INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY
WORLD HEALTH ORGANIZATION
SAFETY EVALUATION OF CERTAIN FOOD
ADDITIVES AND CONTAMINANTS
WHO FOOD ADDITIVES SERIES: 44
Prepared by the Fifty-third meeting of the Joint FAO/WHO
Expert Committee on Food Additives (JECFA)
World Health Organization, Geneva, 2000
IPCS - International Programme on Chemical Safety
POTENTIAL ALLERGENICITY OF REFINED FOOD PRODUCTS
PEANUT OILS AND SOYA BEAN OILS
First draft prepared by Dr J.B. Greig
Joint Food Safety & Standards Group, Department of Health, London,
Protein content and immunoreactivity of crude and refined oils
Observations in humans
Soya bean oil
The allergenicity of foodstuffs has not previously been
considered by the Committee. The Codex Committee on Food Labelling
considered allergens in food on a number of occasions after 1993 and
issued a list of foods and food ingredients known to cause allergy.
That list, with modifications, was revised at an FAO Technical
Consultation in 1995 (FAO, 1995). After debate in the Codex Committee
(FAO, 1998), the list was forwarded at Step 8 for adoption by the
Codex Alimentarius Commission. The Joint FAO/WHO Expert Committee on
Food Additives was asked by the Codex Committee on Food Labelling
(FAO, 1998) to provide scientific advice on this issue, including the
development of criteria for identifying products of foods on the list
for which labelling of the food source is not necessary. An Expert
Panel was convened to assist the Committee (see Annex 6 of the report
of the present meeting: Annex 1, reference 143).
In its report, the Panel developed criteria for excluding a
product from the list, which included: (i) evidence that a clinical
study in which neither the investigators not the participants were
aware of who received the test substance and who received placebo
(double-blind) had been conducted in which challenge with the specific
product did not elicit allergic reactions in a group of patients with
clinical allergy to the parent foodstuff, and (ii) the existence of
specifications for the product and its manufacturing process that
demonstrate that the process can yield a consistently safe product.
The Panel identified only two foodstuffs that it considered may
currently fulfil the criteria for inclusion on a list of products of
foodstuffs for which labelling of the allergen-containing food source
is unnecessary: refined peanut oil and refined soya bean oil. The
Panel recommended that these be considered by the Committee at its
2. BIOLOGICAL DATA
2.1 Protein content and immunoreactivity of crude and refined oils
Peanut oil is commonly marketed as a refined oil or in a crude
form, sometimes called 'gourmet' oil, which may be used for flavouring
purposes. The refined oil is produced by a process which involves
degumming, neutralization, bleaching, filtration, and deodorization to
achieve a concentration of free fatty acids of < 0.1%. Two
alternative processes can be used to achieve this specification, one
physical and the other chemical. A code of practice for the production
and labelling of peanut oil in connection with peanut allergy was
adopted in the United Kingdom and was subsequently adopted by the
European Association FEDIOL, to be implemented no later than 1 January
1998. All refined vegetable oils undergo the same degree of processing
(VCH Veriagsgesellschaft mbH, 1987; Seed Crushers' and Oil Processors'
The mean protein content of peanut oil is reduced at each stage
of the refining process described above. Although the results differed
depending on whether the Lowry assay or the Pierce Micro bicinchoninic
acid assay was used for protein analysis, there was a consistent
downward trend with each analytical procedure. The mean protein
content in extracts of oil prepared by two separate procedures and
analysed by each assay were 187 µg/ml of crude peanut oil; 60 µg/ml of
oil after alkali refining, neutralization, and washing; 15 µg/ml of
oil after bleaching and filtering; and 2.2 µg/ml of oil after
deodorization (Skinner & Haynes, 1998).
The presence of soya bean proteins in phosphate-buffered saline
(pH 7.2) extracts of eight soya bean oils of unspecified origin was
tested by enzyme-linked immunosorbent assay. Three of the oil extracts
showed measurable quantities, two contained less than 50% of the
maximum concentration tested, and three were considered to contain no
protein in comparison with negative controls (Porras et al., 1985).
The protein in a refined, deodorized, alkali-refined, and
steam-refined peanut oil, a refined, deodorized, alkali-refined, and
steam-deodorized soya bean oil, and an alkali-refined and bleached
soya bean oil was extracted by chromatography on a diethylaminoethyl
cellulose column. After elution of non-polar, polar, and acid lipids,
the proteins were eluted with 1 mol/L aqueous sodium chloride and 0.2
mol/L aqueous sodium hydroxide. Only an acid hydrolysate of the sodium
hydroxide extract of the alkali-refined and bleached soya bean oil had
a detectable content of amino acids, estimated as being equivalent to
0.96 µg/g of oil (Tattrie & Yaguchi, 1973).
Extracts of soya bean oil, crude and processed by unspecified
means, and of five types of processed peanut oil from various sources
were extracted with 0.15 mol/L phosphate-buffered saline (pH 7.4). The
protein content of the extracts was measured with Bradford dye-binding
reagent, and the lectin-like activity was assayed by agglutination of
human erythrocytes. The crude and processed soya bean oils had protein
contents of 1900 and 720 µg/kg and lectin-like activities of 860 and
45 µg/kg, respectively, while the protein content and lectin-like
activity of the peanut oils ranged from 120 to 580 µg/kg and 26 to 55
µg/kg, respectively (Klurfeld & Kritchevsky, 1987).
Oils from walnuts, almonds, hazelnuts, and macadamia nuts and
refined and unrefined peanut oil were each extracted with 0.2 mol/L
ammonium bicarbonate solution. The protein concentrations of the
aqueous extracts were measured and binding to immunoglobulin (Ig) E
was assayed by slot-blot and western immunoblotting. The IgE was
derived from atopic sera in pooled samples from a serum bank
consituted from persons with food allergies. Standard extracts of the
nut and legume proteins were prepared from nuts and peanut oils from
various sources. The analytical results indicated that those oils that
had undergone the least processing at lower maximum temperatures had
higher protein concentrations. Of the peanut oils, those that were
unrefined elicited immunoreactive bands in the western immunoblotting
assay, whereas such bands were absent from the lanes corresponding to
the refined and deodorized oils. Confirmation of this observation was
provided by the slot-blot immunoassay (Teuber et al., 1997).
Various commercially available refined peanut oils from the
European market were extracted with 0.1 mol/L sodium bicarbonate (pH
8.0), and the protein content of the extract was analysed by the
Pierce bicinchoninic acid method after further complex purification
steps. Whereas the crude, cold-pressed oil had a protein content of
3.4 µg/g, a neutralized oil contained 0.2 µg/g and the protein
content of five commercial refined oils ranged from 0.1 to 0.2 µg/g.
Although protein extracts released histamine from the leukocytes of
peanut-allergic patients, as did a micellar casein-peanut oil
preparation in most instances, the results were not totally concordant
and did not completely correlate with the presence of a positive
reaction to a double-blind challenge with peanut oil. An
immunoblotting experiment with protein extracted from two of the oils
revealed that the extract could inhibit binding of IgE from a patient
allergic to peanuts. Western blot analysis of the protein extract of
refined oil showed that the reactive species was an 18-kDa protein
with a relative molecular mass similar to that of one of the major
peanut allergens (Olszewski et al., 1998).
Since the Lowry and bicinchoninic acid methods have been found to
be subject to interference, fully refined peanut and soya bean oils
were assayed for protein content by aqueous extraction and amino acid
analysis of an acid hydrolysate. The peanut oil had a protein content
of 0.83 mg/kg, and three samples of soya bean oil had protein contents
of 0.08-0.22 mg/kg (Institute of Shortening and Edible Oils, 1999).
2.2 Observations in humans
The allergenicity of edible oils has been reviewed (Hefle &
2.2.1 Peanut oil
Seven male and three female patients aged 17-45 years with a
known history of immediate hypersensitivity reaction to peanuts were
recruited into a double-blind, placebo-controlled food challenge study
in which a commercial peanut oil containing no detectable protein and
olive oil (the placebo) were administered in a cross-over design. All
the patients gave an immediate response in skin-prick tests with two
crude peanut extracts, and all the patients had elevated serum IgE
antibody levels to peanut allergens in a radioallergo-sorbent test
with a crude peanut extract. The serum from nine of the 10 patients
showed two to 11 times more binding than negative control serum to the
major peanut allergen, peanut-1 or Ara h I, by the
Neither peanut oil nor olive oil elicited any reaction in
skin-prick tests. In food challenge tests performed on two separate
occasions at least 14 days apart, the patients being randomly
allocated to receive peanut oil or olive oil on the first day, none of
the patients experienced any adverse immediate or delayed reaction
when challenged with sequential doses of 1, 2, or 5 ml of peanut or
olive oil in gelatin capsules (Taylor et al., 1981).
In a double-blind, placebo-controlled food challenge study in
which four patients who were sensitive to peanuts were exposed to 30
ml of commercial peanut oil, no reaction was recorded (Bock & Atkins,
1989), but no additional details were given.
Vitamin D supplementation may be administered in an oil
preparation in early infancy. After skin-prick testing of 122 children
aged 7-60 months who had been referred to an allergy clinic, the
children were classified according to whether they had received a
vitamin D preparation without peanut oil, one containing peanut oil
that had been administered monthly, or one containing peanut oil that
had been administered daily. Although the groups did not differ in
respect of allergic status, statistically significantly children more
showed a positive reaction to peanut if they had been exposed to a
peanut oil-containing vitamin preparation. The peanut oil used in the
vitamin preparations was not specified (de Montis et al., 1993).
In a study in France, two male and two female infants aged 4-13
months who had received a diagnosis of atopic dermatitis were found to
react to peanut allergens during skin-prick testing or labial
challenge with peanut extract, peanut butter, or peanut oil. In a
single blind oral challenge test with peanut oil, the infants reacted
with a rash to doses of 1 or 5 ml of peanut oil. In each case, the
infant was receiving a formula containing peanut oil in such an amount
that it contributed 67 or 80% of the lipids. In general, the condition
of the children improved when they were placed on an exclusion diet,
but the posssibility that concurrent medication contributed to the
recovery was not ruled out (Moneret-Vautrin et al., 1994). An earlier
letter from the same group may have concerned two of these infants
(Moneret-Vautrin et al., 1991).
The same group reported the results of studies in a group of six
male and five female allergy patients aged 2-17 years. Positive
results in skin-prick tests were found for 10/10 patients challenged
with native peanut, 9/9 with commercially available peanut protein
extracts, 0/9 with refined peanut oil, and 5/7 with protein extracts
of crude or refined peanut oil. Only four of the 11 patients reacted
adversely in a double-blind food challenge with peanut oil (not
specified whether crude or refined) (Olszewski et al., 1998).
A randomized, double-blind, cross-over challenge study was
conducted to determine the allergenicity of crude and refined peanut
oil in a group of 15 men and 54 women of a mean age of 26 years
(range, 14-48 years) who had participated in a questionnaire study of
peanut allergy. Each individual was subjected to skin-prick tests with
peanut extract and with the crude and refined peanut oils. Those 62
individuals who gave positive reactions in the skin-prick test with
peanut (weal diameter equal to or greater than that elicited by 1%
histamine) were tested on the same day by oral challenge with the
oils, which were administered in a random order determined by a person
who was not involved in assessing the participants' reactions, at
increasing doses of 1, 5, and 10 ml. The flavour was disguised, six
subjects being offered the oil with bread, one with soya milk, and the
remainder with rice pudding. The onset of symptoms was monitored
during an observation period of 10-15 min between doses. If a reaction
occurred, at least 1 h was allowed to elapse before administration of
the next dose.
None of the subjects reacted to the refined oil, but six subjects
reacted to the crude peanut oil. One subject reacted with wheeze to a
dose of 1 ml of crude oil, four subjects reacted with oral itch,
throat itch, or lip swelling to a dose of 5 ml of crude oil, and one
subject reacted with oral itch to the 10-ml dose of crude oil. The
reactions of four of these six subjects were subjective, with no
observable or measurable sign of reaction. When the 58 individuals who
showed reactions on skin-prick testing were challenged with peanuts,
two ate a cumulative dose of 32 peanuts with no evidence of reaction.
The other 56 subjects all responded to the challenge with reactions
ranging from severe (after labial challenge) to mild (after
consumption of four nuts). The study authors suggested that refined
peanut oil does not induce an allergic reaction and that the true
incidence of measurable reaction to crude peanut oil among patients
who are allergic to peanuts may be 3.3% rather than the 10% suggested
by the reactions of six individuals, which in four instances were
possibly psychologically mediated (Hourihane et al., 1997).
2.2.2 Soya bean oil
A person who worked and lived near a soya bean mill experienced
asthma which was dependent on the wind direction. He gave a positive
reaction to skin scratch tests with soya bean products, including soya
bean oil of unspecified grade. The report stated that when the oil was
'filtered through stone' it ceased to cause a reaction (Duke, 1933).
A study of the allergenicity of soya bean food products,
including two brands of soya bean oil, was carried out in children who
were passively sensitized by injection of serum from a patient
allergic to soya beans. The patient, a woman, had experienced shock
and collapse after eating food containing a soya bean filler. She was
presumed to have been sensitized by inhalation of dust from a soya
bean processing plant opposite her home. The presence of antibodies to
soya bean in her serum was confirmed by passive sensitization of eight
adult volunteers who were challenged one to seven days later with soya
bean extract at the site of sensitization. All gave a positive
Eight children (sex and age unspecified) were injected
intradermally with the antibody-containing serum at two sites on the
forearm. Before breakfast 24 and 72 h later, they drank 40-55 g of a
soya bean oil. One brand of the oil was described as crude, but
details of the other brand and the numbers of children ingesting each
oil were not specified. No reactions were observed at the sensitized
skin sites over the next 1-24 h. All of the children had positive
reactions after ingestion of a suspension of soya bean flour in water
(Ratner et al., 1955).
Three men and four women aged 18-63 years who were sensitive to
soya beans were recruited into a double-blind placebo-controlled
cross-over study of the allergenicity of soya bean oil. The time since
the last exposure of the subjects that had resulted in an allergic
reaction ranged from < 1 to 10 years. The oils tested were partially
hydrogenated, unhydrogenated, and cold-pressed soya bean oils; the
placebo was an olive oil. The sequence of administration of the oils
during the study was randomized. Before the start of the study, all
the subjects reacted to a skin-prick test with soya bean extract, but
none gave a positive reaction to a skin-prick test with the test oils.
The percent binding of serum IgE antibody to soya bean allergens,
assessed in a radioallergosorbent test in six of the seven subjects,
was 230-2800% that of a pooled control serum. On the second day of the
study, the subjects were challenged with 2, 5, or 8 ml of the assigned
oil administered in gelatin capsules, these doses being equivalent to
the amount that might be ingested during a meal. Each dose was
followed by a 30-min observation period. Challenges to each of the
other oils were made after intervals of at least six days. None of the
subjects experienced an immediate or delayed adverse reaction, whether
typical or atypical of an allergic reaction, to any of the soya bean
oils (Bush et al., 1985).
The Committee recognized that the allergenicity of vegetable oils
is highly dependent on the processes used to extract and then refine
the oils. It was aware that several steps are involved in the refining
process and that different producers may use variations of the basic
procedures. In addition, in any clinical trial of the oils, the mode
of administration, the allergic sensitivity of the subjects to the
source material, and use of double-blind protocols can affect the
outcome of the trial.
The Committee was aware of studies of challenges of
peanut-sensitive individuals with various grades of peanut oil, all
involving a double-blind procedure. In a study from the United States
involving 10 male and female patients with known sensitivity to
peanuts, all gave a positive reaction in skin-prick tests with peanut
extracts and had elevated serum titres of antibodies to peanut
allergens. A cross-over challenge with commercial peanut oil and olive
oil did not elicit adverse reactions, although the Committee noted
that use of gelatin capsules to administer the oils may have masked
any reactions of the lips and oral cavity.
In a study in France, 11 children with symptoms possibly due to
allergies were found to react to skin-prick tests with peanut or
peanut protein extracts. Four of the patients reacted to a
double-blind oral challenge with peanut oil. The origin or grade of
the peanut oil used was not defined, and the Committee recognized that
it may have been obtained before adoption of a revised code of
practice for the refining of vegetable oils by the continental
European industry. Earlier studies of French infants suggested that
peanut oils used as a vitamin carrier or in infant formulas may have
contained allergenic proteins.
A randomized, double-blind, cross-over challenge study with crude
and refined peanut oils involved a group of 62 patients in the United
Kingdom who had reacted to skin-prick tests with peanut suspensions.
None of the subjects reacted to challenge with refined peanut oil,
although six reacted to the crude oil. Sixty of the 62 also reacted to
an open oral challenge with peanuts. The Committee considered that the
study was well designed and had good statistical power and recognized
the value of the confirmation of the sensitivity of the subjects to
peanuts after the double-blind challenge had been completed. Although
the study provided adequate evidence for lack of allergenicity of the
oil used, appropriate descriptions of the manufacturing process and
the consequent specifications of the oil were not provided and the
results could not be extrapolated to other oils.
A double-blind, placebo-controlled, cross-over challenge study of
the allergenicity of hydrogenated, partially hydrogenated, and
cold-pressed soya bean oils was conducted in a group of seven
individuals who had experienced allergic reactions after exposures
that had occurred up to 10 years previously. All had positive
reactions to a skin-prick test with soya bean extract. The titres of
serum immunoglobulin E binding to soya bean proteins were increased in
six of the seven patients. None of the subjects reacted to increasing
volumes of any of the oils, although the Committee noted that use of
gelatin capsules to administer the oils may have masked any reactions
of the lips and oral cavity. Although the study provided some evidence
that the oil used was not allergenic, appropriate descriptions of the
manufacturing process and the consequent specifications of the oil
were not provided, and the results could not be extrapolated to other
The Committee noted the absence of clear descriptions of the
processes that had been used to refine the peanut and soya bean oils
tested. Additionally, comparable data on the protein content of those
oils that were clinically tested were not available. Furthermore, the
Committee expressed reservations about the quality of the analytical
procedures used and the lack of validation of the methods to determine
the concentrations of residual protein in the oils. In view of these
considerations, it concluded that distinct processes that would
consistently yield safe products have not been defined.
The Committee therefore indicated that the results of studies of
representative refined peanut and soya bean oils would be required for
a full evaluation. Such studies should provide extensive information
on a wide range of oils representing refining procedures throughout
the world. Full descriptions of the refining process used and evidence
for lack of allergenicity of these oils as determined by appropriately
designed clinical studies should be provided. Evidence for the nature
and quantities of protein in the oils would be essential for ensuring
the representative nature of the oils tested.
Bock, S.A. & Atkins, F.M. (1989) The natural history of peanut
allergy. J. Allergy Clin. Immunol., 83, 900-904.
Bush, R.K., Taylor, S.L., Nordlee, J.A. & Busse, W.W. (1985) Soybean
oil is not allergenic to soybean-sensitive individuals. J. Allergy
Clin. Immunol., 76, 242-245.
Duke, W.W. (1933) Soy bean as a possible important source of allergy.
J. Allergy, 5, 300-302.
FAO (1995) Report of the FAO Technical Consultation on Food
Allergies, Rome, Italym 13-14 November 1995, Rome, Food and
Agricultural Organization of the United Nations.
FAO (1998) Report of the Twenty-sixth Session of the Codex Committee
on Food Labelling, Ottawa, 26-29 May 1998 (unpublished FAO document
ALINORM 99/22), Rome, Food and Agricultural Organization of the United
Hefle, S.L. & Taylor, S.L. (1999) Allergenicity of edible oils.
Food Technol., 53, 62-70.
Hourihane, J.O'B., Bedwani, S.J., Dean, T.P. & Warner, J.O. (1997)
Randomised, double blind, crossover challenge study of allergenicity
of peanut oils in subjects allergic to peanuts. Br. Med. J., 314,
Institute of Shortening and Edible Oils (1999) The protein content of
peanut and soybean oils. Unpublished study. Submitted to WHO by R.M.
Reeves, Institute of Shortening and Edible Oils, Inc., Washington DC.
Klurfeld, D.M. & Kritchevsky, D. (1987) Isolation and characterisation
of lectins from vegetable oils. Lipids, 22, 667-668.
Moneret-Vautrin, D.A., Hatahet, R., Kanny, G. & Ait-Djafer, Z. (1991)
Allergenic peanut oil in milk formulas. Lancet, 338, 1149.
Moneret-Vautrin, D.A., Hatahet, R. & Kanny, G. (1994) Risks of milk
formulas containing peanut oil contaminated with peanut allergens in
infants with atopic dermatitis. Pediatr. Allergy Immunol., 5,
de Montis, G., Gendrel, D., Chemillier-Truong, M. & Dupont, C. (1993)
Sensitisation to peanut and vitamin D oily preparations. Lancet,
Olszewski, A., Pons, L., Moutété, F., Aimone-Gastin, I., Kanny, G.,
Moneret-Vautrin, D.A. & Guéant, J.L. (1998) Isolation and
characterization of proteic allergens in refined peanut oil. Clin.
Exp. Allergy, 28, 850-859.
Porras, O., Carlsson, B., Fällström, S.P. & Hanson, L.Å. (1985) Int.
Arch. Allergy Appl. Immunol., 78, 30-32.
Ratner, B., Untracht, S., Crawford, L.V., Malone, H.J. & Retsina, M.
(1955) Allergenicity of modified and processed foodstuffs. V. Soybean;
influence of heat on its allergenicity; use of soybean preparations as
milk substitutes. Am. J. Dis. Child., 89, 187-193.
Seed Crushers' and Oil Processors' Association (1999) Submission to
WHO dated 4 March 1999.
Skinner, J.M. & Haynes, C. (1998) Measurement of protein content in
various oils during stages of refining. Unpublished report H6672 from
Leatherhead Food Research Association, Leatherhead, United Kingdom.
Submitted to WHO by the Seed Crushers' and Oil Processors'
Tattrie, N.H. & Yaguchi, M. (1973) Protein content of various
processed edible oils. J. Inst. Can. Sci. Technol. Aliment., 6,
Taylor, S.L., Busse, W.W., Sachs, M.I., Parker, J.L. & Yunginger, J.W.
(1981) Peanut oil is not allergenic to peanut-sensitive individuals.
J. Allergy Clin. Immunol., 68, 372-375.
Teuber, S.S., Brown, R.L. & Haapanen, L.A.D. (1997) Allergenicity of
gourmet nut oils processed by different methods. J. Allergy Clin.
Immunol., 99, 502-507.
VCH Veriagsgesellschaft mbH (1987) Fats and fatty oils. In: Ullman's
Encyclopedia of Industrial Chemistry, 5th Ed., Weinheim, Vol. A10,