeting (Annex 1, reference 62). The view stated at
the twenty-sixth meeting (Annex 1, reference 59), that the tolerable
daily intake should not be used as a guide for fortifying processed
foods, was reiterated. This monograph discusses the safety of
NaFeEDTA for food fortification in developing countries.

2. BIOLOGICAL DATA

2.1 Biochemical aspects

The biochemistry of EDTA metal complexes is inextricably tied
to their chemical properties. An understanding of these chemical
properties is essential in interpreting the biochemistry and
toxicology of EDTA metal complexes. A brief discussion of the
chemical properties of EDTA metal complexes is presented here to
facilitate understanding of the material that follows.

Ethylenediaminetetraacetic acid (EDTA) is a hexadentate
chelator capable of combining stoichiometrically with virtually
every metal in the periodic table (Chaberck and Martell, 1959). With
divalent or trivalent metal ions a neutral or anionic metal chelate
results. The metal is largely prevented from reacting with
competing anions and its solubility is greatly increased. The
effectiveness of EDTA as a chelate for a particular metal ion is
given by its stability constant with the metal ion. Chelation
potential is affected by pH, the molar ratio of chelate to metal
ion, and the presence of competing metal ions capable of forming
complexes with EDTA (Plumb et al., 1950; Martell, 1960; Hart,
1984). The stability constants for different metal-EDTA complexes
vary considerably and any metal which is capable of forming a strong
complex with EDTA will at least partially displace another metal.

Of the nutritionally important metals, Fe³⁺ has the highest
stability constant (log k of 25.1), followed by Cu²⁺ with 18.4,
Zn²⁺ with 16.1, Fe²⁺ with 14.6, Ca²⁺ with 10.6, mg²⁺ with 8.7
and Na⁺ with 1.7 (West and Sykes, 1960). The situation is somewhat
complicated by each metal having an optimum pH for chelate formation
ranging from pH 1 for Fe³⁺, to pH 3 for Cu²⁺, pH 4 for Zn²⁺, pH
5 for Fe²⁺, pH 7.5 for Ca²⁺, and pH 10 for mg²⁺ (West and
Sykes, 1960). When NaFeEDTA is ingested with foods, the Fe^{3+ ion}
would be expected to remain firmly bound to the EDTA moiety during
passage through the gastric juice, but could be exchanged for
Cu²⁺, Zn²⁺, Fe²⁺ or Ca²⁺ in the duodenum. Similarly when
Na₂EDTA and Na₂CaEDTA are consumed with foods, the Na⁺ and Ca²⁺
ions would be predominantly exchanged in the gastric juice for Fe³⁺
ions, which could again in turn be exchanged for Cu²⁺, Zn²⁺, Fe²⁺
or Ca²⁺ further down the gastrointestinal tract. The extent to
which the metal EDTA complexes form is dependent on the pH and the
concentration of the competing metals as well as competing ligands.
The lower stability constant and higher pH optimum of the mg-EDTA
chelate make reaction with this metal less likely.

An appreciation of the chelating properties of EDTA with
respect to iron provide the basis of our understanding of the
observed effects of EDTA on food iron absorption. Ferric food iron
is poorly absorbed by human beings because it is precipitated from
solution above pH 3.5 unless suitable complexing agents are present.

It may therefore be partially insoluble in the upper small intestine
where most nonhaem iron is absorbed (Conrad and Schade, 1968;
MacPhail et al., 1981).

When EDTA is present in a meal, iron (primarily Fe³⁺) remains
complexed with EDTA under the acidic conditions prevailing in the
stomach. The chelate holds the iron in solution as the pH rises in
the upper small intestine, but the strength of the complex is
progressively reduced allowing at least partial exchange with other
metals and the release of some of the iron for absorption. There is
convincing evidence that iron chelated by EDTA (NaFeEDTA) is
available for absorption via the physiologically regulated pathways
responsible for iron uptake (Candela et al., 1984). The results of
absorption studies with NaFeEDTA indicate that iron is dissociated
from the EDTA moiety prior to absorption. The results of these
studies are summarized in Section 2.1.1.

2.1.1 Absorption and excretion

2.1.1.1 Absorption and excretion of Fe from NaFeEDTA - Injection
studies

When NaFeEDTA is injected intravenously into rats most of the
iron (70-90%) is lost through the urine within 24 hours (Najarajan
et al., 1964; Anghileri, 1967). A small proportion enters the
physiological iron pool destined primarily for haemoglobin synthesis
probably because of the slow release of iron to the iron transport
protein, transferrin, in the circulation (Bates et al., 1967).
After intramuscular or intraperitoneal injection a greater
proportion of the iron is available for physiological exchange with
compartments in the bone marrow and liver. The longer contact time
between transferrin and EDTA, allows for greater transfer of iron
from the chelate to the physiological transport protein (Rubin
et al., 1970).

FeEDTA administered intravenously to humans was almost
quantitatively excreted in urine (Lapinleimu and Wegelius, 1959).

2.1.1.2 Absorption and excretion of Fe from NaFeEDTA - oral studies

Human iron deficiency anaemia was successfully treated with
FeEDTA given orally with 84% of labelled FeEDTA excreted in the
faeces and none in the urine. Red cells, however, contained
labelled Fe and reticulocytosis occurred (Will and Vilter, 1954).

Studies carried out in swine using a doubly labelled
Na⁵⁵Fe[2-¹⁴C]EDTA preparation demonstrated rapid transfer of ⁵⁵Fe
to the plasma with a peak at 1 hour and subsequent incorporation of
4.6% of the administered dose into circulating haemoglobin (Candela
et al., 1984). A small fraction (0.3%) of the ⁵⁵Fe administered
was excreted in the urine. In contrast to the ⁵⁵Fe only a small

percentage of the ¹⁴C could be detected in the plasma at any time.
Absorption occurred over an extended period (5-20 hours). A total
of about 5% of the ¹⁴C labelled EDTA was eventually absorbed in the
duodenum and jejunum and quantitatively excreted in the urine.

In a parallel experiment when 5 mg Fe as Na⁵⁹FeEDTA was given
to six fasting human volunteers, mean radioiron absorption as
measured by red blood cell utilization was 12.0%. Only 0.3% of the
administered dose of iron was excreted in the urine. The studies
demonstrate that Fe and EDTA are absorbed independently when
NaFeEDTA is administered by mouth (Candela et al., 1984).

Similar conclusions were reached in an earlier human absorption
study carried out by MacPhail et al. (1981). Na⁵⁹FeEDTA was
administered to human volunteers. Between 3 and 25% of the ⁵⁹Fe
was absorbed, but less than 1% of the administered ⁵⁹Fe appeared in
the urine over the subsequent 24 hours. All ⁵⁹Fe absorbed in the
form of the intact Na⁵⁹FeEDTA complex would be expected to be
excreted in the urine within 24 hours (based on the results of
Nagarajan et al., 1964 and Anghileri, 1967) demonstrating again
that most of the iron is released from the EDTA complex before
absorption. Similar conclusions have been reached with another iron
chelator (nitrilotriacetic acid), the properties of which have been
studied extensively in experimental animals (Simpson and Peters,
1984).

2.1.1.3 Absorption and excretion of EDTA from EDTA metal chelates

Rats

¹⁴C-labelled CaNa₂EDTA, when fed to rats at 50 mg/kg bw, was
absorbed only to an extent of 2 to 4%; 80 to 90% of the dose
appeared in the faeces within 24 hours, and absorption was still
apparent at 48 hours. At the low pH of the stomach the calcium
chelate is dissociated with subsequent precipitation of the free
acid (EDTA), and this is only slowly redissolved in the intestine
(Foreman et al., 1953).

In feeding experiments in rats receiving disodium EDTA at
dietary levels of 0.5, 1.0 or 5.0%, the faeces contained 99.4, 98.2
and 97.5% of the excreted material (Yang, 1964).

Similar experiments conducted also in rats gave essentially the
same results. Thirty-two hours after a single dose of 95 mg
disodium EDTA/rat, 93% was recovered from the colon. After doses of
47.5, 95.0 and 142.5 mg disodium EDTA the amount of EDTA recovered
in the urine was directly proportional to the dose given, suggesting
that EDTA was absorbed from the gastrointestinal tract by passive
diffusion. The motility of the intestine was not affected by the
compound (Chan, 1964).

When 200 mg CaNa₂EDTA was introduced into the duodenum of
rats an absorption rate of 6.5 to 26% was observed (Srbrova and
Teisinger, 1957).

The maximum radioactivity in the urine after application of
¹⁴C-labelled CaNa₂EDTA to the skin was only 10 ppm (0.001%)
(Foreman and Trujillo, 1954).

Humans

Experiments in humans also revealed poor absorption; only 2.5%
of a 3 g dose given was excreted in the urine (Srbrova and
Teisinger, 1957). These authors also confirmed the dissociation of
the calcium chelate in the stomach. A dose of 1.5 mg of
¹⁴C-labelled CaNa₂EDTA given in a gelatine capsule to normal
healthy men was absorbed to an extent of 5% (Foreman and Trujillo,
1954).

The absorption of the EDTA moiety from orally administered
NaFeEDTA has not been measured directly in humans. However
physicochemical considerations indicate that EDTA absorption from
NaFeEDTA should be similar to that from other metal complexes, such
as CaNa₂EDTA and CrEDTA. As described above, poor absorption of
the intact NaFeEDTA can be inferred from the measurements of urinary
radioiron excretion after the oral administration of Na⁵⁹FeEDTA
made by MacPhail and coworkers in 1981.

Similar results have been obtained with a tightly bound
chelate, ₅₁CrEDTA from which any released metal is very poorly
absorbed (Bjarnason et al., 1983; Aabakken and Osnes, 1990). Only
1-5% of a dose of ₅₁CrEDTA given in a fasting state is absorbed by
the healthy intestinal mucosa. In the presence of disorders of the
gastrointestinal tract the absorption may be doubled. The
₅₁CrEDTA that is absorbed appears to be taken up through
intercellular junctions as the intact complex. The amount absorbed
has been used as a measure of the integrity of the bowel mucosa.

In summary, most of the iron in NaFeEDTA is released to the
physiological mucosal uptake system before absorption. Only a very
small fraction of the NaFeEDTA complex (less than 1%) is absorbed
intact and this is completely excreted in the urine. An additional
small fraction (less than 5%) of the EDTA moiety is absorbed,
presumably bound to other metals in the gastrointestinal tract, and
is also completely eliminated in the urine.

2.1.1.4 Bioavailability of iron from NaFeEDTA

The results of iron absorption studies comparing the
bioavailability of iron from FeSO₄ and NaFeEDTA fortified foods are
listed in Table 1. For purposes of comparison the individual
absorption values have been standardized to a reference absorption

of 40% to remove the influence of varying iron requirements in
different subjects. A reference absorption value of 40% is assumed
to represent borderline iron deficiency (Hallberg et al., 1978).
The bioavailability of iron from FeSO₄ varies over a wide range
and correlates with the relative proportions of enhancers and
inhibitors known to be present in the meals.

Enhanced bioavailability was most marked in meals with poor
FeSO₄ bioavailability (FeSO₄ absorptions less than 4%). Between
2.1 to 2.9 times as much iron was absorbed under such circumstances.
This point is exemplified by the results of the study by Viteri
et al. (1978) (see Table 1) in which iron absorption from a
NaFeEDTA-fortified meal of beans, maize, and coffee was 2.7 times
greater than that from the same meal containing FeSO₄ (Viteri
et al., 1978).

In contrast, the absorption of iron from NaFeEDTA eaten with
identical meals varies only two to three fold. More iron was
absorbed from the meals containing NaFeEDTA in all but one case in
which Na₂EDTA and FeSO₄ were eaten with sugar cane syrup (see
Table 1).

The absorption of Fe from NaFeEDTA has been studied in a wide
variety of meals. Comparisons with iron absorption from simple iron
salts have not always been made. However, some studies provide
useful information about the suitability of three staple food items
as potential vehicles for fortification with NaFeEDTA. This
information is summarized in Table 2. Some studies listed in Table
1 have been included under the appropriate categories (all values
are corrected to 40% reference absorption, or a serum ferritin of
27 µg/l). It is evident that approximately 10% of the fortification
iron added would be absorbed by iron deficient individuals if these
staple foods were used as the vehicle for delivering the
fortificant.

Table 1. Comparison of iron absorption from meals of different iron bioavailability
fortified with ferrous sulfate or NaFeEDTA; Standardized Iron Absorption (%)^a

Components of Meal A B Ratio Reference
FeSO₄ NaFeEDTA B/A

1. Rice Milk 1.7 4.5 2.6 Viteri et al., 1978
2. Beans, Maize, Coffee 2.0 5.3 2.7 Viteri et al., 1978
3. Egyptian flat bread^b 2.1 5.3 2.5 el Guindi et al., 1988
4. Bran 2.7 7.8 2.9 MacPhail et al., 1981
5. Beans, Plantain, Rice, 3.1 7.0 2.3 Layrisse and Martinez-Torres, 1977
Maize, Soy^c
6. Rice 3.9 11.5 2.9 MacPhail and Bothwell, unpublished, 1992
7. Maize Meal 4.0 8.2 2.1 MaPhail et al., 1981
8. Beans, Plantain, Rice, 4.2 7.4 1.8 Layrisse and Martinez-Torres, 1977
Maize, Soy, Orange Juice^c
9. Beans, Plantain, Rice, 4.3 9.6 2.2 Layrisse and Martinez-Torres, 1977
Maize, Soy, Meat^c
10. Potato 5.9 7.3 1.2 Lamparelli et al., 1987
11. Wheat 6.2 14.6 2.3 Martinez-Torres et al., 1979
12. Milk 10.2 16.8 1.6 Layrisse and Martinez-Torres, 1977
13. Sweet Manioc 14.1 16.6 1.2 Martinez-Torres et al., 1979
14. Sugar cane Syrup^c 33.1 10.8 0.3 Martinez-Torres et al., 1979

a. Geometric means standarized to a reference (Ferrous ascorbate) absorption of 40%
b. A mixture of FeSO₄ and Na₂EDTA was used in this study.
c. Comparison between FeSO₄ and NaFeEDTA not in the same individuals.

Table 2. Percentage iron absorption from meals containing
NaFe(III)EDTA

Vehicle No. of Standardized References
Studies Iron Absorption
(Range)

Wheat 4 10.1 (5.3 - 14.6) Martinez-Torres et al.,
1979 and el Guindi et al.,
1988

Maize 7 9.1 (7.6 - 12.0) Martinez-Torres et al.,
1979 and MacPhail et al.,
1981

Cassava 3 13.5 (11.0 - 16.4) Martinez-Torres et al.,
1979

2.1.1.5 Effect of NaFeEDTA on bioavailability of intrinsic food
iron

Conclusions drawn from much of the experimental work on food
iron absorption and iron fortification are based on the observation
that soluble iron added to a meal and the intrinsic nonhaem food
iron behave as a common pool, which is equally susceptible to
enhancers and inhibitors of iron absorption present in the meal
(Cook et al., 1972; Hallberg and Bjorn-Rasmussen, 1972); NaFeEDTA
shares this property. When Na⁵⁹FeEDTA was added to meals
containing foods labelled intrinsically with ⁵⁵Fe, the ratio
between the proportions of iron absorbed from the two sources was
close to unity (Layrisse and Martinez-Torres, 1977; Matrinez-Torres
et al., 1979; MacPhail et al., 1981), with the exception of one
study in which Na⁵⁹FeEDTA fortified sugar was sprinkled onto
⁵⁵Fe-labelled maize immediately before it was eaten (MacPhail
et al., 1981). These results indicate that the Na⁵⁹FeEDTA
equilibrates with the common pool, since without such equilibration,
the amount of food iron absorbed would be much lower than the amount
absorbed from NaFeEDTA (MacPhail et al., 1981). Inadequate mixing
of the NaFeEDTA-fortified sugar with the maize meal probably
accounted for the lack of equilibration in the one inconsistent
study reported by MacPhail et al. (1981). These results reveal
another important property of NaFeEDTA. Equilibration of NaFeEDTA
with the common pool iron improves the bioavailability of the
intrinsic food iron as well. Therefore NaFeEDTA improves iron

balance by supplying iron in a form less affected by dietary
inhibitors, but also improves the absorption of nonhaem iron in the
meal derived from other sources.

This point is further illustrated by the results of a number of
studies which demonstrate that the positive effects of EDTA on iron
absorption are shared by other elements of the common pool, such as
another iron salt added to the meal. When FeSO₄ and NaFeEDTA were
fed to humans on separate days in the same type of meal (maize
porridge), iron absorption from the NaFeEDTA fortified meal was
significantly better. However the iron from FeSO₄ was as well
absorbed as that from NaFeEDTA when they were fed together in the
same meal (MacPhail et al., 1981; Martinez-Torres et al., 1979).
More direct evidence of reciprocal exchange between food iron and
iron added as NaFeEDTA was provided by experiments in which subjects
were given maize porridge fortified with equimolar quantities of
⁵⁹FeSO₄ and Na⁵⁵FeEDTA (McPhail et al., 1981). The ratio
between the two isotopes was almost the same in the meal and the
urine. This implies that exchange of iron between FeSO₄ and
NaFeEDTA must occur before absorption of the chelate, since only the
small amount of iron (less than 1%) absorbed as the intact chelate
would subsequently appear in the urine (for explanation see section
2.1).

2.1.1.6 The effect of Na₂EDTA on iron absorption

Na₂EDTA is widely used as a food additive to prevent
oxidative damage by free metals. Since Na₂EDTA readily chelates
iron in the gut to form NaFeEDTA, its effect on iron absorption is
of interest. In a recent study (el Guindi et al., 1988) Na₂EDTA
was added, together with an equimolar quantity of iron as FeSO₄, to
bread with a high concentration of phytate (an inhibitor of iron
absorption). The combination was associated with a 2.6x enhancement
in iron absorption when compared with results with FeSO₄ used
alone. Mean percentage iron absorption was approximately equivalent
to that reported in other similar studies using NaFeEDTA. It is
evident that the same effect on iron absorption can be achieved in
meals containing compounds that inhibit iron absorption by adding
Na₂EDTA and a soluble iron salt as is the case for adding
NaFeEDTA.

The effects of Na₂EDTA on iron absorption appear to be
influenced by the molar ratio of EDTA to iron. Earlier work
suggested that increasing the molar ratio of Na₂EDTA to Fe was
associated with a progressive reduction in iron absorption (Cook and
Monsen, 1976).

These observations have been extended recently: iron absorption
from a series of rice meals containing Na₂EDTA and iron in a molar
ratio of 1:1 was compared to rice containing Na₂EDTA and iron in
molar ratios (EDTA:Fe) ranging from 0:1 to 4:1. Statistically
significant enhancement of absorption occurred at ratios of Fe:EDTA
between 1:4 and 1:1. The enhancing effect of EDTA on iron
absorption appeared to be maximal at a molar ratio (EDTA:Fe) of
approximately 1:2, not 1:1 as previously assumed. At this molar
ratio over three times as much iron was absorbed from the EDTA
containing meal as was the case for the control meal containing no
EDTA (MacPhail and Bothwell, unpublished data, 1992).

2.1.2 Distribution

After parenteral administration to rats, 95 to 98% of injected
¹⁴C-labelled CaNa₂EDTA appeared in the urine within six hours.
All the material passed through the body unchanged. Peak plasma
levels were found approximately 50 minutes after administration.
Less than 0.1% of the material was oxidized to ¹⁴CO₂, and no
organs concentrated the substance. After i.v. injection,
CaNa₂EDTA passed rapidly out of the vascular systems to mix with
approximately 90% of the body water, but did not pass into the red
blood cells and was cleared through the kidney by tubular excretion
as well as glomerular filtration (Foreman et al., 1953). The same
was also found in man using ¹⁴C-labelled CaNa₂EDTA. Three
thousand milligrams were given i.v. to two subjects and were almost
entirely excreted within 12 to 16 hours (Srbrova and Teisinger,
1957). These results indicate that intact CaNa₂EDTA, and presumably
other EDTA metal complexes are rapidly excreted and do not
accumulate.

2.1.3 Biotransformation

Neither the iron nor the EDTA moiety of NaFeEDTA undergoes
biotransformation. Evidence for this conclusion comes from studies
discussed in the previous section which indicated that both EDTA and
iron are excreted unchanged following ingestion of NaFeEDTA.

2.1.4 Influence of EDTA compounds on the biochemistry of metals

EDTA removes about 1.4% of the total iron from ferritin at pH
7.4 to form an iron chelate (Westerfeld, 1961). Transfer of Fe from
Fe-transferrin to EDTA in vitro occurs at a rate of less than 1%
in 24 hours. In vivo studies in rabbits demonstrated transfer of
iron only from FeEDTA to transferrin and not the reverse. It
appeared that tissue iron became available to chelating agents
including EDTA only when an excess of iron was present (Cleton
et al., 1963). Equal distribution between a mixture of EDTA and
siderophilin was obtained only at EDTA:siderophilin ratios of
20-25:1 (Rubin, 1961).

Addition of 1% Na₂EDTA to a diet containing more than optimal
amounts of iron and calcium lowered the absorption and storage or
iron in rats and increased the amount present in plasma and urine.
The metabolism of calcium, however, was apparently unaffected
(Larsen et al., 1960). A diet containing 0.15 mg of iron, 4.26 of
calcium and 1 mg of EDTA/rat (equivalent to 100 ppm (0.01%) in the
diet) for 83 days had no influence on calcium and iron metabolism,
e.g. the iron content of liver and plasma (Hawkins et al., 1962).

Copper absorption and retention were improved at 500 mg EDTA/kg
but not at 200 mg or 1 000 mg EDTA/kg. Apart from a very small
increase in urinary copper excretion, dietary EDTA had no influence
on copper metabolism (Hurrell et al., 1993).

CaNa₂EDTA increased the excretion of zinc (Perry and Perry,
1959), and was active in increasing the availability of zinc in
soybean containing diets to poults (Kratzer et al., 1959).
CaNa₂EDTA enhanced the excretion of Co, Hg, Mn, Ni, Pb, Ti and W
(Foreman, 1961). The treatment of heavy metal poisoning with
CaNa₂EDTA has become so well established that its use for more
commonly seen metal poisonings, e.g. lead, is no longer reported in
the literature (Foreman, 1961). EDTA could not prevent the
accumulation of ⁹⁰Sr, ¹⁰⁶Ru, ¹⁴¹Ba and ²²⁶Ra in the skeleton.
⁹¹Y, ²³⁹Pu and ²³⁸U responded fairly well to EDTA, the excretion
being accelerated (Catsch, 1961).

Food fortification with NaFeEDTA may be expected to increase Zn
and Cu absorption and retention but not Ca nor Mg. A diet
containing RDA quantities of each metal (800 mg Ca and, 350 mg,
10 mg Zn, and 2 mg Cu) which was fortified with 10 mg Fe as NaFeEDTA
would contain a 1.5 molar excess of EDTA over Zn, an 8-fold molar
excess of EDTA over Cu, but 80 times less EDTA than Ca and 50 times
less EDTA than mg on a molar basis. The small quantity of chelate
with respect to Ca and mg would be unlikely to have any detrimental
effect. Both NaFeEDTA and NaEDTA may increase the absorption and
retention of Zn and Cu when added to low bioavailability diets.
This conclusion is supported by experiments with turkey poults
(Kratzer et al., 1959), chicks (Scott and Ziegler, 1963) and rats
(Forbes, 1961) which have demonstrated that Zn bioavailability and
animal growth is improved when Na₂EDTA is added at 150-300 mg/kg
to animal rations based on soybean protein isolate. The enhancing
effect of EDTA on zinc absorption in these studies can be explained
by a combination of two factors. Firstly, EDTA forms soluble
chelates with Zn from which the metal is potentially absorbable, and
secondly, Zn is prevented from forming non-absorbable complexes with
phytic acid. EDTA does not enhance Zn absorption when absorption
inhibitors are absent from the meal as evidenced by the observation
that Na₂EDTA (1 000 mg/kg) improved Zn absorption in rats fed a
casein-based diet with added phytic acid, but had no effect in the
absence of phytic acid (Oberleas et al., 1966).

Other chelating substances can also enhance Zn absorption from
low bioavailability diets. Vohra and Kratzer compared the growth
promoting effect of chelates with stability constants (log k) for Zn
varying from 5.3 to 18.8 in turkey poults fed zinc deficient diets
based on soy protein isolate. They found that
ethylenediaminediacetic acid-dipropionic acid, hydroxyethyl-EDTA,
and EDTA (stability constants 14.5, 14.5 and 16.1, respectively)
were the most effective (Vohra & Kratzer, 1964).

These earlier observations were made with Na₂EDTA. However,
NaFeEDTA has been shown to have similar properties in a recent
study. Zinc, copper, and calcium balances were performed in rats
fed low Zn (6.1 mg/kg) soybean based diets containing 36 mg/kg added
Fe as either ferrous sulfate or NaFeEDTA. In some experimental
groups additional Na₂EDTA was added to the diet containing
NaFeEDTA to give dietary EDTA levels of 200, 500 and 1 000 mg/kg.
Changing the iron compound in the diet from ferrous sulfate to
NaFeEDTA at a level of 200 mg/kg increased apparent Zn absorption,
urinary Zn excretion and Zn retention significantly (p < 0.05), but
caused no changes in Cu nor Ca absorption or excretion. Increasing
the dietary EDTA level to 500 mg/kg (molar ratio EDTA:Zn, 19:1) and
1 000 mg/kg (molar ratio EDTA:Zn, 38:1) further increased both Zn
absorption and urinary Zn excretion. At the highest dietary EDTA
level (1 000 mg/kg), Zn retention was significantly higher than with
no dietary EDTA, but lower than with 500 mg/kg EDTA. This resulted
from an increase in urinary excretion of Zn to 15.6% of intake.
Similar results were obtained with a Zn-sufficient (30 mg/kg)
soybean diets, but more EDTA was required to achieve optimal ratios
for improved absorption.

These studies demonstrate that an 11-fold molar excess of EDTA
over Cu increased Cu absorption and retention but that neither a 4.5
nor 23-fold molar excess had a significant effect. A human diet
containing the RDA for Zn and Cu which was fortified with 10 mg Fe
as NaFeEDTA would be expected to contain a 1.5 molar excess of EDTA
over Zn and an 8-fold molar excess of EDTA over Cu. NaFeEDTA
fortification would therefore be expected to have very little effect
on Zn and Cu balance. A small beneficial effect could occur in
meals containing little Zn or Cu or large quantities of phytate
(Hurrell et al., 1993).

The applicability of the observations made in experimental
animals to human nutrition has been confirmed by recent observations
made by Hurrell's group. The metabolism of Zn and Ca was studied
using a stable isotope technique in 10 adult women fed a breakfast
meal of bread rolls made from 100 g high extraction wheat flour and
fortified with 5 mg Fe as FeSO₄ or NaFeEDTA. The test meals
contained a 3.3 molar excess of EDTA over Zn but some 10-fold less
EDTA than Ca. Changing the Fe fortification compound from ferrous
sulfate to NaFeEDTA significantly increased ⁷⁰Zn absorption
(p<0.05) from this meal from 20.9% to 33.5%. Urinary ⁷⁰Zn

excretion also rose from 0.3% to 0.9%. Calcium metabolism was
similar with the two different iron compounds (Davidsson et al,
1993).

Earlier studies using less precise methodology have led to
similar conclusions. Adding NaFeEDTA to a low bioavailability
Guatemalan meal did not influence Zn absorption by human subjects.
However, as these workers measured Zn absorption based on plasma Zn
concentrations after ingesting 25 mg Zn with a meal, the molar
concentration of EDTA was some 10-fold less than that of Zn and an
improvement in Zn absorption would not be expected (Solomons et al.
(1979).

Finally, no significant changes in plasma Zn concentration were
observed in field studies in which NaFeEDTA was used as a food
fortificant over a two year period (Viteri et al., 1983, Ballot
et al., 1989b).

2.1.5 Effects on enzymes and other biochemical parameters

EDTA had a lowering effect on serum cholesterol level when
given orally or i.v. It may have acted by decreasing the capacity
of serum to transport cholesterol (Gould, 1961). Disodium EDTA had
a pyridoxin-like effect on the tryptophan metabolism of patients
with porphyria or scleroderma, due to a partial correction of
imbalance of polyvalent cations (Lelievre and Betz, 1961).

In vitro, 0.0033 M EDTA inhibited the respiration of liver
homogenates and of isolated mitochondria of liver and kidney
(Lelievre and Betz, 1961). The acetylation of sulfanilamide by a
liver extract was also inhibited (Lelievre, 1960). EDTA stimulated
glucuronide synthesis in rat liver, kidney and intestines but
inhibited the process in guinea-pig liver (Pogell and Leloir, 1961;
Miettinen and Leskinen, 1962). Of the heavy metal-containing
enzymes, EDTA at a concentration of about 10^-3 M inhibited aldehyde
oxidase and homogentisinase. Succinic dehydrogenase, xanthine
oxidase, NADH-cytochrome reductase and ceruloplasmin (oxidation of
p-phenylenediamine) were not inhibited (Westerfeld, 1961). Disodium
EDTA was found to be a strong inhibitor for delta-aminolevulinic
acid dehydrogenase, 5.5 x 10^-6 M causing 50% inhibition (Gibson
et al., 1955). The i.p. injection of 4.2 mmol/kg bw (equivalent
to 1722 mg/kg bw) CaNa₂EDTA caused in rats an inhibition of the
alkaline phosphatase of liver, prostate and serum up to four days
depending on the dose administered; zinc restored the activity
(Nigrovic, 1964).

In vitro, EDTA inhibited blood coagulation by chelating Ca.
The complete coagulation inhibition of human blood required
0.65-1.0 mg/ml. The i.v. injection of 79-200 mg EDTA/rabbit had no
effect on blood coagulation (Dyckerhoff et al., 1942).

I.v. injections of Na₂EDTA and CaNa₂EDTA had some
pharmacological effect on the blood pressure of cats; 0-20 mg/kg bw
CaNa₂EDTA (as Ca) produce a slight rise; 20-50 mg/kg, a biphasic
response; and 50 mg/kg, a clear depression (Marquardt and
Schumacher, 1957).

One per cent Na₂EDTA enhances the absorption of ¹⁴C-labelled
acidic, neutral and basic compounds (mannitol, inulin,
decamethenium, sulfanilic acid and EDTA itself) from isolated
segments of rat intestine, probably due to an increased permeability
of the intestinal wall (Schanker and Johnson, 1961).

2.2 Toxicological studies

2.2.1 Acute toxicity studies

The results of acute toxicity studies with disodium EDTA are
summarized in Table 3.

Table 3. Results of acute toxicity studies with disodium EDTA.

Animal Route LD₅₀ References
(mg/kg bw)

Rat oral 2 000 - 2 200 Yang, 1964

Rabbit oral 2 300 Shibata, 1956

i.v. 47^a Shibata, 1956

^a Dose depending on the rate of infusion

The results of acute toxicity studies with Ca-disodium EDTA are
summarized in Table 4.

Table 4. Results of acute toxicity studies with Ca-disodium EDTA.

Animal Route LD₅₀ References
(mg/kg bw)

Rat oral 10 000±740 Oser et al., 1963

Rabbit oral 7 000 approx. Oser et al., 1963

i.p. 500 approx. Bauer et al., 1952

Dog oral 12 000 approx. Oser et al., 1963

The oral LD₅₀ in rats is not affected by the presence of food
in the stomach or by pre-existing deficiency in Ca, Fe, Cu or Mn
(Oser et al., 1963).

Oral doses of over 250 mg/animal cause diarrhoea in rats
(Foreman et al., 1953).

There are many reports in the literature on kidney damage by
parenteral over-dosage of CaEDTA. A review was given by Lechnit
(1961). Lesions simulating "versene nephrosis" in man have also
been produced in rats. Disodium EDTA in doses of 400-500 mg i.p.
for 21 days caused severe hydropic degeneration of the proximal
convoluted tubules of the kidneys. CaNa₂EDTA produced only
minimal focal hydropic changes in 58% of animals, disappearing
almost two weeks after stopping the injections (Reuber and
Schmieller, 1962).

2.2.2 Short-term toxicity studies

2.2.2.1 Rats

Groups of five male rats received 250 or 500 mg/kg bw/dy
CaNa₂EDTA i.p. daily for three to 21 days and some were observed
for an additional two weeks. Weight gain was satisfactory and
histology of lung, thymus, kidney, liver, spleen, adrenal, small gut
and heart was normal except for mild to moderate renal hydropic
change with focal subcapsular swelling and proliferation in
glomerular loops at the 500 mg level. There was very slight
involvement with complete recovery at the 250 mg level. Lesions
were not more severe with simultaneous cortisone administration
(Reuber and Schmieller, 1962).

Groups of three male and three female rats were fed for four
months on a low mineral diet containing one-half the usual portion
of salt mixture (i.e. 1.25% instead of 2.50%) with the addition of
0% and 1.5% CaNa₂EDTA. The test group showed a reduced weight
gain, but there was no distinct difference in general condition of
the animals (Yang, 1964).

Groups of five male rats were given 250, 400 or 500 mg/kg bw/dy
disodium EDTA i.p. daily for three to 31 days; some groups were
observed for another two weeks. At the 500 mg level all rats became
lethargic and died within nine days, the kidneys being pale and
swollen, with moderate dilatation of bowel and subserosal
haemorrhages. Histological examination of a number of organs showed
lesions only in the kidneys. Animals at the 400 mg level died
within 14 days, kidney and bowel symptoms being similar to the 50 mg
level. One rat at the 250 mg dose level showed haemorrhage of the
thymus. All three groups showed varying degrees of hydrophic
necrosis of the renal proximal convoluted tubules with epithelial
sloughing: recovery occurred in all groups after withdrawal of
disodium EDTA (Reuber and Schmieller, 1962).

2.2.2.2 Rabbits

Eight groups of three rabbits were given either 0.1, 1, 10 or
20 mg/kg bw/dy disodium EDTA i.v., or 50, 100, 500 or 1 000 mg/kg
bw/dy orally for one month. All animals on the highest oral test
level exhibited severe diarrhoea and died. In the other groups body
weight, haemograms, urinary nitrogen and urobilinogen were
unaffected. Histopathological examination of a number of organs
showed degenerative changes in the liver, kidney, parathyroid and
endocrine organs and oedema in muscle, brain and heart at all levels
of treatment (Shibata, 1956).

2.2.2.3 Dogs

Four groups of one male and three female mongrels were fed
diets containing 0, 50, 100 and 200 mg/kg bw/dy CaNa₂EDTA daily
for 12 months. All appeared in good health, without significant
change in blood cells, haemoglobin and urine (pH, albumin, sugar,
sediment). Blood sugar, non-protein nitrogen and prothrombin time
remained normal. Radiographs of ribs and of long bones showed no
adverse changes at the 250 mg level. All dogs survived for one
year. Gross and microscopic findings were normal (Oser et al.,
1963).

2.2.3 Long-term toxicity/carcinogenicity studies

2.2.3.1 Mice

Groups of 50 male and 50 female B6C3F₁ mice received trisodium
EDTA (Na₃EDTA) in the diet at concentrations of 3 750 or 7 500 ppm
for 103 weeks, followed by one week during which standard diet
without EDTA was fed. A control group consisting of 20 mice of each
sex received the standard diet. Food was available ad libitum and
fresh food was provided three times per week.

Animals were examined for signs of toxicity twice per day, and
were weighed and palpated for masses regularly (schedule not
stated). Gross and microscopic pathological examinations were
performed on animals found dead or moribund and on those sacrificed
at the end of the study. Microscopic examinations were conducted on
the following tissues and organs: skin, lymph nodes, mammary gland,
salivary gland, bone marrow, trachea, lungs and bronchi, heart,
thyroid, parathyroids, oesophagus, stomach, small intestine, liver,
gallbladder, pancreas, spleen, kidneys, adrenals, urinary bladder,
prostate or uterus, testis or ovary, brain and pituitary.

Survival rates were comparable among treated and control
animals of both sexes. No treatment-related clinical signs of
toxicity were noted during the study. Body weight gain was decreased
in high-dose males during the second year of the study (no
statistical analysis). From the graphical representation of the
data, it appears that the body weights in the high-dose group were
approximately 10% below that of controls during the last nine months
of the study. In females, average body weights in treated groups
were consistently lower than the average control body weight for
most of the study period, however, the differences among the three
groups were very slight. No tumours or non-neoplastic lesions
attributable to treatment were observed (NCI, 1977).

2.2.3.2 Rats

Rats were fed for 44 to 52 weeks on a diet containing 0.5%
disodium EDTA without any deleterious effect on weight gain,
appetite, activity and appearance (Krum, 1948).

In another experiment three groups of 10 to 13 males and
females were fed a low-mineral diet (0.5% Ca and 0.013% Fe) with the
addition of 0, 0.5 and 1% disodium EDTA for 205 days. At the 1%
level some abnormal systems were observed: growth retardation of the
males, lowered erythrocyte and leucocyte counts, a prolonged blood
coagulation time, slightly but significantly raised blood calcium
level, a significantly lower ash content of the bone, considerable
erosion of the molars and diarrhoea. Gross and histological
examination of the major organs revealed nothing abnormal. Rats fed

for 220 days on an adequate mineral diet containing 1% disodium EDTA
showed no evidence of dental erosion (Chan, 1964).

In a two-year study, five groups of 33 rats each were fed 0,
0.5, 1 and 5% disodium EDTA. The 5% group showed diarrhoea and
consumed less food than the rats in other groups. No significant
effects on weight gain were noted nor were blood coagulation time,
red blood cell counts or bone ash adversely affected. The mortality
of the animals could not be correlated with the level of disodium
EDTA. The highest mortality rate occurred in the control group.
Gross and microscopic examination of various organs revealed no
significant differences between the groups (Yang, 1964).

Four groups of 25 male and 25 female rats were fed diets
containing 0, 50, 125 and 250 mg/kg bw/dy CaNa₂EDTA for two years.
Feeding was carried on through four successive generations. Rat
were mated after 12 weeks' feeding and allowed to lactate for three
weeks with one week's rest before producing a second litter. Ten
male and 10 female rats of each group (F₁ generation) and similar
F₂ and F₃ generation groups were allowed to produce two litters.
Of the second litters of F₁, F₂, and F₃ generations only the
control and the 250 mg/kg bw/dy groups were kept until the end of
two-years' study on the F₀ generation. This scheme permitted
terminal observation to be made on rats receiving test diets for 0,
0.5, 1, 1.5 or 2 years in the F₃, F₂, F₁ and F₀ generations,
respectively. No significant abnormalities in appearance and
behaviour were noted during the 12 weeks of the post weaning period
in all generations. The feeding experiment showed no statistically
significant differences in weight gain, food efficiency,
haematopoiesis, blood sugar, non-protein nitrogen, serum calcium,
urine, organ weights and histopathology of liver, kidney, spleen,
heart, adrenals, thyroid and gonads. Fertility, lactation and
weaning were not adversely affected for each mating. Mortality and
tumour incidence were unrelated to dosage level. The prothrombin
time was normal. There was no evidence of any chelate effect on
calcification of bone and teeth. Liver xanthine oxidase and blood
carbonic anhydrase activities were unchanged (Oser et al., 1963).

Groups of 50 male and 50 female Fisher F344 rats received
trisodium EDTA (Na₃EDTA) in the diet at concentrations of 3 750 or
7 500 ppm for 103 weeks, followed by one week during which standard
diet without EDTA was fed (NCI, 1977). A control group consisting of
20 rats of each sex received the standard diet of Wayne Lab Blox
Meal. Food was available ad libitum and fresh food was provided
three times per week.

the following tissues and organs: skin, lymph nodes, mammary gland,
salivary gland, bone marrow, trachea, lungs and bronchi, heart,
thyroid, parathyroids, oesophagus, stomach, small intestine, liver,
gallbladder, pancreas, spleen, kidneys, adrenals, urinary bladder,
prostate or uterus, testis or ovary, brain and pituitary.

Survival was comparable among control and treated groups of
male rats. There was a significant dose-related increase in survival
in treated groups of females compared to controls. Body weights were
comparable among treated and control groups, and there were no
clinical signs of toxicity in treated animals. No tumours or
non-neoplastic lesions attributable to treatment were observed (NCI,
1977).

2.2.4 Reproduction studies

2.2.4.1 Rats

Groups of six rats were maintained for 12 weeks on diets
containing 0.5, 1 and 5% disodium EDTA. No deaths occurred and
there were no toxic symptoms except diarrhoea and lowered food
consumption at the 5% level. Mating in each group was carried out
when the animals were 100 days old. Mating was repeated 10 days
after weaning the first litters. Parent generation rats of 0, 0.5
and 1% levels gave birth to normal first and second litters. The
animals given 5% failed to produce litters (Yang, 1964).

To elucidate possible teratogenic effects, daily doses of
20-40 mg EDTA/rat were injected i.m into pregnant rats at days six
to nine, 10 to 15 and 16 to the end of pregnancy. A dose of 40 mg
was lethal within four days but 20 mg was well tolerated, allowing
normal fetal development; 40 mg injected during days six to eight or
10 to 15 produced some dead or malformed fetuses, especially
polydactyly, double tail, generalized oedema or circumscribed head
oedema (Tuchmann-Duplessis and Mercier-Parot, 1956).

In a four generation study, groups of rats received CaNa₂EDTA
at doses of 50, 125 or 250 mg/kg/day via the diet. No reproductive
or teratogenic effects were observed in any of the three generations
of offspring (Oser et al., 1963). This study is discussed in
greater detail in Section 2.2.3 of this monograph.

Groups of pregnant Sprague-Dawley rats were fed Na₂EDTA in
standard diet at levels of 2 or 3% from day 1 to 21 of gestation.
Another group of pregnant rats received 3% Na₂EDTA in standard
diet from day 6 to 14 of gestation. A third group received 3%
Na₂EDTA and 1 000 ppm zinc in the diet from day 6 to 21 of
gestation. Controls received standard diet, which contained 100 ppm
zinc. The number of mated animals per group ranged from 5 to 16. on
day 21 of gestation fetuses were removed, fixed in Bouin's solution
and stored in 70% ethanol. Fetuses were examined under a dissecting

microscope for gross external abnormalities. Razor cut sections were
examined for abnormalities of the eye and head. In rats fed 2% EDTA
during pregnancy, litter size was normal and fetuses were alive.
Gross congenital malformations were apparent in 7% of the treated in
fetuses, compared to 0% in controls. In rats fed 3% EDTA during
pregnancy, almost half of the implantation sites had dead fetuses or
resorptions. Full term young were significantly smaller than
controls and 100% of them were malformed. Maternal toxicity as
manifested by diarrhoea was observed in rats fed 2 or 3% EDTA in the
diet. Malformations included severe brain malformations, cleft
palate, malformed digits, clubbed legs and malformed tails. The
detrimental effects of EDTA were prevented by supplementation of the
diet with 1 000 ppm zinc. These findings suggest that the
teratogenic effects observed in rats fed EDTA at very high levels in
the diet are due to zinc deficiency (Swenerton and Hurley, 1971).

Groups of pregnant CD rats were treated with Na₂EDTA via the
diet at a dose of 954 mg/kg/day (3% in the diet; 42 rats), by
gastric intubation at doses of 1 250 mg/kg/day (split dose of
625 mg/kg twice/day; 22 rats) or 1 500 mg/kg/day (split dose of
750 mg/kg twice/day; 8 rats), or by subcutaneous injection at a dose
of 375 mg/kg/day (25 rats). Animals were dosed on gestation day 7
through 14. The number of control animals for each exposure route
were: diet, 38; gavage, 20; subcutaneous injection, 14. Fetuses were
removed at day 21 of gestation. One third of the fetuses from each
litter (including all stunted fetuses and those with external
malformations) were dissected and examined for visceral
abnormalities. All fetuses surviving to the time of sacrifice were
fixed and examined for skeletal malformations. Maternal toxicity as
evidence by decreased food consumption, diarrhoea and diminished
weight gain was observed in groups treated by all three dose routes.
In the dietary exposure group, there were no maternal deaths, but
there was a significant increase in fetal death and 71% of the
fetuses were malformed. In the group administered 625 mg/kg/day by
gavage, only 64% of the dams survived treatment. In those surviving,
the number of fetal resorptions was similar to controls and 20.5% of
the fetuses were malformed. Seven out of eight of the dams
administered 750 mg/kg/day by gavage failed to survive. In the group
administered EDTA by subcutaneous injection, 76% of the dams
survived, the number of resorptions was significantly increased
above control levels and the proportion of malformed fetuses was
similar to controls. The types of malformations were consistent with
those observed by Swenerton and Hurley, although these former
workers only evaluated external malformations. The results of this
study indicate that the route of exposure to EDTA is an important
factor in determining its lethality and teratogenicity (Kimmel,
1977).

Groups of 20 pregnant CD rats were administered EDTA,
Na₂EDTA, Na₃EDTA, Na₄EDTA or CaNa₂EDTA by gavage at a total
dose of 1 000 mg EDTA/kg/day in two divided doses per day during
gestation day 7 through 14. All fetuses were subjected to gross
examination. One third were sliced and examined for visceral
abnormalities and the other two thirds were dissected, processed and
examined for skeletal abnormalities. The incidence of diarrhoea was
increased in all treated groups. Food intake was decreased in
treated groups as was weight gain during the treatment period.
Litter size and fetal mortality were unaffected by treatment in all
groups. No treatment-related teratogenic effects were observed in
any group (Schardein et al., 1981).

2.2.5 Special studies on embryotoxicity

2.2.5.1 Chickens

Disodium EDTA injected at levels of 3.4, 1.7 and 0.35 mg/egg
gave 40, 50 and 85% hatch, respectively. At the highest level, some
embryos which failed to hatch showed anomalies (McLaughlin and
Scott, 1964).

2.2.6 Special studies on genotoxicity

Na₃EDTA was tested for mutagenicity in the L5178Y tk⁺/tk^-
mouse lymphoma cell forward mutation assay. Two experiments were
conducted with S9, and three without S9, using EDTA concentrations
of up to 5 000 ug/ml. No mutagenicity was observed with or without
S9 (McGregor et al., 1988).

Na₃EDTA was tested for mutagenicity in Salmonella
typhimurium strains TA98, TA100, TA1535, TA1537 and TA1538 as well
as in Escherichia coli WP uvrA, in the presence and absence of
S9. Concentrations of up to 1 mg/plate were tested. No evidence of
mutagenicity was found in either of these bacterial systems, by four
independent laboratories (Dunkel et al., 1985).

2.2.7 Special studies on skin sensitization

Groups of 10 Hartley guinea-pigs received topical application
of Na₃EDTA, ethylene diamine (EDA) or epoxy resin (positive
control) four times over 10 days to a shaved and depilated area on
the back. Following a two week recovery period, animals received a
challenge on the clipped flank. Animals originally treated with EDTA
were not sensitized to EDTA. Animals originally treated with EDA
were sensitized to EDA, but not to EDTA. The results of this study
indicate a lack of sensitizing potential of EDTA and a lack of
cross-sensitization between EDA and EDTA (Henck et al. 1985).

2.3 Observations in humans

Three comprehensive field trials have been carried out using
NaFeEDTA as an iron fortificant in fish sauce (Garby and Areekul,
1974), off-white sugar (Viteri et al., 1983) curry powder (Ballot
et al., 1989b). The salient features of these trials are listed
in Table 5. All three trials were preceded by some estimate of the
iron status of the population and care was taken to establish the
acceptability and bioavailability of iron from the chosen vehicle
prior to the trials (Garby and Areekul, 1974, Viteri et.al, 1983,
Lamparelli et.al, 1987, Ballot et.al, 1989a). The choice of food
vehicle in each case reflected the dietary habits of the population.

2.3.1 NaFeEDTA fortified fish sauce

Fortified fish sauce was provided for a period of one year to
the population of a Thai village. The packed red cell volume (PCV)
values before and after the fortification program showed a
significant increase as compared to a control village supplied with
unfortified fish sauce. The biggest mean change (+4.7) was seen in
a sub-group of women who were anaemic at the start of the trial
(initial PCV < 33). Although a similar sub-group of women in the
control group also improved during the year (mean change +2.1) the
increase in PCV in the fortified group was significantly greater.
The same pattern was seen in both men and children with low initial
PCV values.

In terms of iron nutrition the increase of 4.7 PCV units over
initial values represents an increase of about 187 mg iron, in total
body iron or an increase in daily absorption of about 0.5 mg over
the duration of the trial (1 year). This is 64% of the expected
increase in body iron of 0.8 mg/day calculated on the basis of an
anticipated absorption of 8% and an assumed daily intake of 10 ml
fortified fish sauce (10 mg Fe). Iron stores were not measured in
this trial and the calculation does not take into account any
absorbed iron which may have been laid down in stores. The
calculated value therefore would be an underestimate of the total
amount of iron actually absorbed. Nevertheless it illustrates that
fortification with NaFeEDTA is a highly effective method for
improving iron status.

Overall this trial demonstrated that fortification of fish
sauce at modest levels using NaFeEDTA is feasible, and that it can
produce a significant improvement in iron status as assessed by a
single simple criterion (PCV) (Garby and Areekul, 1974).

2.3.2 NaFeEDTA fortified sugar

The design of this trial makes interpretation of the results
difficult. The analysis is based on the comparative changes in iron
status observed in four communities. Three (#13, #14, #16) were
test sites, and one was a control site (#15). The initial iron
status of individuals drawn from test community #14 was
significantly worse than that of individuals from the other test
communities and the control community (#15). Unfortunately
compliance was poor in this community and also in test community
#13. Furthermore seventy percent of the families in test
communities number 13 and 14 used fortified sugar for only half of
the time. The remaining 30% used it for 80% of the time. Finally,
subjects with severe anaemia were given therapeutic iron to improve
their iron status prior to the trial. Despite the presence of these
confounding factors, the haemoglobin values rose in both males and
females after 20 months of fortification, although the values did
not reach statistical significance. Only the children (5-12 years)
in communities #13 and #16 showed a significant improvement in
haemoglobin levels when compared to children in the control
community #15 (+2.2±1.7 and +2.2±1.5 respectively vs +1.6±1.2 g/dl).
The greater benefit observed in children may have resulted from the
fact that sugar consumption was greater in children than in adults
when considered relative to body weight. Mean serum ferritin which
is a measure of the size of iron stores increased in each of the
test communities, but not in the control community. In conclusion
it should be noted that the relatively modest improvement in iron
status noted in this trial may also have been due, in part, to the
fact that the fortification level was considerably less than in the
other two trials (4.3 vs 10-15 and 7.7 mg/person/day) (Viteri
et al., 1983).

Table 5. Outline of field trials using NaFeEDTA to fortify various
food vehicles

References Garby and Areekul, 1974 Viteri, et al., 1983 Ballot, et al., 1989a,b

Geographical region Thailand Central America South Africa

Population studied Two rural villages 4 rural Guatemalan Urban Indian community in a
communities municipal housing estate

Design of trial Controlled (one village) Controlled (community #15) Controlled (random allocation
not blinded not blinded by families) double-blinded

Sample studies Test village (284) control #13 - 186 #14 - 306 263 Families (672 subjects)
village (330) #15 - 234 #16 - 296 severe 129 control families 134 fortified
anaemics treated prior to families Hb < 9 g/dl excluded
trial

Food vehicle Fish-sauce (salt substitute) Off-white sugar distribution: Masala (curry powder) distributed
30 g NaCl/l, 10 mg Fe/l sold to store keepers. Purchased directly to families monthly free
distributed by village by participants (poor compliance of charge
head-man as required #13 and #14)

Cons. of food vehicle 10 - 15 ml/person/day 33 g/person/day; children 5.5 g/person/day
highest consumption

Fe absorption 8% 8% 10%

Level of fortification 1 mg Fe/ml 13mg Fe/100g 1.4 mg Fe/g
and intake 10 - 15 mg Fe/person/day 4.29 mg Fe/person/day 7.7 mg Fe/person/day

Acceptability No changes Barely perceptible yellowing Slight darkening of food

Table 5 (contd).

References Garby and Areekul, 1974 Viteri, et al., 1983 Ballot, et al., 1989a,b

Duration of trial 12 months 20 months 24 months

% Abnormal iron status 30 - 50 of population anaemic; Low Low Low Females Males
prior to trial 34% initial PCV below normal Comm PCV Sat
Ferr IDA 24 4
#13 31 34 52 ID 53 24
#14 43 58 72
#15 35 12 37
#16 21 23 34

Measurements taken Packed cell volume (PCV) Haemoglobin, PCV, %Sat, FEP, Haemoglobin, %Sat, Serum Ferritin
Serum Ferritin, Cu, Zn

IDA = Iron Deficiency anaemia; ID = Iron Deficiency; % Sat = % Saturation of Transferrin; FEP =
Free Erythrocyte Protoporphyrin; PCV = Packed Cell Volume; Comm = Community

2.3.3 NaFeEDTA fortified masala

The design of the most recent fortification trial differed from
those of earlier studies in that it was conducted in a single
community with families randomly assigned to control and test
groups. The groups were matched for iron status. It was also
double-blinded and care was taken to ensure that cross-over between
groups did not occur. Fortified or unfortified masala was
distributed directly to each family. In addition to evaluating
fortification the usual indices of improving iron status (increasing
haematocrit or haemoglobin and ferritin) in each individual by using
a composite of haemoglobin concentration, percent saturation of
transferrin, and serum ferritin concentration, an attempt was made
to estimate the total body iron (in mg) in each individual by using
a composite of haemoglobin concentration, percent saturation of
transferrin and serum ferritin concentration (Cook, et al, 1986).
This comprehensive index of iron nutrition made it possible to
compare subjects with wide variations in iron status and thus to
assess both the beneficial and potentially adverse effects of
additional iron i.e. development of iron overload (Ballot, et al.,
1989a, b).

Significant improvement in body iron as assessed by the index
was detectable in the group of women receiving fortified masala
after one year of the program (Ballot, et al., 1989a, b). This
improvement continued during the second year when the rise in
haemoglobin concentration became significantly greater than in the
control group. The prevalence of iron deficiency dropped
dramatically in the women receiving fortified masala. Iron
deficiency anaemia was detected in 22% of individuals at the start
of the study, but only to 4.9% after two years. The most
significant improvement in iron status was noted in women who
entered the trial with iron deficiency (especially in those with
anaemia). They showed an increase in calculated body iron of 505 mg
which is equivalent to the absorption of an additional 0.7 mg
iron/day. The latter figure is close to the predicted improvement
in iron balance of 0.8 mg daily based on isotopic absorption studies
using NaFeEDTA fortified masala (Lamparelli et al., 1987).

In iron-replete males the rise in calculated body iron was
modest and only reached significance in alcohol abusers receiving
fortified masala. This suggests that iron-replete males are unlikely
to accumulate excessive amounts of iron under these fortification
conditions.

3. COMMENTS

The Committee was concerned about over-fortification or misuse
of this product and did not recommend its availability for general
use by individuals.

The Committee noted that sodium iron EDTA dissociates in the
intestine, and iron in this form is approximately twice as
bioavailable as iron in the form of iron sulfate. The available
studies indicated that only a fraction, if any, of the iron EDTA
chelate is absorbed as such, that EDTA from sodium iron EDTA is only
poorly absorbed and that the majority is excreted in the faeces.
The portion that is absorbed (<5%) is rapidly excreted in urine.
The proposed supplementation programme would result in intakes of
iron and EDTA of approximately 0.2 and 1.34 mg/kg bw/day,
respectively.

4. EVALUATION

Based on previous evaluations of both iron and EDTA and the
available bioavailability and metabolism data, the Committee
provisionally concluded that use of sodium iron EDTA meeting the
tentative specifications prepared at the present meeting in
supervised food fortification programmes in iron-deficient
populations does not present a safety problem. The Committee
requested that additional studies be conducted to assess the site of
deposition of iron administered in this form and further studies to
assess the metabolic fate of sodium iron EDTA following long-term
administration.

The Committee emphasized that its evaluation pertains only to
the use of sodium iron EDTA as a dietary supplement to be used under
supervision and expressed its concern about the potential for over-
fortification because of the enhanced bioavailability of iron in
this form.

The Committee developed new tentative specifications for sodium
iron (III) ethylenediaminetetraacetate (NaFeEDTA). In preparing the
specifications, the Committee was aware that food-grade NaFeEDTA is
not commercially available. However, the Committee was advised,
that this substance is being evaluated for its usefulness in
fortifying the diet in areas of the world where iron deficiency in
the population is endemic, prepared specifications, which it
believed would assist in the evaluation. The Committee obtained
analytical data and other information on fertilizer-grade NaFeEDTA,
which is widely available, and considered this together with the
existing specifications for disodium ethylenediaminetetra-acetate
and calcium disodium ethylenediaminetetraacetate in formulating the
specifications. Because further information on assay and purity
data for food-grade material and on analytical methodology is still
needed, the specification was designated as tentative.

5. REFERENCES

AABAKKEN, L. & OSNES, M. (1990). ⁵¹Cr-ethylenediaminetetraacetic
acid absorption test. Effects of Naproxen, a non-steroidal,
antiinflammatory drug. Scan. J. Gastroenterol. 25, 917-24.

ANGHILERI, L.J. (1967). Fate of intravenously injected iron
compounds:ferric-fructose complex, iron-EDTA, ferric hydroxide, and
iron-albumin labeled with ⁵⁹Fe. Biochemical Pharmacology 16,
2033-36, 1967.

BALLOT, D.E., MACPHAIL, A.P., BOTHWELL, T.H., GILLOOLY, M., & MAYET,
F.G. (1989a). Fortification of curry powder with NaFe(III)EDTA in an
iron deficient population: Initial survey of iron status. Am. J.
Clin. Nutr. 49, 156-61.

BALLOT, D.E., MACPHAIL, A. P., BOTHWELL, T.H., GILLOOLY, M. & MAYET,
F.G. (1989b). Fortification of curry powder with NaFe(III)EDTA in an
iron-deficient population: Report of a controlled iron-fortification
trial. Am. J. Clin. Nutr. 49, 162-69.

BATES, G.W., BILLUPS C. & SALTMAN, P. (1967). The kinetics mechanism
of iron (III) exchange between chelates and transferrin. J. Biol.
Chem. 242, 2816-21.

BAUER, R.O., Rullo, F., Spooner, C., and Woodman, E. (1952) Acute
and subacute toxicity of ethylene diamine tetraacetic acid (EDTA)
salts Fed. Proc. 11, 321.

BJARNASON, I., PETERS, T.J. & VEALL, N. (1983). A persistent defect
in intestinal permeability in coeliac disease demonstrated by a
⁵¹Cr-labelled EDTA absorption test. Lancet 1, 323-325.

BOTHWELL, T.H., CHARLTON, R.W., COOK, J.D. & FINCH, C.A. (1979).
Iron Metabolism in Man. Blackwell Scientific Publications:
Oxford, England.

CANDELA, E., CAMACHO, M.V. MARTINEZ-TORRES, C., PERDOMO, J.,
MAZZARRI, G., ACURERO, G. & LAYRISSE, M. (1984). Iron absorption by
humans and swine from Fe(III)-EDTA. Further studies. J. Nutr. 114,
2204-11.

CATSCH, A. (1961). Fed. Proc. 20 (Suppl. 10), 206.

CHABERCK, S.A. & MARTELL, A. E. (1959). Organic sequestering
agents. John Wiley and Sons, Inc., New York.

CHAN, M.S. (1964). Fd. Cosmet. Toxicol. 2. 763-765.

CLETON, F., TURNBULL, A. & FINCH, C.A. (1963). Synthetic chelating
agents in iron metabolism J. Clin. Invest. 42, 327.

CONRAD, M.E. & SCHADE, S.G. (1968). Ascorbic acid chelates in iron
absorption: a role for hydrochloric acid and bile. Gastroenterology
55, 35-45.

COOK, J.D. & MONSEN E.R. (1976). Food iron absorption in man II. The
effect of EDTA on the absorption of non-heme iron. Am. J. Clin.
Nutr. 29, 614-20.

COOK, J.D. & REUSSER. (1983). Iron fortification: an update. Am. J.
Clin. Nutr. 38, 648-59.

COOK, J.D., LAYRISSE, M., MARTINEZ-TORRES, C., WALKER, R.B., MONSEN
E. & FINCH, C.A. (1972). Food iron absorption measured by an
extrinsic tag. J. Clin. Invest. 51, 805-815.

COOK, J.D., SKIKNE, B.S., LYNCH S.R., & REUSSER, M.E. (1986).
Estimates of iron sufficiency in the U.S. population. Blood 68,
726-31.

DAVIDSSON, L., KASTENMEYER, P., AND HURRELL, R. (1993) Sodium iron
EDTA (NaFeIIIEDTA) as a food fortificant: the effect n the
absorption and retention of calcium and zinc for humans. Am. J.
Clin. Nutr., submitted for publication.

DUNKEL, V.C., ZEIGER, E. BRUSICK, D. M^cCOY, E. M^cGREGOR, D.,
MORTLEMANS, K. ROSENKRANZ H.S., & SIMMON V.F. (1985).
Reproducibility of microbial mutagenicity assays:II. Testing of
carcinogens and noncarcinogens in Salmonella typhimurium and
Escherichia coli. Environ. Mutagen. 7(Supp.5:1), 1-248.

DYCKERHOFF, H., MARX., R. & LUDWIG, B. (1942). Z. ges. exp. Med.
110, 412.

el GUINDI, M., LYNCH, S.R. & COOK, J.D. (1988). Iron absorption from
fortified flat breads. Br. J. Nutr. 59, 205-13.

FORBES, R.M. (1961). Excretory patterns and bone deposition of zinc,
calcium and magnesium in the rat as influenced by zinc deficiency,
EDTA and lactose. J. Nutr. 74, 193-200.

FOREMAN, H. (1961). Fed. Proc. 20, (Suppl. 10), 191.

FOREMAN, H., & TRUJILLO, T.T. (1954). The metabolism of C¹⁴ labeled
ethylenediaminetetraacetic acid in human beings. J. Lab. Clin. Med.
43, 566-71.

FOREMAN, H., VIER, M., & MAGEE, M. (1953). The metabolism of C¹⁴
labeled ethylenediaminetetraacetic acid in the rat. J. Biol. Chem.
203, 1045-53.

GARBY, L. & AREEKUL, S. (1974). Iron supplementation in Thai
fish-sauce. Ann. Trop. Med. Parasitol. 68(4), 467-76.

GIBSON, K.D., NEUBERGER, A. & SCOTT, J.C. (1955). The purification
and properties of delta-aminolaevulic acid dehydrase. Biochem. J.
61, 618-629.

GOULD, R.G. (1961). Fed. Proc. 20 (Suppl. 10), 252.

HALLBERG, L. & BJÖRN-RASMUSSEN, E. (1972). Determination of iron
absorption from whole diet. A new two-pool model using two radioiron
isotopes given as haem and non-haem iron. Scand. J. Haematol. 9,
193-97.

HALLBERG, L., BJÖRN-RASMUSSEN, E., GARBY, L. PLEEHACHINDA, R. &
SUWANIK, R. (1978). Iron absorption from South-East Asian diets and
the effect of iron fortification. Am. J. Clin. Nutr. 31, 1403-08.

HART, J.R. (1984) EDTA-type chelating agents in everyday consumer
products: some medicinal and personal care products. Journal of
chemical education 61(12).

HAWKINS, W.W. et al. (1962). Can. J. Biochem. 40, 391.

HENCK, J.W., LOCKWOOD, D.D. & OLSON, K.J. (1985). Skin sensitization
potential of trisodium ethylenediaminetetraacetate. Drug Chem.
Toxicol. 3, 99-103.

HURRELL, R.F., RIBAS, R. & DAVIDSSON, L. (1993). Sodium iron EDTA as
a food fortificant: Influence on zinc, calcium and copper metabolism
in the rat. Brit. J. Nutr. in press.

KIMMEL, C.A. (1977). Effect of route of administration on the
toxicity and teratogenicity of EDTA in the rat. Tox. Appl.
Pharmacol. 40, 299-306.

KRATZER, F.H., ALLRED, J.B., DAVIS, P.N., MARSHAL, B.J., & VOHRA, P.
(1959) The effect of autoclaving soybean protein and the addition of
ethylenediaminetetraacetic acid on biological availability of
dietary zinc for turkey poults. J. Nutr. 68, 313-22.

KRUM, J.K. (1948). Thesis, University of Massachusetts.

LAMPARELLI, R.D., MACPHAIL, A.P. BOTHWELL, T.H. BALLOT, D.E.
DANILEWITZ, M.D. MACFARLANE, B.J., MAYET F. & BAYNES, R.D. (1987).
Curry powder as a vehicle for iron fortification: effects on iron
absorption. Am. J. Clin. Nutr. 46, 335-40.

LAPINLEIMU, K. & WEGELIUS, R. (1959). Antibiotic Med. Clin. Ther.
(Br. Edit.) 6, 151.

LARSEN, B.A., BIDWELL, R.G.S. & HAWKINS, W.W. (1960). Can. J.
Biochem. 38, 51.

LAYRISSE, M. & MARTINEZ-TORRES, C. (1977). Fe(III)EDTA complex as
iron fortification. Am. J. Clin. Nutr. 30, 1166-74.

LECHNIT, V. (1961). Arch. Gewerbepath. Gewerbehyg. 18, 495.

LELIœVRE, P. (1960). C.R. Soc. Biol. (Paris) 154, 1890.

LELIœVRE, P. & BETZ, E.H. (1961). C.R. Soc. Biol. (Paris) 155,
199.

MACPHAIL, A.P. & T.H. BOTHWELL. (1992). The prevalence and causes
of nutritional iron deficiency anemia. In: Nutritional anemias
Eds. Fomon, S.J. and Zlotkin. S. Nestle Nutrition Workshop Series
30: 1-12, Nestec Ltd., Vevey, Switzerland / Raven Press, New York.

MACPHAIL, A.P., BOTHWELL, T.H., TORRANCE, J.D., DEOMAN, D.P.,
BEZWODA, W.R. & CHARLTON, R.W. (1981). Factors affecting the
absorption of iron from Fe(III)EDTA. Br. J. Nutr. 45: 215-227.

MARQUARDT, P. & SCHUMACHER, H. (1957). Arzneimittelforsch, 7, 5.

MARTELL, A.E. (1960). Chelation: Stability and selectivity. Ann.
New York Academy of Sciences 88, 284-92.

MARTINEZ-TORRES, C., ROMANO, E.L., & LAYRISSE M. (1979). Fe(III)EDTA
complex as iron fortification. Further studies. Am. Clin. Nutr.
32, 809-16.

M^cGREGOR, D.B., BROWN, A., CATTANACH, P., ET. AL. (1988).
Responses of the L5178Y tk⁺/tk mouse lymphoma cell forward
mutation assay: III. 72 coded chemicals. Environ. Mol. Mutagen.
12, 85-154.

M^cLAUGHLIN, J., Jr. & SCOTT, W.F. (1964). Toxicity of some food
packaging chemicals measured by the chick embryo technique. Fed.
Proc. 23, 406.

MIETTINEN, T.A. & LESKINEN, E. (1962). Ans. Med. Exp. Fenn., 40,
427.

NAGARAJAN, B., SIVARAMAKRISHNAN, V.M., & BRAHMANANDAM, S. (1964).
The distribution of ⁵⁹Fe in albino rats after intravenous
administration in the ionic or chelated form. Biochemical Journal
92, 531-7.

NCI. (1977). Bioassay of trisodium ethylenediaminetetraacetate
trihydrate (EDTA) for possible carcinogenicity. National Cancer
Institute. Carcinogenesis Tech. Report Series No. 11, 1977.

NIGROVIC, V. (1964). Arch. Exp. Pathol. Pharmacol. 249, 206.

OBERLEAS D., MUHRER, M.E. & O'DELL, B.L. (1966). Dietary metal
complexing agents and zinc availability in the rat. J. Nutr. 90,
56-62.

OSER, B.L., OSER, M. & SPENCER, H.C. (1963). Safety evaluation
studies of calcium EDTA. Tox. Appl. Pharmacol. 5, 142-62.

PERRY, H.M. & PERRY, E.F. (1959). Normal concentrations of some
trace metals in human urine: changes produced by EDTA. J. Clin.
Invest. 38, 1452-63.

PLUMB, R.C., MARTELL, A.E. & BERSWORTH, F.C. (1950).
Spectrophotometric determination of displacement series of metal
complexes of the sodium salts of ethylenediamine-tetraacetic acid.
J. Phys. Colloid. Chem. 54, 1208-15.

POGELL, B.M. & LELOIR, L.F. (1961). Nucleotide activation of liver
microsomal glucuronidation. J. Biol. Chem. 236, 293-8.

REUBER, M.D. & SCHMIELLER, G.C. (1962). Edetate lesions in rats.
Arch. Environ. Health 5, 430-36.

RUBIN, M. (1961). Fed. Proc. 20 (Suppl 10.), 149.

RUBIN, M., PACHTMAN, E. ALDRIDGE, M. ZAPOLSKI, E.J. BAGLEY, Jr.,
D.H., & PRINCIOTTO, J.V. (1970). The metabolism of parenteral iron
chelates. Biochemical Medicine 3: 271-88.

SCHANKER, L.S. & JOHNSON, J.M. (1961). Biochem. Pharmacol. 8,
421.

SCHARDEIN, J.L., SAKOWSKI, R., PETRERE, J., & HUMPHREY, R.R. (1981).
Teratogenesis studies with EDTA and its salts in rats. Tox. Appl.
Pharmacol. 61, 423-28.

SCOTT, M.L. & ZIEGLER, T.R. (1963). Evidence for natural chelates
which aid in the utilization of zinc by chicks. Agr. Fd. Chem. 11,
123-25.

SHIBATA, S. (1956). Folio Pharmacol. Jap. 52, 113.

SIMPSON, R.J. & PETERS, T.J. (1984) Studies of Fe⁺⁺⁺ transport
across intestinal brush border membrane of the mouse. Biochemica
et Biophysica Acta 772, 220-26.

SIMMONS, W.K. & GURNEY, J.M. (1980). Nutritional anemia in the
English-speaking Caribbean and Surinam. Kingston: Caribbean Food
and Nutrition Institute J-46-80. Pan American Health Organization.

SOLOMONS, N.W., JACOB, R.A. PINEDA, O. & VITERI, F.E. (1979).
Studies on the bioavailability of zinc in man. Effect of the
Guatemalan rural diet and the iron-fortifying agent, NaFeEDTA.
J. Nutr. 109, 1519-28.

SRBROVA, J. & TEISINGER, J. (1957). Arch. Gewerbepathol. 15, 572.

SWENERTON, H. & HURLEY, L.S. (1971). Teratogenic effects of a
chelating agent and their prevention by zinc. Science 173, 62-64.

TUCHMANN-DUPLESSIS, H. & MERCIER-PAROT, L. (1956). C.R. Acad. Sci.
243, 1064.

VITERI, F.E., GARCIA-IBANEZ, R. & TORUN, B. (1978). Sodium iron
NaFeEDTA as an iron fortification compound in Central America.
Absorption studies. Am. J. Clin. Nutr. 32, 961-71.

VITERI, F.E., ALVARES, E. & TORUN, B. (1983). Prevention of iron
deficiency by means of iron fortification of sugar. In: Underwood,
B.A., ED. Nutrition intervention strategies in national
development. New York, Academic Press, 287-314.

VOHRA, P. & KRATZER, F.H. (1964). Influence of various chelating
agents on the bioavailability of zinc. J. Nutr. 82, 249-56.

WEST, T.S. & SYKES, A.S. (1960). Diamino-ethane-tetra-acetic acid.
In: Analytical applications of Diamino-ethane-tetra-acetic acid.
Poole, England: The British Drug Houses, Ltd. 9-22.

WESTERFELD, W.W. (1961). Fed. Proc. 20, (Suppl 10.), 158.

WILL, J.J. & VILTER, R.W. (1954). A study of the absorption and
utilization of an iron chelate in iron-deficient patients. J. Lab.
& Clin. Med. 44:499-505.

YANG, S.S. (1964). Toxicology of EDTA. Fd. Cosmet. Toxicol. 2,
763-67.

See Also: Toxicological Abbreviations