TOXICOLOGICAL EVALUATION OF CERTAIN FOOD ADDITIVES
WHO FOOD ADDITIVES SERIES 10
The evaluations contained in this document were prepared by the
Joint FAO/WHO Expert Committee on Food Additives*
Rome, 21-29 April 1976
Food and Agriculture Organization of the United Nations
World Health Organization
*Twentieth Report of the Joint FAO/WHO Expert Committee on Food
Additives, Geneva, 1976, WHO Technical Report Series No. 599, FAO Food
and Nutrition Series No. 1.
Butylated hydroxyanisole was evaluated for acceptable daily
intake for man by the Joint FAO/WHO Expert Committee on Food Additives
in 1961 and 1973 (see Annex 1, Ref. No. 6, p. 41; No. 33, p. 148).
Since the previous evaluation, additional data have become
available and are summarized and discussed in the following monograph.
The previously published monograph has been expanded and is reproduced
in its entirety below.
Absorption, distribution and excretion
Butylated hydroxyanisole (BHT) was absorbed from the
gastrointestinal tract, and there was some evidence that the feeding
of amounts 100-500 times the levels generally permitted in fats for
human consumption (in the United States of America 200 mg/kg fat)
caused deposition in depot fat, the stability of which was thereby
increased (Johnson et al., 1958). However, there was no evidence of
cumulation in other tissues (Astill et al., 1960; Bunnell et al.,
1955; Hodge et al., 1964). In the rabbit, BHA was conjugated mainly
with glucuronic acid or sulfuric acid (Dacre et al., 1956); a small
amount of unchanged BHA was excreted in the urine.
In rats the 2-tert-butyl isomer was chiefly excreted as
glucuronide, while the 3-tert-butyl isomer was excreted mainly as
ethereal sulfate (Astill et al., 1960). Thus, these animals
effectively detoxicated BHA. The changes described occurred in the
liver. No evidence has been found to suggest that BHA produces any
adverse biochemical or metabolic effect in the animal body (Dacre,
Dogs excreted 60% of a 350 mg/kg dose unchanged in the faeces
within three days. The remainder was excreted in the urine mainly as
sulfate conjugates of BHA, tert-butyl-hydroquinone and an unidentified
phenol. Only 5.5% of the dose was excreted in urine as the glucuronide
(Astill et al., 1962). Rats were injected intra-peritoneally with a
single dose of tritium-labelled BHA. Approximately 90% of the
radioactivity was recovered in the urine within four days (Golder et
Pigs fed 0.1% BHA in the diet for four months, and pullets fed
0.1% BHA in the diet for eight weeks, showed no accumulation in
muscle, liver, kidney or the reserve fat (Francois & Pihet, 1960).
Effects on enzymes and other biochemical parameters
Rats administered 500 mg/kg bw (seven daily doses) of BHA, showed
no change in liver glucose-6-phosphatase activity. BHA at dose levels
of 100 mg/kg or more (seven daily doses) caused increase in liver
weight of male rats, but in females only at doses greater than
200 mg/kg bw (seven daily doses). Liver weight was comparable to
control within 14 days of withdrawal of BHA from the diet. No fatty
changes were observed (Feuer et al., 1965a), Rats administered BHA
(500 mg/kg/day) for two days showed no increased activity of
microsomal processing enzymes (aminopyrine demethylase) hexabarbitine
oxidase and nitro-anisol demethylase (Gilbert & Golberg, 1965). In
another study, rats fed 0.1, 0.25 or 0.5% BHA in the diet for 12 days,
showed no increased liver weight, but there was an increase in liver
biphenyl-4-hydroxylase activity in the 0.5% group (Creaven et al.,
Groups each of eight rats (SPF Carworth strain) equally divided
by sex were administered by intubation daily for one week, BHA
dissolved in arachis oil, at a dose level equivalent to 0, 50, 100,
200, or 500 mg/kg body weight. BHA had no effect on the growth of
the animals. Twenty-four hours after administration of the final
dose the animals were sacrificed and liver preparations assayed
for glucose-6-phosphatase, glucose-6-phosphate dehydrogenase,
hexabarbitone oxidase, nitro-anisol demethylase and aminopyrine
demethylase activities. BHA had no effect on these enzyme activities.
Histochemical studies showed that BHA caused no fatty changes in liver
(Feuer et al., 1965b).
Groups each of two to three rhesus monkeys (macaca mulatta) one
month old infant or sexually immature juvenile, of both sexes, were
administered BHA at a dose level of 500 mg/kg body weight for four
weeks. Another group of juveniles received 50 mg BHA/kg body weight
for the same period. Controls received equivalent amounts of corn oil.
Urine and blood analyses were normal with the exception of serum
cholesterol which was elevated at the high dose at week 3 of the
study, but which returned to normal at week 4. The livers of all test
animals were enlarged. No other organ showed any pathological change.
Ultrastructurally, infant and juvenile monkeys treated at the high
dose level showed pronounced proliferation of the hepatic smooth
endoplasmic reticulum. Infants and juveniles treated with BHA
(500 mg/kg) had lower levels of liver lipids than corn oil controls.
Nitro-anisol demethylase activity was increased and glucose-6-
phosphatase activity was decreased in BHA treated juveniles but
unaffected in infant monkeys (Allen & Engblom, 1972). No changes were
observed in DNA, RNA and cytochrome P450 levels.
Additional biochemical tests were carried out on the liver and
plasma of these monkeys. BHA treated animals showed higher plasma
triglyceride levels than controls. There was a significant decrease
in the level of liver cholesterol, as well as a lowering of the
cholesterol/lipid-phosphorus ratio of the liver, in the test animals.
Liver succinic dehydrogenase was also lowered in BHT treated animals
(Branen et al., 1973).
Hepatic microsomal preparations were made from female mice (A/HcJ
strain) fed a diet of 0 and 0.5% BHA for 14 days. The aryl hydrocarbon
hydroxylase activity (AHH) of the preparations was similar. However,
microsomal preparations from the BHA fed mice showed greater
sensitivity to in vitro inhibition of AHH activity by alpha-
naphthafluorine, and continued increased amounts of cytochrome
P450 per unit weight, than preparations from control mice (Speier
& Wattenberg, 1975).
Rats were maintained on a diet supplemented with 20% lard, and
containing 0, 0.1, 0.2, 0.3, 0.4, 0.5% BHA, for a period of six weeks.
BHA caused an increase in the total serum cholesterol at the 0.1%
level, but no further elevations occurred at higher doses. There was a
relatively greater increase in the amount of serum-free cholesterol
than of ester-cholesterol. BHA produced enlarged adrenals in males at
all levels, but no histological changes were observed. Increased liver
weight at the higher dietary BHA levels was accompanied by an increase
in the absolute lipid content of the liver. However, BHA had no effect
on the concentration in liver of total and esterified cholesterol, or
the composition of the polyunsaturated fatty acids (Johnson & Hewgill,
Special studies on carcinogenicity
Groups each of 100 mice (equally divided by sex) were given
single s.c. injections (10 mg/mouse) of BHA in trioctanoin and
observed for up to 575 days. Another group was given weekly skin
applications of 0.1 mg or 10 mg of BHA in acetone, for a period of
309-459 days. Microscopic examination of the skin from the test mice
showed no evidence of tumours (Hodge et al., 1966).
BHA in lanolin applied to the ears of guinea-pigs once daily for
periods of two to six weeks, resulted in a microinvasion of basal cell
pseudopods with destruction of the superficial connective tissue and
fragmentation of the collagen. BHA alone did not cause these changes
(Riley & Seal, 1968).
Special studies on mutagenicity
Butylated hydroxyanisole (BHA) was investigated at concentrations
of 2.0, 20.0 and 200 µg/ml, in vitro, employing WI-38 human
embryonic lung cells for anaphase abnormalities. It was also
investigated in vivo by the cytogenetic analysis of metaphase cells
from rat bone marrow at dosages of 15, 150 and 1500 mg/kg. BHA did not
produce any significant increases in abnormalities above the control
values in either assay (Fabrizio, 1974).
(b) Host-mediated assay
In vitro-Salmonella TA-1530, and G-46, together with
Saccharomyces D-3 were employed. A 10% concentration was tested. BHA
was non-mutagenic for Salmonella TA-1530 and G-46. Tests with
Saccharomyces D-3 demonstrated a biologically significant increase
in the frequency of recombinants. This result could not be repeated
upon subsequent testing and was therefore thought to be spurious
In vivo-BHA was tested at 15, 150 and 1500 mg/kg in ICR Swiss
mice employing as indicator organisms Salmonella G-46, and TA-1530,
and Saccharomyces D-3. BHA was non-mutagenic for Salmonella but
demonstrated a biologically significant increase in the frequency of
recombinants. In as much as the host-mediated assay is no longer
recommended for routine use, this suggested mutagenic effect was
investigated using more sensitive (in terms of detection) procedures.
In this study BHA was investigated employing Salmonella typhimurium
strains TA-1535, TA-1537 and TA-1538, and Saccharomyces D-4 with and
without metabolic activations, in plate and suspension tests. The
percent concentrations (w/v) employed were .00375, .0075, .0150 for
Salmonella and 0625, 1250 and 02500 for Saccharomyces. Under the
conditions of this investigation BHA was non-mutagenic (Fabrizio,
(c) Dominant lethal test
Sprague-Dawley C-D strain male rats were used. Dosages of 15, 150
and 1500 mg/kg were employed.
Acute study - a single dose was administered with subsequent
mating for each of eight weeks. BHA produced random statistical
increases in dominant lethality. These were discounted due to the
unusually low negative control values (Fabrizio, 1974).
Subacute - five daily doses were administered (5 × 15, 5 × 150
and 5 × 1500 mg/kg) and males subsequently mated for each of seven
weeks. BHA produced a statistically significant increase in
pre-implantation loss in weeks 6 and 7. This effect occurring
alone is not demonstrative of mutagenicity (Fabrizio, 1974).
Special studies on teratogenicity
Groups of rats or mice of various strains were given BHA in
accordance with one of three different regimes, viz., daily
administration for seven weeks, before pairing continuing until day 18
of pregnancy or daily administration on days 1-20 of pregnancy or
single administration on day 9, 11 or 13 of pregnancy. Dosage ranged
from 250 to 1000 mg/kg bw. No teratogenic effects were observed at
dose levels as high as 300-500 mg/kg bw administered for as long as
seven weeks, although under these conditions the mortality was 25%.
At a higher dose level (750 mg/kg) mortality was 75% (Clegg, 1965).
Pregnant rats receiving a total dose of 0.5 g of BHA in the diet
showed less resorptions than rats on control diets (Telford et al.,
Other special studies
BHA at concentrations as low as 8 × 10-10 mole/litre can inhibit
the guinea-pig's smooth muscle contraction caused by bradykinin
(Posati & Pallanich, 1970).
Groups of mated pairs of Swiss-Webster mice (Mus musulus) were
maintained on diets containing 0, or 0.5% BHA. The litters obtained
from the mated pairs were weaned at 21 days and then maintained on a
diet similar to that of their mother. At six weeks of age the mice
were subjected to behavioural tests. The BHA treated offsprings showed
increased exploration, decreased sleeping, decreased self-grooming,
slower learning and a decreased orientation complex than did the
control group (Stokes & Scudder, 1974).
Using an in situ method of perfusion for rat intestine, BHA at
a level of 2 mg/ml has been shown to reduce the absorption of glucose
and methionine, but not butyric acid (Fritsch et al, 1975a). BHA at
levels of 400 µg/ml caused an inhibition of the metabolism (as
measured by gas evolution) of culture of bacteria isolated from the
cecal flora of rats (Fritsch et al, 1975b).
Animal Route (mg/kg bw) References
Mouse oral 2 000 Bunnell et al., 1955;
Lehman et al., 1951
Rat oral 2 200-5 000 Bunnell et al., 1955;
Lehman et al., 1951
No effect on potassium excretion, as described below for the
rabbit was observed in a short-term feeding study in the rat (Dacre,
Groups of seven recently weaned rats were fed for six months on
rations containing 0, 0.5, 1, 2 and 3% of BHA. The rats at the 3%
level did not eat enough to gain weight and were put on to the 2%
diet for a time, then returned to 3%. Even at the 2% level, food
consumption was not optimal. Histopathological examination revealed no
pathological condition attributable to BHA (Wilder & Kraybill, 1948).
Combinations of BHA with other food additives, such as chlorine
dioxide, sodium propionate, propyl gallate, or polyoxyethylene-8-
stearate, at 50 times the normal levels of use in bread, had no
deleterious effects when they were fed in bread to groups of 26 rats
for a period of 32 weeks. The treated bread formed 75% of the animals'
diet. The daily dosage levels of BHA were from 3.3 to 7.0 mg/kg bw
(Graham et al., 1954; Graham & Grice, 1955).
Rats were maintained on test diets containing 0, or the
equivalent of 500-600 mg/kg bw BHA (1/5 of the LD50), for a
period of 10 weeks. The test animals showed decreased growth rate,
and reduced activity of the blood enzymes, catalase, peroxidase and
cholinesterase. Chemical analysis of livers of test animals showed a
decrease in the amount of phospholipid as compared to controls, but
there was no lipid accumulation. Histological examination of the
tissues and organs did not show any compound-related effects
In rabbits, a dose of 1 g given daily for five to six days by
stomach tube caused a ten-fold increase in sodium excretion and a 20%
increase in potassium excretion in the urine. Extracellular fluid
volume fell, and this prevented any marked change in the plasma sodium
level. The serum potassium fell after five days' treatment and
potassium was being replaced by sodium in muscle cells. In heart
muscle the changes occurred later than in skeletal muscle and were
less marked. The antioxidant may have a direct effect on the kidney;
the adrenal cortex showed changes in the zona glomerulosa and there
was increased excretion of aldosterone in the urine, associated with
the sodium and potassium loss (Denz & Llaurado, 1957).
When BHA was fed to dogs at dose levels of 0, 0.3, 30 and
100 mg/kg bw for one year, no ill-effects were observed. Renal
function, haematology and histopathology of the main tissues were
normal. Organ weights were within normal limits and there was no
demonstrable storage of BHA. The urine did not contain a demonstrable
increase of reducing substances, even when 100 mg/kg bw of BHA was
fed. Groups of three dogs were used at each dose level for these
experiments (Hodge et al., 1964).
Groups each of four weanling dogs were fed BHA at 0, 5, 50 and
250 mg/kg for 15 months. General health and weight gains of the dogs
were within normal range, as were haematologic parameters. Urine from
test dogs contained higher ratios of total to inorganic sulfate, and
glucuronates than controls. At autopsy, microscopic examination of
tissues and organs showed that three/four of the animals at the
highest dose level tested had a liver cell degeneration and a diffuse
granulocytic infiltration. The lobular structure of the livers of
these animals was normal, and there was no excessive connective tissue
proliferation (Wilder et al., 1960).
A group of six adult female rhesus monkeys were maintained on a
test diet containing a mixture of BHT and BHA that provide an intake
equivalent to 50 mg BHT and 50 mg BHA/kg body weight. Another group of
six adult female rhesus monkeys were used as controls. The monkeys
were fed the diet for one year prior to breeding and then for an
additional year, including a 165 day gestation period. Haematologic
studies including haemoglobin, haematocrit, total as well as
differential white blood cell count, cholesterol, Na+, K+, total
protein, serum glutamic pyruvic transaminase, and serum glutamic
oxylacetic transaminase, were carried out at monthly intervals. Body
weights were taken at monthly intervals. Records of menstrual cycles
were maintained through the test period.
After one year the females were bred to rhesus males not
receiving test diets. During pregnancy complete blood counts were done
on days 40, 80, 120 and 160 of gestation and on days 30 and 60 post-
partum. A total of five infants were born to the experiment monkeys
and six to the control monkeys. Haematological evaluations were made
on infants of the test and control monkeys at days 1, 5, 15, 30 and
60, and observations of the infants were continued through two years
of age. Two experimental and two control infants, three months of age,
were, removed from their mothers for one month of psychological home
cage observations. No clinical abnormalities were observed in parent
and offspring during the period of study. The gestation of test
animals was free of complications and normal infants were delivered.
Adult females continued to have normal infants. Infants born during
the exposure period remained healthy, with the exception of one infant
that died from unrelated causes. Home cage observations at the third
month of life did not reveal any behavioural abnormalities (Allen,
Groups of 15 or more newly weaned rats were placed on diets
containing 0, 0.05, 0.5 and 1% BHA in lard (0, 0.003, 0.03 and 0.06%
of the total diet) for 22 months. Weight gain was comparable in all
groups. Reproduction was normal, and young rats kept on the same
ration grew normally. Number, size, weight, weight gain and mortality
of the litters were comparable for animals of all groups. After one
year on test, the colony suffered from an infectious respiratory
disease and many died. There was no significant difference in
mortality among the groups. After 22 months, the remaining animals
were killed; histopathological examination revealed no changes
attributable to the antioxidant (Wilder & Kraybill, 1948).
A similar series of tests was undertaken, with an additional
group on a diet of 2% BHA in lard (0.12% of the total diet). There
were 17 rats in each group. After 21 months, the survivors were
killed. Histopathological examination revealed no significant
differences compared with the control animals. The rate of gain in
weight during the growing period was unchanged, and all rats appeared
normal in every respect (Wilder & Kraybill, 1948).
In another rat feeding test carried out over a period of two
years on groups of 40 rats, there was a small reduction in the mature
weight and an increase in relative liver weight in some cases with the
highest level of BHA used (0.5% of the diet), but there were no
effects on any of the following: the reproductive cycle; histology of
the spleen, kidney, liver, or skin; ratio of weight of heart, spleen,
or kidneys to total body weight:mortality. The toxicity of BHA was not
affected by the dietary fat load (Brown et al., 1959).
Rats were maintained on diets containing 0, and the equivalent of
500-600 mg/kg bw BHA (1/5 of the LD50) for a period of one year.
During the course of this study, rats were bred to produce three
successive generations. Two generations were maintained on the test
diet for six months. BHA had no effect on reproductive performance, as
measured by litter size, birth weight, date of appearance of incisors,
and opening of eyes. Autopsy and histological examination of tissues
and organs of parents and offsprings at the termination of the study
did not reveal any compound-related effect (Karplyuk, 1962).
OBSERVATIONS IN MAN
Human volunteers were dosed with 0.5-0.7 mg/kg bw BHA. 22-77% was
excreted in the urine as the glucoronide within 24 hours. Less than 1%
was excreted in the urine as unchanged BHA, and no dealkylation or
hydroxylation products were detected (Astill et al., 1962). In
another study, human volunteers were administered a single dose of
14C-labelled BHA (approximately 0.5 mg/kg bw). 60-70% of the
radioactivity was excreted in the urine within two days, and by day
11 post-dosing, 80-86.5% of the radioactivity was recovered in the
urine (Daniel et al., 1967).
Several metabolic studies with orally administered BHA are
available in rats, mice and monkeys. In the rat, at the levels tested,
there appeared no change in the activity of the liver enzymes studied.
In the case of mice, although there appeared no increase in aryl
hydrocarbon hydroxylase, there was an increase in cytochrome P450 and
the nature of the enzyme appears altered. With monkey, on the other
hand, cytochrome P450 levels appeared unaffected whereas some enzyme
activities were affected. In this regard, infant monkeys, as opposed
to juveniles, did not show these changes. BHA appeared to lower liver
cholesterol whereas plasma triglycerides were increased.
Female monkeys maintained on a dietary intake of 50 mg/kg body
weight for one year prior to breeding, and bred to untreated males,
gave birth to normal offspring. BHA had no effect on any of the
indices tested. These included haematology and clinical biochemistry
as well as behavioural observations on the young.
Mutagenic activity of BHA has been tested in several in vitro
and in vivo systems. The collective assessment of these tests does
not show BHA to be mutagenic.
The previously stated requirement that studies on the effect on
reproduction of mixtures of BHA, BHT and propyl gallate was considered
to be no longer necessary.
Level causing no toxicological effect
Rat: 5000 ppm (0.5%) in the diet equivalent to 250 mg/kg bw.
Estimate of acceptable daily intake for man
0-0.51 mg/kg bw.2
FURTHER WORK OR INFORMATION
Required by 1980.
A multigeneration reproduction study.
Allen, J. R. (1976) Annual Report of the University of Wisconsin Food
Research Institute 1974, pp. 308-315
Allen, J. R. & Engblom, J. F. (1972), Fd. Cosmet. Toxicol., 10, 769
Astill, B. D. et al. (1962) J. Ag. & Food Chem., 10, 315
Astill, B. D., Fassett, D. W. & Roudabush, R. L. (1960) Biochem. J.,
Branen, A. L. et al. (1973) Fd. Cosmet. Toxicol., 11, 797
Brown, W. D., Johnson, A. R. & O'Halloran, M. W. (1959) Aust. J. exp.
Biol. med. Sci., 37, 533
Bunnell, R. H. et al. (1955) Poultry Sci., 34, 1068
Clegg, D. J. (1965) Fd. Cosmet. Toxicol., 3 387
Creaven, P. J., Davies, W. H. & Williams, R. T. (1966) J. Pharm.
Pharmacol., 18, 485
Dacre, J. C. (1960) N. Z. J. Inst. Chem., 24, 161
Dacre, J. C., Denz, F. A. & Kennedy, T. H. (1956) Biochem. J., 64, 777
Daniel, J. W. et al. (1967) Fd. Cosmet. Toxicol., 5, 475
Denz, F. A. & Llaurado, J. G. (1957) Brit. J. exp. Path., 38, 515
1 As BHA, BHT or the sum of both.
Fabrizio, D. P. A. (1974) Mutagenic evaluation of compound FDA 71-24:
Butylated hydroxyanisole. Unpublished report from Litton
Bionetics, Inc., Kensington, Md. US, submitted to the World
Health Organization by the US Food and Drug Administration
Feuer, G. et al. (1965) Fd. Cosmet. Toxicol., 3 457
Feuer, G., Golberg, L. & LePelley, J. R. (1965a) Fd. Cosmet. Toxicol.,
Francois, A. C. & Pihet, A. (1960) Ann. ind. natl. recherche agron.
Ser., D9, 195
Fritsch, P., de Saint-Blanquat, G. & Derache, R. (1975a) Eur. J. Tox.,
Fritsch, P., Lamboeuf, Y. & de Saint-Blanquat, G. (1975b) Toxicology,
Gilbert, D. & Golberg, L. (1965) Fd. Cosmet. Toxicol., 3, 417
Golder, W. S., Ryan, A. J. & Wright, S. E. (1962) J. Pharm.
Pharmacol., 14, 268
Graham, W. D. & Grice, H. C. (1955) J. Pharm. (Lond.), 7, 126
Graham, W. D., Teed, H. & Grice, H. C. (1954) J. Pharm. (Lond.), 6,
Hodge, H. C. et al. (1966) Toxicol. appl. Pharmacol., 9, 583
Hodge, H. C. et al. (1964) Toxicol. appl. Pharmacol., 6, 512
Johnson, A. R. & Hewgill, E. R. (1961) Aust. J. exp. Biol. med. Sci.,
Johnson, A. R., O'Halloran, M. W. & Hewgill, F. R. (1928) J. Amer. oil
Chem. Soc., 35, 496
Karplyuk, I. A. (1962) Tr. 2-01 (Vtoroi) Nauchn. Konf. po Vopr. Probl.
Zhira v Pitanii, Leningrad, 1962, 318
Lehman, A. J. et al. (1951) Advanc. Food Res., 3, 197
Posati, L. P. & Pallansch, M.J. (1970) Science, 168, 121
Riley, P. A. & Seal, P. (1968) Nature, 220, 922
Speier, J. L. & Wattenberg, L. W. (1975) J. nat. Cancer Inst., 55, 469
Stokes, J. D. & Scudder, C. L. (1974) Development Psychobiology, 7,
Telford, I. R., Woodruff, C. S. & Linford, R.H. (1962) Amer. J. Anat.,
Wilder, O. H. M., Ostby, P. C. & Gregory, B. A. (1960) J. Ag. & Food
Chem, 8, 504
Wilder, O. H. M. & Kraybill, H. R. (1948) Summary of toxicity studies
on butylated hydroxyanisole, American Meat Institute
Foundation, University of Chicago