BUTYLATED HYDROXYANISOLE (BHA)
This substance was evaluated for acceptable daily intake for man
(ADI) by the Joint FAO/WHO Expert Committee on Food Additives in 1961,
1965, 1973, 1976, 1977, 1980 and 1982 (see Annex I, Refs. 6, 11, 32,
40, 43, 54 and 60). Toxicological monographs were issued in 1961,
1973, 1976 and 1980 (see Annex I, Refs. 6, 33, 41 and 55).
Since the previous evaluation, additional data have become
available and are summarized and discussed in the following monograph.
Effects on enzymes and other biochemical parameters
3-tert-butyl-4-hydroxyanisole (3-BHA) the major isomeric
constituent of BHA (93% or more) or 2-tert-butyl-4-hydroxyanisole
(2-BHA) were fed to mice for two weeks at dietary concentration 0.5%
either singly or as mixtures of the isomers in different ratios. In
the liver the cytosolic GSH-S-transferase was induced to a higher
level by 3-BHA than by 2-BHA, whereas in the forestomach the induction
characteristics of the isomers were reversed. Sulfhydryl levels in
these tissues followed the same trend. Significant synergistic effects
of the two isomers were observed in the forestomach but not in the
liver. The epoxide hydrolase activity of isolated liver microsomes
from the treated mice was enhanced by 3-BHA, but only marginally by
2-BHA (Lam et al., 1981). Cytosolic and microsomal glutathione
transferase activity (GT) in the liver of mice was markedly increased
when BHA was incorporated into the diet. In a parallel study in the
rat, only a minor increase (in GT) was observed (Morgenstern & Dock,
Special studies on carcinogenicity
In a 12-week range finding study, groups of Fischer-344 rats were
maintained on diets containing 0, 0.25, 0.5, 1, 2, or 4% BHA. Four per
cent. in the diet caused a marked growth retardation. At 2% there was
only slight growth retardation. Decrease in body weight gain was used
as an end-point in determining the maximum tolerated. A high incidence
of hyperplasia and hyperkeratosis of the forestomach epithelium was
reported at the higher dose levels (2% and 4%). No changes were seen
except at the limiting ridge (Ito, 1983).
Groups each of 51 or 52, five-week-old F-344 rats of both sexes
were maintained on diets in which had been incorporated 0, 0.5 or 2.0%
BHA. The diet mixtures were pelletized and one analysis of the pellets
showed that the actual concentrations of BHA in the pellets were 0.25%
and 1.07%, respectively. The loss of BHA was considered to be due to
sublimation during the manufacturing process, which included a heating
and steaming process (Ishiwata et al., 1982). In a subsequent study on
the BHA content of the diet, it was shown that the actual dosage
levels were similar to those initially reported (0.5% and 2%), and
that no BHA was lost during the processing. However, the processing
could result in a reduction of vitamin A levels (MAFF, 1983). No
unusual degradation products were reported to be present in the diet.
There was a slight increase in the amount of dimer present. The rats
were maintained on the test diets for 104 weeks, and then on basal
diets until week 112. Body weight and food consumption was measured
weekly for the first 26 weeks, and then every two weeks. One week
prior to sacrifice urine samples were collected from 10 rats/sex/group
for the usual urinalysis. At termination of the study blood samples
were collected for a haematological evaluation and blood serum
chemistry. At week 112, surviving animals were killed and examined
grossly. The weights of brain, salivary gland, heart, lung, liver,
kidney, spleen, testis, ovary and pituitary were measured. Tissues
were examined histologically. There were no differences in appearance
and behaviour of test and control animals during the course of the
study. Food intake for male test rats was similar to that of control,
but the high-dose females showed a slightly lower food intake. The
body weight of males and females in the high-dose group was
approximately 10% less than controls. The percentage of survival of
animals in week 112 was, for male, control 35/51 (68.6%), low dose
35/51 (68.6%), high dose 32/52 (67.3%), for female, control 33/51
(64.7%), low dose 34/51 (66.7%), high dose 40/51 (78.4%).
Haematological and serum chemistry analyses showed only minor changes
related to BHA administration, and no compound-related effects were
observed in the urinalysis. A significant decrease was observed in the
brain weights of males in the test group, and in the relative weights
of salivary glands and hearts of females. There was a very significant
increase in the incidence of forestomach hyperplasia and neoplasia
(papillomas and squamous cell carcinomas) in the test animals which
was clearly dose-related. There was an almost complete absence of
irritation or inflammation of the forestomach in the dosed animals.
The first squamous cell carcinoma was found in a male rat in the
high-dose level group at week 59 and in a female in the high-dose
group at week 82. Two males and one female in the high-dose group had
metastases to the lymph node, and one female in the high-dose group
had an invasive lesion of the liver.
No. of rats Changes in forestomach (%)
Sex Group surviving Squamous
more than Hyperplasia Papilloma cell
41 weeks) carcinoma
Male Control 51 0 0 0
Low 50 13 (26%) 1 (2.0%) 0
High 52 52 (100%) 52 (100%) 18 (34.6%)
Female Control 51 0 0 0
Low 51 10 (19.6%) 1 (2.0%) 0
High 51 50 (98%) 49 (96.1%) 15 (29.6%)
The incidence of tumours in other organs was not significantly
different between test and control animals.
Bile duct proliferation, which was present in all groups, was
reduced in a dose-related manner in the BHA groups. No other compound-
related effects were reported (Ito et al., 1982).
A subsequent review of the pathology slides by an international
group concurred with these findings, with the exception that the
papillomas observed in the low-dose group were considered to be no
more than marked hyperplasia, and the two high-dose males in which the
lesions were reported to be papillomas were considered to be squamous
Groups each of 20-30 male or female Fischer-344 rats were fed
pelleted diets containing 0, 0.04, 0.2 or 1.0% BHA, for 96 weeks. Food
consumption and water intake was similar for test and control animals.
Haematological and serum biochemical parameters were similar for test
and control animals, with the exception of LDH activity which was
elevated in the 0.2% and 1.0% groups when compared to control.
Survival, body weight, and organ weights were not significantly
affected by the dietary BHA. Tumours reported included pituitary
adenomas and subcutaneous sarcomas in the test and control groups,
hepatohaemangiomas in the test animals, one renal cell carcinoma in a
female in the 0.2% group, and an adenocarcinoma of the glandular
stomach of a male in the 0.4% group, one thyroid adenoma and one
interstitial tumour of the testis. Papillomas of the forestomach were
found in the 1.0% BHA group (3/15 males and 2/18 females). None was
present in the control group (nine males/13 females). The occurrence
of papillomas is statistically significant if both sexes are combined
(P = 0.007 Fischer Exact Test) (Tomii & Aoki, 1982).
In another study, BHA at levels of 0.5, 1.0 or 2.0% in the diet
of Fischer-344 rats (male) was shown to induce proliferative changes
in the forestomach of rats, using either visual, histopathological or
radioautographic (thymidine-H3-labelled incorporation) observations.
Changes were observed nine days after administration of the test diet,
and occurred when the BHA was dissolved in corn oil prior to addition
to the diet, or when the BHA was ground into the diet, or if the diet
were prepared in a manner similar to that described in the Ito study.
The "no observed effect level" for the change was 0.25% BHA in the
diet (Lok et al., 1983).
Potentiation or inhibition of carcinogenesis
Groups of rats were administered N-butyl-N-(4-hydroxybutyl)
nitrosamine in drinking-water at 0.01% or 0.05%. From week 4 to 36 of
the study, BHA (2%), BHT (1%), or L-ascorbic acid (5% or 1%) was
administered in the diet. Both BHA and BHT enhanced hyperplasia of the
bladder in the rats exposed to 0.01% BBN, and enhanced bladder
papillomas and cancer incidence in the 0.05% BBN group. Two per cent.
ascorbate enhanced hyperplasia in the 0.01% BBN group and papilloma
formation in the 0.05% BBN group. One per cent. ascorbate was without
effect in either group (Ito, 1983).
In another study, male six-week-old, F-344 rats were administered
i.p. methylnitrosurea (MNU) at dose levels equivalent to 20 mg/kg,
twice a week for four weeks, and then maintained on diets containing
2% BHA or 1% BHT or 5% sodium L-ascorbate for 32 weeks. Dietary BHA
and ascorbate enhanced the incidence of cancer of the forestomach in
the MNU treated rats. BHT had no effect (Ito, 1983).
In a study of the incidence of gamma-glutamyl transferase foci
induced by diethylnitrosamine (200 mg/kg) in partially hepatectamized
rats, diets containing 2% BHA or 1% BHT or 1% ethoxyquine inhibited
the formation of foci in the liver. Five per cent. ascorbate or 0.5%
disulfuram had no effect (Ito, 1982).
Male A/J mice were dosed i.p. with a single dose of urethan, and
then fed diets containing either 0.75% BHT or 0.75% ethoxyquine,
either once a week or continuously for eight weeks. Lung rumour yield
was scored four months after the urethan treatment. Dietary BHT, but
not BHA or ethoxyquine, enhanced lung tumour formation under either
conditions of the test. When the mice were prefed with diets
containing either BHA or BHT for two weeks, prior to urethan
treatment, and then maintained on conventional laboratory diets for
four months, the BHT diet had no effect on tumour yield, but the BHA
diet significantly reduced the average number of lung tumours
Groups each of 90-130 female CF1 mice (five weeks of age) were
fed control diets or diets containing 300, 1000, 3000 or 6000 ppm
(0.03, 0.1, 0.3 or 0.6%) BHA. After two weeks on the diet all but 10
mice in each group were injected with 20 mg of methylazomethanol
acetate (MAM - a colon-specific cancer compound). A second injection
was given four days later. The BHA protected against death caused by
MAM. The mortality of the MAM-treated mice on control diets was 80%
and 92%, and those fed diets with BHA were inversely related to the
amount of BHA in the diet being 0% and 1% at the 6000 ppm (0.6%)
dietary level. Dietary BHA at the 3000 or 6000 ppm (0.3 or 0.6%) also
reduced and eliminated histopathological changes observed in the
livers and other organs of the mice. Microsomal preparations from the
livers or test animals, showed a marked increase in the levels of
cytochrome P-450 and b5, when compared to controls (Reddy et al.,
Female ICR/Ha mice, nine weeks of age, were maintained on diets
containing 0, 0.03 or 0.06 mmol/g of the isomers of BHA. On day 8 of
the study, the mice were given the first of eight doses (twice a week
for four weeks) of 1 mg of benzo(alpha)pyrene in corn oil by p.o.
intubation. The mice were maintained on the test diet during the
period of administration, of the carcinogen, and then three more days.
The mice were then fed control diets for another 113 days when they
were killed, and the number of tumours of the forestomach determined.
The most potent inhibitor was the 2-BHA (the minor isomer present in
BHA) reducing the number of mice with tumours of the forestomach from
90-100% (controls) to 65%, and also markedly decreasing the number of
tumours per mouse (1/6). The major isomer of BHA (3-BHA) reduced the
number of mice with tumours to 79% and almost halved the number of
tumours/mouse (Wattenberg et al., 1980).
In a study when BHA was fed to ICR/Ha mice prior to
administration of benzo(alpha)pyrene it was shown that the formation
of DNA adducts in the forestomach was inhibited to a similar degree
reported for inhibition of tumour formation. However, BHA did not have
any effects on the formation of the adduct in the lung (Ioannou et
Special studies on dermal effects
Application of BHA (up to 20%) in lanolin to the ears of an
inbred strain of black guinea-pigs for a period of six weeks, resulted
in ultrastructural morphological lesions of the epidermis. This
consisted of microinvasion by basal cell pseudopods, with associated
destruction of the superficial connective tissue and fragmentation of
the collagen (Riley & Seal, 1968).
Special studies on effect on gastric mucosa
Simultaneous administration of BHA (50 or 100 mg/kg) with
indomethican resulted in a potentiation of the damaging effect of
indomethican to the gastric mucosa. Under similar conditions butylated
hydroxytoluene reduced damage to the gastric mucosa in a dose-
dependent manner (Krupinska et al., 1980).
In a study using an in situ method of perfusion of intestine to
measure the effect of a number of compounds on intestinal absorption,
BHA at a concentration of 2 mg/ml of perfusion fluid caused a
reduction in the absorption of glucose And methionine, but not butyric
acid. Histological changes were observed in the intestinal mucosa
(Fritsch et al., 1975).
Special studies on reproductive effects
Groups of approximately 10 young adult gilts (Danish Landacre),
were maintained on test diets containing the equivalent of 0, 50, 200,
or 400 mg/kg bw BHA for three weeks and then artificially inseminated.
Feeding of the test diet was continued throughout the study. Body
weight and food consumption were reported weekly. Haematological
analyses were performed two days prior to commencement and on weeks 2,
4 and 14 of the dosing period. Caesarean sections were performed on
day 110. The foetuses were weighed, sexed and autopsied and X-rayed
after removal of the internal organs. A complete autopsy was performed
on the dams. BMA had no effect on food consumption, although the body
weight of dams in the highest dosage group was significantly lower
than controls. BHA had no effect on reproductive performance and there
were no teratogenic effects. A dose-related increase in absolute and
relative organ weight was found for the liver and thyroid. No
compound-related histology was observed in the liver, but the thyroid
gland in test animals showed large follicles with flattened epithelium
containing thyroglobulin. Analysis of fixed thyroid from the control
and high dose animals showed a similar level of iodine in both groups
(0.9 mg/100 g tissue). The iodine content of all diets satisfied the
iodine requirements of the pig (Hansen et al., 1982).
In a lifetime study in rats, 2% BHA produced carcinomas and
papillomas of the forestomach in virtually all the test animals. At
the lower level tested (0.5% of the diet) neoplasia was not observed
but hyperplasia of the forestomach was present in approximately 20% of
the treated rats. A low, but significant incidence of papillomas of
the forestomach was also reported in another lifetime study in the
rat, in which the level of dietary BHT added to the diet was 1%. Other
lifetime studies in rats, in which the BHA was dissolved in oil prior
to addition of the diet and fed at a maximum level of 0.5%, did not
show a carcinogenic response. Dietary BHA caused a rapid onset of
proliferative changes in the forestomach of rats, with an apparent
no-effect level of 0.25%. In these studies the effect was observed
when the BHA was dissolved in oil prior to addition to the diet, or
added dry to the diet.
The significance of these findings to man raises a number of
questions. There is uncertainty as to the relevance of the forestomach
of rat to man since the cell type (squamous epithelium), is found in
the oesophagus of man rather than in the stomach. The pathological
changes observed in the rat appear to commence at the limiting ridge
of the stomach and may represent an area of special cellular activity.
The lack of carcinogenicity at the lowest dose level tested suggests a
very steep dose response and the possibility of a no-effect level. The
lack of genotoxic effects of BHA suggests that the observed
carcinogenic response may be a secondary effect. Studies on the
mechanism of the effect of BHA on the forestomach, as well as studies
to determine hyperplasia of the stomach of species which do not have a
forestomach, will assist in the interpretation of these findings.
There is information to show that BHA can act both as an
inhibitor or enhancer of the carcinogenic effect of certain chemicals.
BHA inhibits benzo(alpha)pyrene-induced forestomach neoplasia in mice.
Preliminary studies have shown that in the rat BHA enhances
forestomach tumorogenesis initiated with methyl-nitrosamine, and
urinary bladder carcinogenesis initiated with N-butyl-N-
(4-hydroxybutyl)-nitrosamine. BHT was inactive in these tests. BHA has
shown a lack of enhancing effects in systems in which BHT is active,
e.g., urethan-induced lung tumours in susceptible mice. Detailed
mechanistic studies will assist in the interpretation of these
effects, and how they can be used to evaluate the safe use of BHA.
A reproduction study in pigs showed that BHA was not teratogenic.
However, it caused a dose-dependent increase in the weight of the
liver and thyroid of the dams. A no-effect level was not established.
The effects on the thyroid have not been reported in lifetime feeding
studies in rats.
The present ADI for BHA provides a wide margin of safety between
the lowest dose fed in the Ito study (a dose that did not result in a
carcinogenic effect) and dietary intake of BHA.
The present ADI is retained.
Estimate of a temporary acceptable daily intake for man
0.05* mg/kg bw.
FURTHER WORK OR INFORMATION
Required by 1986
(1) Studies to show whether or not hyperplasia is induced in the stomach
of species that do not have a forestomach - such as the dog and monkey.
(2) Studies to determine the mechanism involved in the effect of BHA on
(3) Multigeneration reproduction study.
* Group ADI: As BHT and TBHQ, singly or in combination.
Fritsch, P. et al. (1975) Effet de l'anisole de l'anethole, du butyl
hydroxyanisole et du safrole sur l'absorption intestinal chez le
rat, Toxicology, 4, 341-346
Hansen, E. V., Meyer, O. & Olsen, P. (1982) Study on toxicity of
butylated hydroxyanisole (BHA) in pregnant gilts and their
foetuses, Toxicology, 23, 79-83
Ioannou, Y. M., Wilson, A. G. E. & Aderson, M. W. (1982) Effect of
butylated hydroxyanisole, x-angelica lacton, and B-napthoflavone
on benzo(alpha)pyrene: DNA adduct formation in vivo in the
forestomach, lung and liver of mice, Cancer Research, 42,
Ishiwata, H. et al. (1982) Decreases of butylated hydroxyanisole added
to diet for a carcinogenicity test in rats and mice, Bull.
Environm. Contam. Toxicol., 28, 397-383
N. (1983) Unpublished data. Submitted to WHO by the Ministry of Health
and Welfare, Tokyo, Japan
Ito, N. et al. (1982) Induction of squamous cell carcinoma in the
forestomach of F-344 rats treated with butylated hydroxy-toluene,
Gann, 73, 332-334
Krupinska, J. et al. (1980) Antioxidants as agents reducing the
toxicity of indomethican, Acta Biol. Med. Germ., 39, 717-721
Lam, L. K. T. et al. (1981) Effects of 2- and 3-tert-butyl-4
hydroxyanisole on glutathione S-transferase and epoxide hydrolase
additives and sulfhydryl levels in liver and forestomach of mice,
Cancer Research, 41, 3940-3943
Lok, E. et al. (1983) Studies on effect of butylated hydroxyanisole
(BHA) on cell proliferation in rat forestomach. Unpublished
report submitted to the World Health Organization by Health
Protection Branch, Health and Welfare Canada, Canada
MAFF (1983) Unpublished report of the Ministries of Agriculture,
Fisheries and Food
Morgenstern, R. & Dock, L. (1982) A comparison of microsomal
glutathione S-transferase activity in the liver of the mouse and
rat by dietary 2(3)-tert-butyl-4-hydroxyanisole (BHA), Acta
Chemica Scandinavica, B36, 255-279
Reddy, B. S. et al. (1982) Effect of dietary butylated hydroxy-anisole
on methylazoxymethanol acetate-induced toxicity in mice, Fd.
Chem. Tox., 20, 853-859
Riley, P. A. & Seal, P. (1968) Micro-invasion of epidermis caused by
substituted anisoles, Nature, 220, 922-923
Tomii, S. & Aoki, Y. (1982) Chronic toxicity study of BHA. Unpublished
report prepared for the Food Chemistry Division of the Ministry
of Health and Welfare, Japan, March 1977
Wattenberg, L. W., Coccia, J. B. & Lam, L. K. T. (1980) Inhibitory
effects of phenolic compounds on benzo(alpha)pyrene induced
neoplasia, Cancer Research, 40, 2820-2823
Witschi, H. P. (1981) Enhancement of tumor formation in mouse lung by
dietary butylated hydroxytoluene, Toxicology, 21, 95-104