Following a request from the European Commission, the Panel on Food Additives and Nutrient Sources added to Food (ANS) was asked to deliver a scientific opinion re-evaluating the safety of iron oxides and hydroxides when used as food additives.
Iron oxides and hydroxides are a group of inorganic pigments collectively allowed for use as food additives (E 172) in the European Union (EU) and previously evaluated by the EU Scientific Committee for Food (SCF) in 1975 and the Joint FAO/WHO Expert Committee of Food Additives (JECFA) in 1974, 1975, 1978, 1980 and 2000 (JECFA, 1974, 1975, 1978, 1980, 2000).
In the European Commission (EC) specifications for iron oxides and hydroxides (E 172) (Commission Regulation (EU) No 231/2012), there are three different oxides: yellow iron oxide (FeO(OH)·H2O), red iron oxide (Fe2O3) and black iron oxide (FeO·Fe2O3). Brown Iron Oxide has been included in this assessment, for completeness due to its importance as a commercial blend: its colour shades are obtained by mixing different amounts of the aforesaid powdered principles. The Panel considered that only material with brown shades obtained by blending of the iron oxides and hydroxides evaluated in this Opinion would be covered by the present assessment.
As these iron oxides and hydroxides have different physical and chemical properties and they can be used separately, the Panel recommended that a clear differentiation (e.g. by adding a, b, c to the E number) be made between the different iron oxides and hydroxides that are currently all included under E 172.
According to the data previously submitted by industry (Rockwood, 2013a), the average particle sizes of iron oxide particles were 1 677, 318 and 957 nm for yellow iron oxide (FeO(OH)), red iron oxide (Fe2O3) and black iron oxide (Fe3O4), respectively. The Panel noted that the method used by industry for measuring the particle size of iron oxides (Rockwood, 2013a) cannot exclude the presence of particles with one or more dimensions below 100 nm.
More recently, transmission electron microscopy (TEM) analyses were carried out on few E 172 products (Huntsman, 2015). Particle size distributions were found to vary in relation to the chemistry of the product so that the distributions of primary particle sizes changed from FeO(OH) to Fe2O3 to Fe3O4. In all cases, particles that showed at least one dimension in the nanosize range were detected. The Panel had previously noted that, according to the European Food Safety Authority (EFSA) Guidance document, two different methods should be used to examine the particle size distribution (EFSA Scientific Committee, 2011).
In general, the Panel noted that the manufacturing process of powdered or particulate food additives results in material with a range of sizes. While the mean or median size of the particles is generally significantly greater than 100 nm, a small fraction will always be, and has been, with at least one dimension below 100 nm. The material used for toxicological testing would have contained this nano fraction. The test requirements stipulated in current EFSA guidance documents and EC guidelines for the intended use in the food/feed area apply in principle to unintended nano forms as well as to engineered nano material (ENM).
Therefore, the Panel considered that, in principle, for a specific food additive containing a fraction of particles with at least one dimension below 100 nm, adequately conducted toxicity tests should be able to detect hazards associated with this food additive including its nanoparticulate fraction. The Panel considered that for the re-evaluation of food additives this procedure would be sufficient for evaluating constituent nanoform fraction in accordance with the recommendation of the EFSA Nano Network in 2014.
In 1974, JECFA allocated a ‘temporary acceptable daily intake (ADI) not specified’ to iron oxides and hydrated iron oxides due to the lack of information on physiological absorption and iron storage following the use of iron oxides as food pigments. At the 1978 JECFA meeting, this temporary ADI was extended until 1979. In 1980, an ADI of 0-0.5 mg/kg bw/day was established (JECFA, 1980).
The available data indicate that absorption of iron from iron oxides is low. In rats, 0.01–2.3 % of the total oral dose of microsized red iron oxide (Fe2O3) was absorbed and distributed in different organs or excreted in urine. Low absorption of iron (0.01 %) from red iron oxide was observed in humans receiving a diet containing red iron oxide, whereas a higher absorption of yellow iron oxide (1.5–2.4 %) was described in similar populations. In these human studies, the addition of ascorbic acid increased by 5–50 times the iron absorption rates from diets containing either red iron oxide (Fe2O3) or yellow iron oxide (FeO(OH)). The Panel noted that there are no data regarding the biological fate of microparticles of black iron oxide (FeO·Fe2O3).
Concerning toxicological studies, the Panel noted that there is a lack of information on the presence of nanoparticles in iron oxides used in most of the old studies. Regarding acute toxicity, the available data indicate that iron oxides and hydroxides are of low toxicity in rats and mice.
The subacute oral toxicities of nano red iron oxide (Fe2O3-30 nm) and microsized red iron oxide (Fe2O3-Bulk) were compared in rats given 0, 30, 300 or 1 000 mg/kg bw/day for 28 days (Kumari et al., 2012). No decrease in body weight, no change in feed intake, nor any adverse symptoms or mortality were observed in rats exposed to microsized red iron oxide or to 30 or 300 mg/kg bw/day of red iron oxide nanoparticles. However, rats treated with the high dose of nano red iron oxide (1 000 mg/kg bw/day) showed reduced body weight and feed intake, severe toxic symptoms and several disturbances in biochemical parameters and adverse histopathological changes in the liver, kidney and spleen. By contrast, microsized red iron oxide did not induce any significant adverse effects in either biochemical parameters or histopathology in rats given the highest dose. This study indicated that the microsized particles, i.e. bulk material, are less potent than the nanoparticles in causing toxicity in the exposed animals. From this study, the Panel identified a no-observed-adverse-effect level (NOAEL) for microsized red iron oxide of 1 000 mg/kg bw/day, the highest dose tested. No subacute toxicity studies on yellow (FeO(OH)·H2O) and black (FeO·Fe2O3) iron oxides were available.
No subchronic toxicity studies by oral administration of microsized yellow iron oxide (FeO(OH)), red iron oxide (Fe2O3) or black iron oxide (FeO·Fe2O3) were available. A subchronic toxicity study of various orally administered nanoparticles including red iron oxide (Fe2O3 , 60-118 nm) was performed by Yun et al. (2015) according to the OECD Test Guideline (TG) 408 (OECD, 1998). Sprague-Dawley rats received daily doses of 250, 500 or 1000 mg/kg bw/day for 13 weeks by gavage. There were no treatment-related changes in haematological, serum biochemical parameters or histopathological lesions. In blood and all tissues tested including liver, kidney, spleen, lung and brain, the concentration of Fe showed no dose-associated response in comparison to the control groups. The authors stated that the subchronic oral dosing with Fe2O3 nanoparticles showed no systemic toxicity to rats. The Panel agreed with this statement and identified a NOAEL for nanosized red iron oxide of 1000 mg/kg bw/day, the highest dose tested in rats receiving Fe2O3 nanoparticles by gavage. Owing to the presence of nanoparticles in red iron oxide used as food additive, the Panel considered this study as relevant for the assessment of the safety of red iron oxide.
The Panel noted that, using similar range of daily doses, adverse effects were observed in rats subacutely treated (28 days) with red iron oxide nanoparticles, while no effect was described after a subchronic administration (90 days) of such particles to rats. The Panel considered that this difference could be explained by the use of smaller nanoparticles (30 nm) in the subacute study than those used in the subchronic toxicity study (60-118 nm). The former could be more efficiently available to organs and tissues leading to more severe adverse effects.
Red (Fe2O3) and black (FeO·Fe2O3) iron oxides, both in nano- and microform (7–30 nm and >100 nm, respectively), were positive in in vitro genotoxicity assays in mammalian cells, where induction of DNA strand breaks and micronuclei was observed. In vivo oral administration of both nano- and microsized red iron oxides did not elicit genotoxic effects in rat haemopoietic system, while no data are available for the site of contact (gastrointestinal tract). No in vivo genotoxicity studies have been performed on black iron oxide and no genotoxicity studies are available for yellow iron oxide. Due to the limitations of the database, and considering the impossibility to read-across between iron oxides with different redox state, the Panel considered that the genotoxicity of iron oxides cannot be evaluated based on the available data.
Concerning long-term toxicity and carcinogenicity, no adverse effects were reported in ten dogs fed from 1 to 9 years on diets containing iron oxide colourant (unspecified compound); the daily consumption was estimated at 428 mg/dog (unpublished study from Carnation Co., 1967, as reported by JECFA, 1983). In a study from Ralston Purina Cat Care Center (1968), no adverse effects were reported in cats maintained on diets containing 1 900 mg/kg diet (475 mg/kg bw/day) of iron from iron oxide (equivalent to 0.27 % iron oxide) for periods of 2–9 years. The International Agency for Research on Cancer (IARC) Monograph (1987) stated that there was evidence suggesting lack of carcinogenicity of haematite (red iron oxide) and ferric oxide (unspecified compound) to animals, and that there was inadequate evidence of carcinogenicity in humans.
Concerning reproductive and developmental toxicity, no signs of toxicity were observed in an unpublished study (as reported in JECFA, 1983). However, this study was not available and could not be evaluated by the Panel.
The Panel noted that only 10 out of the 49 food categories in which iron oxides and hydroxides (E 172) are authorised could be taken into account in the present exposure estimates and therefore that overall this would result in an underestimation of the actual exposure to iron oxides and hydroxides (E 172) used as food additives in European countries. The Panel noted that due to limited information becoming available on the type of iron oxides and hydroxides (yellow, red or black) used in the authorised food categories, the exposure estimates for E 172 were based on maximum levels/reported use levels irrespectively of the type of iron oxide.
Using the “maximum level exposure assessment scenario”, mean exposure to iron oxides and hydroxides (E 172) from its use as a food additive ranged from 0.1 mg/kg bw/day for infants to 10.5 mg/kg bw/day for toddlers, while the high exposure using this scenario ranged from 0.2 mg/kg bw/day for infants to 26.9 mg/kg bw/day for toddlers.
Using the refined brand-loyal assessment exposure scenario, mean exposure to iron oxides and hydroxides (E 172) from its use as a food additive ranged from 0.1 mg/kg bw/day for infants to 8.9 mg/kg bw/day for toddlers. The high exposure to iron oxides and hydroxides (E 172) using this scenario ranged from 0.2 mg/kg bw/day for infants to 23.1 mg/kg bw/day in toddlers.
Using the refined non-brand-loyal assessment exposure scenario, mean exposure to iron oxides and hydroxides (E 172) from its use as a food additive ranged from 0.03 mg/kg bw/day for infants to 3.7 mg/kg bw/day for toddlers. The high exposure to iron oxides and hydroxides (E 172) from its use as food additive using this scenario ranged from 0.1 mg/kg bw/day for infants to 9.5 mg/kg bw/day for toddlers. Overall, the lowest exposure to iron oxides and hydroxides (E 172) was estimated for infants, while the highest exposure was calculated for toddlers, in all scenarios. The food categories that, at the individual level, had the highest contribution to the total individual exposure to iron oxides and hydroxides (E 172) were fine bakery wares.
In view of assessing the safety of iron oxides and hydroxides, the Panel noted that:
- the particle size distribution of these substances includes particles with one or more dimensions below 100 nm,
- physical-chemical characteristics of the particulate material (redox states, particle size) between black (which contains iron(II) and iron(III)) and red and yellow (which contain iron(III)) iron oxides could be critical toxicological features,
- the toxicological database on yellow and black iron oxides is very limited,
- genotoxicity data on yellow iron oxide are absent,
- in vivo genotoxicity data on black iron oxide are absent.
- in vivo genotoxicity data on red iron oxide at the site of contact are absent,
The Panel further considered that read-across from red iron oxide to black iron oxide should not be performed due to differences in their redox states.
In the absence of data on the genotoxicity of yellow iron oxide (FeO(OH)), the Panel noted that read-across from red iron oxide should not be performed due to marked differences in the shape and size distribution of yellow iron oxide showing a larger fraction of nanosized particles.
Regarding Brown Iron Oxide, the E 172 brown shade is mentioned in Commission Regulation (EU) No 231/2012, although the blend itself is nominally not listed, nor further characterised. The Panel noted that specifications and a reliable toxicological database on yellow, red and black iron oxides are needed in order to assess its safety when used as a food additive.
The Panel concluded that an adequate assessment of the safety of E 172 could not be carried out because a sufficient biological and toxicological database was not available.
The Panel noted that for the food additive iron oxides and hydroxides (E 172), the term ‘iron oxides’ applies sometimes either to iron oxides or iron hydroxides and therefore grouping them together under a single E number is confusing. As these compounds have different physical and chemical properties and they can be used separately, the Panel recommended that a clear differentiation (e.g. by adding a, b, c to the E number) should be made between the different iron oxides and hydroxides that are currently all included under E 172. Furthermore, the Panel noted that concentration data on yellow iron oxide, red iron oxide and black iron oxide alone would be needed for the calculation of exposure estimates for each of the three single iron oxides.
Because of the potential importance of nanoparticles in toxicokinetics and toxicological effects, the Panel considered that the particle size and particle size distribution should be included in the specifications of iron oxides and hydroxides.
The Panel considered that the maximum limits for certain toxic elements (cadmium, arsenic, lead and mercury) present as impurities in the EC specification for iron oxides and hydroxides should be revised in order to ensure that iron oxides and hydroxides (E 172) as food additives will not be a significant source of exposure to these toxic elements in foods. It is also recommended that the limit specified in the EC specifications for chromium should be for the presence of chromium(III) and absence of chromium(VI).
Considering the differences in physical-chemical characteristics of the particulate material (redox state, particle size) between the different iron oxides, the Panel recommended that additional data should be provided on these compounds.
The Panel recommended that the minimum, Tier 1 testing according to the EFSA guidance (2012), should be conducted for the material as marketed as the food additive (E 172):
- red iron oxide: in vivo genotoxicity at the site of contact (gastrointestinal tract) and subchronic toxicity,
- yellow iron oxide: a complete set of genotoxicity studies and subchronic toxicity,
- black iron oxide: absorption, distribution, metabolism and excretion (ADME), in vivo genotoxicity and subchronic toxicity.