INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY
WORLD HEALTH ORGANIZATION
TOXICOLOGICAL EVALUATION OF SOME
FOOD COLOURS, ENZYMES, FLAVOUR
ENHANCERS, THICKENING AGENTS, AND
CERTAIN FOOD ADDITIVES
WHO FOOD ADDITIVES SERIES 6
The evaluations contained in this publication were prepared by the
Joint FAO/WHO Expert Committee on Food Additives which met in Rome,
4-13 June 19741
World Health Organization Geneva 1975
1 Eighteenth Report of the Joint FAO/WHO Expert Committee on
Food Additives, Wld Hlth Org. techn. Rep. Ser., 1974, No. 557.
FAO Nutrition Meetings Report Series, 1974, No. 54.
This compound has been evaluated for acceptable daily intake by
the Joint FAO/WHO Expert Committee on Food Additives (see Annex 1,
Ref. Nos. 10 and 20) in 1964 and 1969.
Since the previous evaluation additional data have become
available and are summarized and discussed in the following monograph.
The previously published monographs have been expanded and are
reproduced in their entirety below.
The metabolic behaviour and excretory pattern for erythrosine
have been studied in adult rats. The colour was given by stomach tube
in log-spaced doses from 0.5-500 mg per kg bw. In five days the
recovery in the excreta was 102%. After an intravenous application of
3 mg per kg bw the urine and bile for the initial two to four hours
was collected, an average of 55% (50.4-58.0%) of the administered
quantity was found in the bile. In the urine 1.3% (0.8-1.8%). No
glucuronic acid conjugation is found.
This colour was found to be largely excreted in the faeces by
rats (55-72%) and despite the presence of two groups capable of
undergoing conjugations, no colour could be identified in the urine. A
small amount of the colour (0.4-1.7%) was excreted in the bile
Consideration has been given to the possibility that iodide may
be liberated from erythrosine and may disturb thyroid function. In the
rat, erythrosine is metabolically stable and 100% of the amount is
ingested and excreted with its iodine content intact after
administration of 500 mg/kg (Webb et al., 1962). Protein-bound and
total blood iodine levels were elevated in rats given erythrosine by
stomach tube twice weekly in a chronic study (Bowie et al., 1966).
However the elevated PBIs (protein-bound iodine) were due to
interference by erythrosine in PBI determinations rather than thyroid
dysfunction in rats and gerbils (USFDA, 1969). In man, oral
administration of 16 mg of erythrosine daily for 10 days resulted in
an increase of protein-bound iodine in the serum from 6-11 µg/100 ml
after 15 to 20 days, followed by a sharp decline in iodine levels in
the next 10 days with gradual return to the initial value in three
months (Anderson et al., 1964). Erythrosine could be an adventitious
source of iodide (Vought et al., 1972).
Large doses of erythrosine labelled with l131 given orally to
rats inhibited uptake of l131 by the thyroid of treated animals. Daily
doses over 1 mg are necessary for this effect (Marignan et al., 1965).
When cherries coloured with erythrosine are stored in plain cans,
fluorescein is readily formed by interaction of the tin-iron couple
present. This does not occur in lacquered cans. The production of
fluorescein (with 4 I atoms) from erythrosine occurs in presence of
metallic iron and/or tin and free organic acid (result of
electrochemical reduction in the can) (Dickinson & Raven, 1962).
It was found that this colour in a concentration of 200-400 mg/l,
inhibited the action of pepsin but had no effect on lipase activity
(Diemair & Hausser, 1951).
It was also found that this colour had in vivo as well as
in vitro haemolytic effect. In the in vivo studies the mouse was
used (Waliszowski, 1952).
Erythrosine was administered to rats in doses of 5, 10, 15 and
50 mg per rat weighing 200-250 g, twice weekly for six months.
Haemoglobin was reduced at three months, also the red cell count. The
cholesterol levels of males were depressed. Excretion of the dye was
mainly in the faeces and predominantly unchanged (Bowie et al., 1966).
Special studies on mutagenicity
This colour was tested for mutagenic activity and showed a very
slight but statistically significant mutagenic effect on Escherichia
coli in concentrations of 0.5 g/100 ml. It was found that xanthene
molecule itself was the causative factor and that the substituent
groups only modify the effect (Lück et al., 1963; Lück & Rickerl,
Special studies on reproduction
Groups of 20 female and 10 male rats were placed on dietary
levels of 0, 1.25, 12.5, 37.5 and 125.0 mg/kg of erythrosine and used
to carry out a three-generation reproduction study. No adverse effects
were noted upon fertility, litter size, litter viability or post-
partum development (Anonymous, 1974).
Special studies on teratogenicity
Groups of 15 to 19 pregnant rats were given orally 25, 85 and
250 mg/kg per day by gavage from days 6 to 15 of gestation and were
sacrificed on day 20. Control groups of 16 to 17 pregnant rats and
positive control groups of 21 rats (given 200 mg sodium salicylate/
kg/day) were used. Body weight, mortality of dams and general
appearance showed no adverse effects. There were no specific effects
due to colour on fetal mortality, fetal weight, external and internal
fetal development and skeletal development. The species was sensitive
to salicylate (Anonymous, 1972a).
Groups of 12 pregnant does were given orally in gelatin capsules
12.5, 40 or 125 mg/kg bw of colour from day 6 to 18 and were
sacrificed on day 29. 10-14 does acted as controls and 12-14 does as
positive control given thalidomide. No adverse effects were seen due
to the colour on body weight gain, mortality of dams, fetal survival,
fetal body weights, internal and external fetal development and
skeletal development. Thalidomide gave the expected result (Anonymous,
Animal Route mg/kg bw Reference
Mouse Oral 6 800 BIBRA, 1973
i.v 370 Waliszewski, 1952, DFG, 1957
Rat i.p. 300 DFG, 1957
Oral 360 BIBRA, 1973
> 2 000 DFG, 1957
7 100 BIBRA, 1973
1 895 Lu. & Lavallee, 1964
1 840 Hansen et al., 1973
Rabbit i.p. 350 BIBRA, 1973
i.v. 200 DFG, 1957
Gerbil Oral 1 930 USFDA, 1969
A group of five young rats were given subcutaneous injections
twice dally for three days. The rats were killed on the fourth day.
The colour was administered in aqueous solution at a level of 250 mg
per kg bw each injection. No oestrogenic activity (normal uterine
weight) was detected (Graham & Allmark, 1959).
In experiments with guinea-pigs it was found that this colour had
no sensitization activity (Bär & Griepentrog, 1960).
In a 90-day study on five groups of 15 male and 15 female rats
erythrosine was given in the diets at 0.25%, 0.5%, 1% and 2%. No
adverse effects were noted as regards body weight, food intake,
haematology, blood and urine analysis which were related to
administration of test substance. Organ weights were normal except
that absolute and relative caecal enlargement was seen at all levels
tested. It was dose-related but histology was normal. Absolute and
relative thyroid weight was increased at the 2% level. Histopathology
showed no abnormalities except pigment deposition in renal tubules in
females only at the 2% level but in males at all levels again in a
dose-related manner. The pigment was identified as protein-bound
erythrosine. In addition total PBI was raised at all levels in a dose-
related manner, protein-bound erythrosine in serum behaved similarly
and non-protein bound iodine also increased with dose levels.
Thyroxine iodine however remained unchanged and 1131 uptake was
reduced (BIBRA, 1973).
Four groups of three male and three female pigs were given 0,
167, 500 and 1500 mg/kg/day in their diet for 90 days. No obvious
adverse effects were noted but detailed reports are not yet available
A total of 122 male and female mice produced by mixed breeding
from five different strains were given a diet containing 1 mg per
animal per day of the colour. Mice at the age of 50-100 days were
used. A number of the animals were sacrificed after an observation
period of 500 days and the survivors of the rest after 700 days. A
total of 168 mice was used as control. Positive control groups which
were given O-aminoazotoluene and dimethylaminoazobenzene were also
included. In these groups the formation of liver tumours was noted
after approximately 200 days. The incidence of tumours in mice
receiving the colour was not significantly greater than in the
controls (Waterman & Lignac, 1958). Chronic feeding studies were
conducted with mice. Seventy mice were fed at 1 and 2%. Because of the
small number of animals surviving the experiment and the small number
of tumours found, no effect of tumour formation could be attributed to
the colour (USFDA, 1964).
Groups of 24 weanling rats, evenly divided by sex, were fed this
colour at 0, 0.5, 1.0, 2.0 and 5.0% for two years. Slight growth
depression was observed in the animals at the 5% level, and those
above 0.5% had distended caecums but microscopically the distended
caecums showed normal histology. The statistical evaluation of the rat
study revealed no changes in organ weights at the highest level. There
was some diarrhoea at the 5% level. There was no difference in
survival (USFDA, 1964). Forty rats were fed 1% of the colour for two
years. Six animals died between time. No tumours were found (DFG,
The colour was fed at a level of 4% of the diet to five male and
five female rats for periods up to 18 months. Gross staining was
observed in the glandular stomach and small intestine and granular
deposits in the stomach, small intestine and colon. Hepatic cirrhosis
was noted in one case out of four rats living up to 12 months. Fifty
control animals observed for 20 months or more failed to develop
tumours, or hepatic cirrhosis (Willheim & Ivy, 1953).
Groups of 12 male and 12 female weanling Osborne-Mendel rats were
fed 0, 0.5, 1.0, 2.0 and 5.0% erythrosine in their diet for two years.
Growth depression was observed in rats given 5%. The relative spleen
weight was depressed in all male test groups and in females at the 5%
level. Slight caecal enlargement was noted at 1% and increased with
dose but the histology of the enlarged caeca was normal. No other
gross or histopathological findings related to colour administration
were noted (Hansen et al., 1973).
Groups of 25 male and 25 female 100-day-old rats and a group of
50 male and 50 female controls were fed 0, 0.5, 1.0, 2.0 and 4.0%
erythrosine in their diet for 86 weeks. Other groups of 25 male and 25
female rats aged 100 days were intubated twice a week for 85 weeks
with erythrosine at 0, 100, 235, 750 and 1500 mg/kg bw. After this
treatment animals were kept on normal diets until two years. Body
weight decreases were seen at 2 and 4%. Elevated PBI, due to
interference by erythrosine with PBI determination rather than thyroid
dysfunction, were seen. Thyroxine-iodine levels were not affected.
There were no other haematological differences and no anaemia was
seen. No adverse gross pathology was noted; histopathology had not
shown any colour related abnormalities (Hansen et al., 1973a).
Feeding with 1% of the colour for two years did not influence the
body weight or fertility of Wistar rats. There was no pathological
damage and no significant difference in tumour incidence from control
rats (Oettel et al., 1965).
Twenty rats were subject to weekly subcutaneous injections of
1 ml of a 5% aqueous solution for 596 days. The total quantity of
colour administered was 2.65 g/animal. Seven rats survived 300 days or
more. No tumours were observed (Umeda, 1956).
Eighteen rats were injected subcutaneously with aqueous solutions
of erythrosine at 12 mg/animal once per week for two years. No tumours
either at the injection sites or in other parts of the body were
observed (Hansen et al., 1973).
Groups of gerbils, 30/test and 60/control, evenly divided by sex
twice a week for 19 months with 0, 200, 750 and 900 mg/kg (those on
900 mg received 1200 mg for the first three months). Body weight
decreases were seen at all feeding levels (only females at 1%).
Elevated PBIs, due to interference by erythrosine with PBI
determination, were seen. No other haematological differences were
seen. No adverse gross pathology was noted. Histopathology has not
been completed (USFDA, 1969).
Two-year feeding studies were conducted with groups of three male
and three female beagles at levels of 0, 0.5, 1.0 and 2.0% in the
diet. All dogs survived the study. No gross or microscopic pathology
related to colour administration was seen (Hansen et al., 1973).
There are adequate long-term studies on several species available
as well as some information.
The observed elevation of PBI levels does not appear to have any
toxicological significance in relation to the thyroid activity. More
extensive metabolic studies in other species preferably including man
and of the mechanism underlying the effect on plasma-bound iodine are
Level causing no toxicological effect
Rat: 0.5% (= 5000 ppm) in the diet equivalent to 250 mg/kg bw.
Estimate of acceptable daily intake for man
0-2.5 mg/kg bw
FURTHER WORK OR INFORMATION
Metabolic studies, preferably including man.
Anderson, C. J., Keiding, N. R. & Nicholson, A. B. (1964) Scand.
J. Clin. Lab. Inrest., 16, 549
Anonymous (1972a) Industrial Biotest Laboratory, Unpublished report
submitted to WHO
Anonymous (1972b) Industrial Biotest Laboratory, Unpublished report
submitted to WHO
Anonymous (1974) Industrial Biotest Laboratories, Unpublished report
submitted to WHO
Bär, F. & Griepentrog, F. (1960) Med. u. Ernahr, 1, 99
BIBRA (1967) Fd Cosmet. Toxicol., 5, 100
Bowie, W. C., Wallace, W. C. & Lindstrom, H. V. (1966) Fed. Proc., 25,
Daniel, J. W. (1962) Toxicol. appl. Pharmacol., 4, 572
Deutsche Forschungsgemeinschaft, Farbstoff Kommission (1957)
Dickinson, D. & Raven, T. W. (1962) J. Sci. Food Agric., 13, 650
Diemair, W. & Hausser, H. (1951) Z. Lebensmitt-Untersuch., 92, 1965
Graham, R. C. B. & Allmark, M. G. (1959) Toxicol. appl. Pharmacol., 1,
Grasso, P. & Golberg, L. (1966) Fd. Cosmet. Toxicol., 4, 269
Hansen, W. H. et al. (1973) Fd. Cosmet. Toxicol., 11(4), 527
Hansen, W. H. et al. (1973a) Fd. Cosmet. Toxicol., 11(4), 535
Lu, F. C. & Lavallee, A. (1964) Canad. pharm. J., 97, 30
Lück, H. & Rickerl, E. (1960) Z. Lebensmitt-Untersuch., 122, 157
Lück, H., Wallnofer, P. & Bach, H. (1963) Path. et Microbiol. (Basel),
Marignan, R., Boucard, M. & Gelis, C. (1965) Trav. Soc. Pharm.
Montpellier, 24, 127
Nelson, A. A. & Hagan, E. C. (1953) Fed. Proc., 12, 397
Umeda, M. (1956) Gann, 47, 51
United States Food and Drug Administration (1969) Unpublished report
Vought, R. L., Brown, F. A. & Wolff, J. (1972) J. Clin. Endocr.
Metal., 34, 747
Waliszewski, T. (1952) Acta Pol. pharm., 9, 127
Waterman, N. & Lignac, G. O. E., as summarized by Genderen, H. van
(1958) Acta Physiol. pharmacol, neerl., 7, 35
Webb, J. M., Fonda, M. & Brouwer, E. A. (1962) J. Pharmacol. exp.
Ther., 137, 141
Willheim, R. & Ivy, A. C. (1953) Gastroenterology, 23, 1