Caballero B. (ed.) Encyclopaedia of Food Science, Food Technology and Nutrition. Ten-Volume Set

Подождите немного. Документ загружается.

0016 In addition to fish, A. hydrophila may be obtained

from the feces of a large range of both wild and

domestic animals. In these studies, the animals were

considered to be healthy. As with fish, poor process-

ing techniques and poor hygienic practices during

slaughter, butchery, dressing, and processing of

meat and poultry products may contribute to the

incidence of this bacterium in raw-flesh foods. (See

Poultry: Chicken; Ducks and Geese; Turkey.)

0017 A. hydrophila is not considered to be a normal

inhabitant of the microbial flora of the human gut

and its presence is usually transient. The fecal

carriage rate for healthy individuals is usually 0–5%.

Although asymptomatic carriers could serve as

vectors for transmission of this bacterium (e.g.,

food handlers) no documented evidence of this is

available.

0018 As a consequence of this bacterium in human and

animal feces, A. hydrophila can be commonly isol-

ated from sewage samples and drainage from animal

lairages. These may also contribute to the incidence in

water samples.

Occurrence in Foods

0019 A. hydrophila may be isolated from a very wide range

of foodstuffs. It is not surprising that it is common (up

to 100% of samples are positive) in a wide range of

seafood, including fish (both wild and farmed),

shrimps (raw and cooked), oysters, crabs, and scal-

lops. The bacterium is also common in raw meats and

may be isolated from a high proportion (50–100%)

of red meat and poultry samples. Cooked meats may

occasionally contain A. hydrophila, indicating that

postprocess contamination has occurred. In fresh

(and cooked) meats and poultry this bacterium is

generally present in low numbers (less than 100 per

gram) but may greatly increase during storage at chill

temperatures, particularly when vacuum or modified-

atmosphere packaging is used. In the past, A. hydro-

phila has been considered as a spoilage bacterium in

these foods. Up to 50% of raw milk samples may

contain. A. hydrophila and it has also been isolated

from pasteurized milk, cream, and some dairy prod-

ucts. The incidence in pasteurized products is much

lower than in raw milk and represents postprocess

contamination. A wide variety of vegetable products

(including parsley, spinach, celery, alfalfa, bean-

sprouts, broccoli, lettuce, and root vegetables) may

contain A. hydrophila and the incidence is usually

between 5 and 40% of positive samples.

0020 In addition to the above foods, A. hydrophila has

been isolated from drinking-water samples and also

from bottled mineral waters, although levels are gen-

erally low (less than 10 per ml). Studies have shown

that when A. hydrophila was inoculated in mineral

water, it was able to survive for 100 days.

0021The high levels of A. hydrophila found in some

foods have also been associated with spoilage,

where it may produce pungent off-odors. In addition

the organisms can produce extracellular lipase or

protease enzymes which are mildly heat-resistant

and still active at temperatures of 0 to 2



Factors Affecting Growth

0022Of particular concern with the incidence of A. hydro-

phila in foods is the ability of the bacterium to grow

at refrigeration temperatures. This bacterium may

grow at temperatures as low as 0



C. Consequently,

refrigerated storage alone cannot be relied upon to

inhibit the growth of A. hydrophila completely. On

occasions, high levels of the bacterium have been

isolated from chilled foods which had been stored

for prolonged periods. Reducing the storage tempera-

ture, however, will reduce the rate of growth of A.

hydrophila and refrigeration may act in combination

with other preservation factors to inhibit growth. The

optimum and maximum temperatures for growth are

28–30 and 42



C, respectively.

0023A. hydrophila is not heat-resistant and has a

decimal reduction time (D value) of between 2.2

and 6.6 min at 48



C. Consequently, this bacterium

will be readily destroyed by processes such as milk

pasteurization and frankfurter processing. It is im-

portant to insure that heat processes are carefully

controlled to insure microbial destruction.

0024This bacterium is not a good competitor with many

other bacteria present in raw foods, but may reach

high levels if the competing bacterial flora has been

removed by heating or its growth is retarded. For

example, vacuum packaging may prevent the growth

of the main spoilage bacteria of raw meats (i.e., Pseu-

domonas species), but A. hydrophila may continue to

grow and even dominate the bacterial flora. The use

of modified atmospheres containing 100% carbon

dioxide will inhibit growth of A. hydrophila, but a

small decrease in this level may permit growth. (See

Chilled Storage: Use of Modified-atmosphere Pack-

aging; Packaging Under Vacuum.)

0025A. hydrophila is not an acid-tolerant bacterium and

at chill temperatures (less than 5



C) it will grow only

poorly, if at all, at pH values less than 6.0. Therefore,

growth would not be expected in even mildly acidic

refrigerated products. At lower pH values, the bacter-

ium will slowly die, although this may require in

excess of 20 days. When stored at 30



C the minimum

pH value for growth has been variously reported to be

between 4.0 and 5.0. As with other bacteria, organic

acids (e.g., acetic, lactic, and citric acids) which occur

AEROMONAS

in foods are more effective at inhibiting growth than

are mineral acids (e.g., hydrochloric acid).

0026 The maximum level of salt permitting growth

depends on the other conditions, but is 6% (w/v)

when other conditions are optimal. With refrigerated

samples (less than 5



C), growth is poor when the salt

concentration exceeds 2% (w/v).

0027 Irradiation, at the doses proposed for foods (i.e.,

10 kGy), would successfully eliminate A. hydrophila

from foods. The decimal reduction value is reported

as 0.14–0.22 kGy. (See Irradiation of Foods: Basic

Principles.) The organism is sensitive to many of the

novel technologies being developed, e.g., high-pres-

sure processing.

0028 Little is known about the effect of freezing on

A. hydrophila, but this process cannot be relied upon

to insure the destruction of the bacterium and it has

been isolated from a wide variety of frozen foods.

0029 The presence of A. hydrophila in chlorinated water

suggests that the bacterium is resistant to this biocide.

Chlorination can be a very effective means of control-

ling A. hydrophila, but the levels used must be care-

fully controlled. Other disinfectants, including those

commonly used in the food industry, and ultraviolet

treatment of water may also be effectively used to

control A. hydrophila. Overall, A. hydrophila is not

a hardy bacterium and is readily controlled by many

of the processes used in the food industry. Care is

needed to insure that postprocess contamination is

minimized.

Detection Methods

0030 A. hydrophila is not difficult to grow and produces

good growth on many laboratory media. A wide

variety of methods are used for this bacterium. Most

of the media used have been developed from those

used for the detection of Enterobacteriaceae or Vibrio

species. It is likely that A. hydrophila has been mis-

identified as an enteric bacterium and so its incidence

and clinical significance will have been underesti-

mated.

0031 The bacterium may be isolated using quantitative

or qualitative procedures. With quantitative methods,

the sample is usually inoculated directly on a solid

selective agar medium. With qualitative methods, the

sample is usually placed in a liquid-selective enrich-

ment medium (to allow multiplication of A. hydro-

phila) prior to inoculation of the agar medium.

Incubation temperatures used for both liquid and

solid media range between 25 and 37



C, although



C is being increasingly accepted.

0032 A variety of agar media have been used for food

and water samples. Most of these contain selective

agents (usually based on ampicillin, bile salts, or

ethanol) and differential agents (e.g., starch, dextrin,

xylose, amino acids). With liquid media, alkaline

peptone water and ampicillin broths are most widely

used. Both liquid and solid media containing ampicil-

lin have been shown to be very useful, although a few

strains are sensitive to this antibiotic.

0033At present, there is no general consensus on the best

media for A. hydrophila and it is likely that the use of

multiple solid and/or liquid media will be required for

optimal recovery of this bacterium. The individual

motile Aeromonas species can be separated using

conventional biochemical tests. Additional tests to

identify biotypes and serotypes are not widely used

for food isolates.

0034Given the complexity of the taxonomy of the genus

and that some isolates, even of A. hydrophila, are not

pathogenic, a variety of molecular techniques have

been proposed for the detection and/or identification

of this organism in clinical, veterinary, and food

situations. These include polymerase chain reaction

(PCR) detection of the enterotoxin and hemolysin

genes, restriction fragment length polymorphism

(RFLP)-PCR to identify the species, and random-

amplified polymorphic DNA (RAPD)-PCR probes

for detection and identification.

See also: Chilled Storage: Use of Modified-atmosphere

Packaging; Packaging Under Vacuum; Dairy Products –

Nutritional Contribution; Fish: Spoilage of Seafood;

Gums: Food Uses; Irradiation of Foods: Basic

Principles; Poultry: Chicken; Ducks and Geese; Turkey;

Shellfish: Contamination and Spoilage of Molluscs and

Crustaceans; Vibrios: Vibrio cholerae; Vibrio

parahaemolyticus; Vibrio vulnificus

Further Reading

Abeyta C and Wekell MM (1988) Potential sources of Aero-

monas hydrophila. Journal of Food Safety 9: 11–22.

Cahill MM (1990) Virulence factors in motile Aeromonas

species. Journal of Applied Bacteriology 69: 1–16.

Janda JM and Duffey PS (1988) Mesophilic aeromonads in

human disease: current taxonomy, laboratory identifica-

tion, and infectious disease spectrum. Reviews of

Infectious Disease 10: 980–997.

Kaznowski A (1998) Identification of Aeromononas strains

of different origin to the genomic species level. Journal

of Applied Microbiology 84: 423–430.

Palumbo SA and Buchanan RL (1988) Factors affecting

growth or survival of Aeromonas hydrophila in foods.

Journal of Food Safety 9: 37–51.

Stelma JR (1989) Aeromonas hydrophila. In: Doyle MP

(ed.) Foodborne Bacterial Pathogens, pp. 1–19. New

York: Marcel Dekker.

Tompkins DS, Hudson MJ, Smith HR et al. (1999) A study

of infectious intestinal disease in England: microbio-

logical findings in cases and controls. Communicable

Disease and Public Health 2: 108–113.

AEROMONAS

AFLATOXINS

M O Moss, University of Surrey, Guildford, UK

Background

0001 The aflatoxins are a group of carcinogenic, acutely

toxic mold metabolites that may be produced in a

range of foods, animal feeds, and the raw materials

used in their manufacture. The discovery, physico-

chemical properties, occurrence, toxicology, control,

analysis, and significance will be discussed.

Discovery

0002 A bird food catalog for the year 2000 has a letter to

customers advising that adverse climate conditions

led to a very poor year for the groundnut crop,

resulting in unacceptable levels of aflatoxin contam-

ination, and that there would be limited supplies for

bird foods this year. Indeed, this particular supplier

withdrew all stocks of peanuts. In 1998, the Euro-

pean Union set a limit of 2 mgkg

1

for aflatoxin B

foods for human consumption, a regulation that was

included in British legislation in 1999. These two

observations make it clear that aflatoxin is an import-

ant contaminant, and it is essential to understand its

origin, occurrence, and significance.

0003 In 1960, there were reports of large-scale deaths of

turkey poults and other poultry in England, which

were initially considered to be due to a viral infection

and were called turkey X disease. The initial symp-

toms were a loss of appetite, feeble fluttering, and

lethargy, death usually occurring within a week of

the appearance of symptoms. Autopsy showed hem-

orrhages and necroses of the liver. It soon became

clear that the birds were poisoned by a contaminant

in the pelleted feeds that formed the major part of

their food. The toxic component was shown to be

associated with groundnut meal contaminated with

the green spored mold Aspergillus flavus, and it was

soon demonstrated that the mold was secreting a

toxin that was called aflatoxin to reflect its source.

As chromatographic analyses improved, it was shown

that aflatoxin is a family of compounds that were

labeled according to the color perceived when thin-

layer chromatography (TLC) plates were examined

under UV light; aflatoxins B

(blue), G

, and G

(green), of which aflatoxin B

is the most toxic.

Chronic toxicity studies in rats demonstrated that,

for this species, aflatoxin was amongst the most

carcinogenic compounds known at the time, and the

study of aflatoxin became truly international, gener-

ating an enormous literature very rapidly.

0004The chemical structures were illucidated within a

few years of the initial outbreak of turkey X disease

(Figure 1), and, with an increased understanding of

their occurrence and properties, it was appreciated

that at least two animal diseases described earlier

were due to the presence of aflatoxins. A serious

liver disorder of dogs known as canine hepatitis and

the presence of liver carcinomas in rainbow trout in

fish farms, where protein rich feeds based on slaugh-

ter house offal were replaced with plant proteins

based on cotton seed meal. Indeed, it is certain that

aflatoxins, either alone or in the presence of other

mycotoxins, have caused many acute and chronic

illnesses in farm animals and several in humans.

One of the best documented outbreaks of acute afla-

toxicosis in humans occurred in India in 1975. Nearly

1000 people were ill, and almost 100 died following

the consumption of contaminated corn.

Physicochemical Properties

0005The aflatoxins are white crystalline solids that are

optically active and have a strong absorbance at

about 365 nm with a fluorescence emission of 415–

450 nm, depending on the solvent or physical status

(Table 1). Fluorescence is especially strong when the

aflatoxins are absorbed on silica gel, making this a

Aflatoxin B

Aflatoxin G

OCH

fig0001Figure 1 Structures of the four most common naturally occur-

ring aflatoxins.

66 AFLATOXINS

sensitive means of detection on thin-layer chromato-

grams. The aflatoxins are soluble in a range of or-

ganic solvents such as chloroform, ethanol, methanol,

and acetone, and insoluble in lipophylic solvents such

as hexane, petroleum ether, and diethyl ether.

Occurrence

0006 Because so many compounds have similar chromato-

graphic and fluorescent properties to aflatoxins, there

were many early reports of their production by molds

from several genera including Penicillium and Rhizo-

pus. However, the only confirmed sources of afla-

toxins are a small group of closely related aspergilli,

namely Aspergillus flavus, A. parasiticus, A. nomius,

and A. ochraceoroseus. There are also reports that an

isolate of A. tamarii produces aflatoxin B

, although

no other isolates of this species have been shown to

produce aflatoxins. This isolate has subsequently

been referred to a new species, A. pseudotamarii.By

far the most commonly isolated aflatoxigenic species

are A. flavus, which produces aflatoxins B

and B

and A. parasiticus, which produces all four aflatoxins

shown in Figure 1. On a world-wide basis, perhaps

35% of A. flavus isolates are aflatoxigenic, whereas

almost all strains of A. parasiticus produce aflatoxins.

0007 The most commonly implicated commodities are

corn and peanuts, although aflatoxins have been

detected in a much wider range of foods. Table 2

provides the results of surveys carried out during the

last few years in a number of different countries.

0008 The highest levels of aflatoxin contamination are

always associated with postharvest spoilage, when

commodities are stored with an inappropriate mois-

ture content and temperature. However, aflatoxin

contamination is not simply a problem of poor

storage but can occur in the field before the crop is

harvested. The spores of these species of Aspergillus

can land on the stigma of the developing plant, ger-

minate and penetrate to the immature seed tissue just

as if they were pollen grains. The mold can establish

an endophytic growth within the tissues of the

plant without causing any perceptible harm to

the plant. Indeed, if the crop continues to develop in

a healthy manner and is harvested and correctly

stored, there may be no detectable aflatoxins pro-

duced, despite the presence of viable mycelium in

the plant tissue. However, if the plant is subject to

some form of stress, and drought stress is the most

common, then low, but significant, levels of aflatoxin

tbl0001 Table 1 Some physico-chemical properties of the aflatoxins

Aflatoxin Formula Molecular weight Melting point (



C) Opticalrotation [a]

(CHCl

) l

max

(nm) e

312.063 268–269 558



223 25 600

265 13 400

362 21 800

314.0790 287–289 430



220 20 500

265 12 700

363 24 000

328.0582 247–250 556



243 11 500

257 9 900

264 10 000

362 16 100

330.0739 230 454



217 28 000

245 12 900

265 11 200

365 19 300

328.0582 299 280



226 23 100

265 11 600

357 19 000

tbl0002Table 2 Reports of the occurrence of aflatoxins during 1991–

1998

Commodity Country Year Incidence

(%)

Range

(mgkg

1

)

Almonds USA 1993 1 tr–372

Brazil nuts USA 1993 17 tr–619

Chillies Pakistan 1995 66 1–79.9

Corn (maize) Argentina 1996 20 5–560

Corn (maize) India 1997 47 5–666

Dried figs Austria 1993 13 2–350

Herbs and spices UK 1996 24 1–51

Nutmeg Japan 1993 43 0.2–16.6

Peanuts Brazil 1998 51 43–1099

Peanuts India 1996 45 5–833

Pistachio nuts Netherlands 1996 59 2–165

Rice Equador 1997 9 6.8–40

Soybeans Argentina 1991 10 1–36

Spices Sweden 1998 90 0.1–62

Wheat Uruguay 1996 20 2–20

AFLATOXINS 67

may be formed in the seed tissue, even though, on

harvest, the crop looks sound and is subsequently

correctly stored.

0009 The molds producing aflatoxins are tropical and

subtropical, and, although their spores occur in

temperate parts of the world, these species do not

compete in a cool climate. Thus, the production of

aflatoxins is not normally a problem in Britain

and Scandinavia. However, in both these parts of

the world, a special problem emerged when high-

moisture barley was prepared as a winter feed by

protecting it with propionic acid, or mixtures of

acetic and propionic acids. When properly applied,

these antifungal agents protect the material from

mold spoilage, but if the dosing is not high enough,

Aspergillus flavus is able to grow and, under these

conditions of partial inhibition, actually produces

enhanced levels of aflatoxin.

0010 Aflatoxin B

is readily metabolized in the animal

body and may be secreted in a number of body fluids.

Thus, when cows are fed on animal feed contamin-

ated with aflatoxin B

, a significant proportion is

secreted in the milk as aflatoxin M

(Figure 2),

which is only slightly less toxic than B

. Aflatoxin

has also been detected in human milk, and be-

cause milk is consumed in relatively large quantities

by both the very young and elderly, there have been

many surveys to provide some guidance about expos-

ure of the human population in different parts of the

world. Table 3 gives some results from such surveys.

Toxicology

0011 The aflatoxins show both acute and chronic toxicity,

and one of the outstanding features in the toxicology

of the aflatoxins is the wide variation in response

amongst different species of animals and even

between the male and female of the same species.

Table 4 shows the variation in ld

, which is a meas-

ure of acute toxicity, and Table 5 shows the variation

in td

as a measure of carcinogenicity. It can be seen

that for some animals, such as the rat, aflatoxins are

very carcinogenic, and yet in other species, it is diffi-

cult to demonstrate carcinogenicity. This consider-

able variation in biological response arises from the

requirement that the mold metabolite itself has to be

metabolized in order that a toxic response occurs, and

the metabolites responsible for acute toxicity differ

from those reponsible for carcinogenicity. There are

several metabolic reactions (Figure 3), such as the

demethylation to aflatoxin P

and hydration to afla-

toxin B

, which may lead to a decrease in toxicity. A

critical reaction associated with the toxic response is

Aflatoxin M

OCH

fig0002 Figure 2 Structure of aflatoxin M

tbl0003Table 3 Selection of surveys for aflatoxin M

in milk

Type of milk Country Date Incidence

(%)

Concentration

range (ng l

1

)

All milks Thailand 1997 93 50–>500

Milk powder UK 1981–1983 13 10–400

Milk powder Sweden 1985 100 6–57

Pasteurized milk Greece 1997 89 1–177

Raw milk India 1995 18 100–3500

Raw milk India 1997 11 100–1000

Raw milk Poland 1997 23 3–25

Raw milk Equador 1997 74 125–6000

Raw milk UK 1980 31 30–520

Raw milk UK 1996 16 10–90

UHT milk Spain 1995 14 10–40

tbl0004Table 4 Oral LD

values for aflatoxin B

Animal species

(mgper kilogram of bodyweight)

Rabbit 0.3

Duckling (1 day old) 0.36

Dog 0.5–1.0

Cat 0.6

Pig 0.6

Baboon 2.0

(Humans 5.0?)

Rat (male) 5.5

Rat (female) 17.9

Macaque monkey 7.8

Mouse 9.0

Hamster 10.2

Chicken embryo 0.025 mg per egg

tbl0005Table 5 TD

values for carcinogenesis of aflatoxin B

Animal species

(mg per kilogramof body weight per day)

Fisher rat 1.3 (male); 7.5 (female)

Wistar rat 5.8 (male); 6.9 (female)

Porton rat 3.1 (male); 12.5 (female)

(Human 132?)

Rhesus monkey 156

Cynomolgus monkey 848

Swiss mouse (male) > 5300

68 AFLATOXINS

the formation of the 8,9-epoxide, which may react

with guanine residues in DNA, one of these being the

third base in codon 249 of the p53 gene, the product

of which is an important component in the tumor

suppressor system. The epoxide itself can be further

metabolized to the 8,9-dihydrodihydroxy derivative

of aflatoxin, which can react with the lysine residues

in protein molecules and is a good candidate for the

acute toxin. An animal that readily forms the epoxide

but does not efficiently metabolize this further may be

especially sensitive to the carcinogenic activity of af-

latoxin. However, an animal that forms the epoxide

but effectively metabolizes this further to the dihy-

droxy derivative may be especially sensitive to the

acute toxicity of aflatoxin. The structure of aflatoxin

(which may not itself be toxic) may facilitate its

ready transport into the liver cell, where it is metab-

olized to these far more hazardous molecules.

0012 Where do humans fit into this picture? By extrapo-

lation from both molecular biological studies, and the

epidemiology of cases of acute aflatoxin poisoning

reported, it would seem that humans are not as sensi-

tive as birds, dogs, or cats but more sensitive than the

rat to the acute toxicity of aflatoxin B

. Similarly, an

analysis of exposure versus liver cancer incidence

again puts humans in the middle of the range of

sensitivity to the carcinogenic activity of aflatoxin

. In the early days of the assessment of the role of

aflatoxin in the epidemiology of liver cancer in

humans, the issue was clouded by the undoubted

role of other agents, such as infection with hepatitis

B virus, in the initiation of liver cancer. It is probable

that these two agents act synergistically, and it is

possible to construct a risk assessment model to

include both. Such a risk assessment has led to the

suggestion that an estimated intake level of between

200 and 400 ng per day would correspond to a

lifetime liver cancer excess risk of 1  10

5

in the

USA. It is clear that it is prudent to assume that

aflatoxins are carcinogenic to humans and that efforts

should be made to control its occurrence in the food

chain.

Control

0013The highest concentrations of aflatoxins are formed

as a result of inappropriate storage conditions post-

harvest, and if commodities can be rapidly dried and

OCH

Aflatoxin B

Milk

Cancer

Acute toxicity

Aflatoxin epoxide

Hydroxylation

Epoxidation

Hydration

Interaction

with DNA

interaction

with enzymes

Dihydro, dihydroxy-aflatoxin

Aflatoxin M

fig0003 Figure 3 Some metabolic products from aflatoxin B

AFLATOXINS 69

stored so that the water activity does not exceed 0.78,

growth of aflatoxigenic molds will be inhibited. Afla-

toxin biosynthesis itself is inhibited at water activities

below 0.83. There is a strong interaction between

water activity and temperature in the production of

aflatoxins, and if the temperature of storage can be

maintained below 15



C, the minimum water activity

for toxin production may be as high as 0.9. Any kind

of damage to the commodity can enhance aflatoxin

formation.

0014 Lower but significant levels of aflatoxin can be

formed before harvest if the crop is subjected to

drought stress, and the soil temperature (in the case

of peanuts) is between 25 and 32



C during the 6 or 7

weeks before harvest. Table 6 shows examples of

preharvest contamination of corn in North Carolina

during the period 1976–1980. It can be seen that

during 1980, approximately 65% of the crop was

contaminated by at least 20 mg per kilogram of afla-

toxins and so could not be used for human consump-

tion leading to an estimated loss of $31 million to the

producers. The real cost was even higher if the prob-

lems for the animal farming industry were also taken

into account.

0015 Because it seemed almost inevitable that contamin-

ation will occur in some years, there is interest in the

possibility of some form of biological control of pre-

harvest contamination by inoculating the soil with

nonaflatoxigenic strains of Aspergillus flavus. Green-

house and field experiments do indicate that such

nontoxigenic strains can outcompete or displace resi-

dent toxigenic strains leading to reduced preharvest

aflatoxin contamination. Alternatively, strains of

commodities such as corn and peanuts can be bred,

in which aflatoxin biosynthesis is inhibited, even if

these aflatoxin producing molds have established an

endophytic relationship with the plant.

0016 A number of physical and chemical methods have

been tested for the removal of aflatoxins from con-

taminated materials. Although pure aflatoxins are

relatively unstable in the presence of water, these

compounds are remarkably stable at elevated tem-

peratures when they are present in the complex

matrix of a food. Cooking is thus not effective in

removing aflatoxins. In particulate commodities

such as peanuts, a significant reduction in contamin-

ation levels can be achieved by the identification and

removal of damaged kernals. Because contaminated

kernals are often (but not always) damaged or dis-

colored, they may be removed by sieving, flotation,

and density segregation, or by electronic sorting using

a laser beam to identify the discolored kernals. Of

several chemical methods tested, the only one that has

shown any promise is the use of aqueous or gaseous

ammonia at moderate temperatures and elevated

pressures. This treatment leads to the production of

several products, two of which have been character-

ized (Figure 4) and shown to be considerably less

toxic than aflatoxin B

0017Many countries have established legislation setting

maximum permissible levels of aflatoxins in food and

animal feeds, and a selection of these are shown in

Table 7. In some countries, there is zero tolerance,

which will be based on the sensitivity of analytical

tbl0006 Table 6 Preharvest contamination in corn by aflatoxin in North

Carolina

Crop year Aflatoxin concentration (mgkg

1

)

0^19 20^100 >10 0

1976 64.2 27.7 8.0

1977 58.1 30.2 11.6

1978 87.0 12.0 1.0

1979 67.3 28.3 4.4

1980 34.3 48.1 17.6

Aflatoxin D

OCH

MW 206

OCH

fig0004Figure 4 Breakdown products from ammonia treatment of af-

latoxin B

tbl0007Table 7 Selection of maximum tolerated levels of aflatoxin B

in foods

Country Maximum

level (mgkg

1

)

Products

Argentina 0 Groundnuts, maize, and products

Brazil 15 All foodstuffs

China 10 Rice and edible oils

Hungary 5 All foods

India 30 All foods

Japan 10 All foods

Nigeria 20 All foods

Poland 0 All foods

South Africa 5 All foods

UK 2 Groundnuts, nuts, dried fruits, and

cereals

USA 20 All foods

70 AFLATOXINS

methods. When the FAO/WHO realized that a valu-

able plant protein such as groundnut meal, used in the

formulation of foods for their food aid programme,

could be contaminated with a potential carcinogen

soon after the discovery of aflatoxin, they set a max-

imum concentration of 30 mgkg

1

. It was recognized

that, in the absence of detailed knowledge of the risk,

this was a compromise that recognized that a more

stringent level could result in there being no material

available for a food aid program. India still accepts

this level. The European Union adopted a maximum

acceptable level of 2 mg of aflatoxin B

per kilogram

for groundnuts, nuts, dried fruit, cereals, and their

products, and 0.05 mg of aflatoxin M

per liter in

milk in 1998.

Analysis

0018 The setting of stringent maximum tolerated levels

implies the ability to obtain analytical results that

are sensitive, specific, quantitative, and reproducible.

In the early days of concern over the presence of

aflatoxin in foods and animal feeds, there was not

sufficient confidence in the chemical analysis of afla-

toxins, and biological tests, such as the 1-day-old

duckling test and the chicken embryo test, were used

because they provided direct evidence of the toxicity

of samples. With increasing confidence in the use of

confirmatory tests in combination with TLC and

high-pressure liquid chromatography (HPLC), these

became increasingly acceptable. Most recently, highly

specific monoclonal antibodies have become com-

mercially available for the aflatoxins and several

other groups of mycotoxins. These can be used in

two ways, either directly as enzyme-linked immuno-

sorbent assays (ELISA) or bound to a solid substrate

as immunoaffinity columns. ELISA is not generally

sufficiently quantitative at very low concentrations

but can be very effectively used as a rapid screen for

positive samples, which can then be analyzed by

physicochemical methods. Immunoaffinity columns

provide an excellent way of obtaining a very clean

concentrated extract from a sample, which can then

be quantified by HPLC.

0019 In many parts of the world, TLC has to be

the method of choice because of cost constraints.

Using silica-gel plates, the four common aflatoxins

can be separated and are sensitively detected be-

cause of their intense fluorescence under long-wave

ultraviolet. Because so many compounds run with

similar R

values (position on the chromoto-

gram relative to the solvent front both measured

from the point of application of the sample) and

fluoresce like aflatoxins, it is essential that positive

samples are confirmed using a confirmatory test

such as derivatization with an acid treatment. Afla-

toxins B

and G

do not fluoresce so intensely in

solution, thus reducing the sensitivity of the fluor-

escence detector in HPLC. However, postcolumn

reaction with either bromine or iodine enhances

fluorescence considerably and provides a confirma-

tory test as well.

0020There have been several collaborative studies in

Europe under the Measurements and Testing Pro-

gramme to validate methods to meet the stringent

analytical requirements of European regulations.

Such a collaborative study has, for example, evalu-

ated the efficiency of using immunoaffinity columns

for the clean-up of samples obtained from a wide

range of food matrices for subsequent determination

of aflatoxin B

and total aflatoxins by liquid chro-

matography. The European Union legislation for

aflatoxin M

is especially stringent (0.05 mgl

1

requiring both the validation of methods and the

proficiency testing of a network of European Union

National Reference laboratories set up to determine

aflatoxin M

in milk.

0021Analysis contains three stages, sampling, extrac-

tion, and quantitation, and a competent analyst

can now achieve the required degree of sensitivity,

precision, and specificity required. The analyst needs

only 50–100 g of material, and the real problems arise

from a consideration of how that sample was obtained

from, for example, a shipment of many tonnes of

commodity. Except in liquid foods, such as milk,

aflatoxins are not uniformly distributed in a com-

modity, and this was beautifully demonstrated when

the UK Ministry of Agriculture, Fisheries and Food

(MAFF) purchased a consignment of whole dried figs

that had been rejected at the port of entry. The con-

signment, of just over 10 tonnes made up of 850

boxes each containing 12 kg, had been rejected be-

cause an analysis based on a 20-kg sample had indi-

cated a contamination level of 33 mgkg

1

. Clearly,

with a statutory limit of 4 mgkg

1

at that time, the

UK had to reject the consignment.

0022Having purchased the consignment, MAFF ana-

lyzed 200 boxes as though each were a single sample.

This gave a mean of 15.4 mgkg

1

, which still implies

rejection but clearly shows that the answer depends

on the method of sampling. A study of the distribu-

tion amongst these 200 boxes showed that more than

70 had less than 4 mgkg

1

and could have been

accepted if only they could have been identified! A

further 50 boxes had between 4 and 10 mgkg

1

and

could have been accepted subject to further process-

ing. Thus, about 67% of these 200 boxes were actu-

ally acceptable, but a few had contamination levels of

> 200 mgkg

1

. Indeed, in a further study, individual

boxes were divided into 12 1-kg samples and each

AFLATOXINS 71

analyzed separately. In one of these boxes, 11 of

the samples had < 10 mgkg

1

, but a single sample

had 2063 mgkg

1

! There is nothing unexpected in

these results, and there are studies on other com-

modities such as peanuts and pistachio nuts that

also demonstrate such a skewed distribution, but

they provide quantitative information on which

to design a sampling plan. Such a sampling plan

has to be acceptable to both the producer and the

consumer.

0023 An example of a sampling plan agreed by the Euro-

pean Union for the analysis of figs and other dried

fruit is outlined below:

0024 Subdivide each lot into units of 25 tonnes (15–30

tonnes).

0025 Each unit is sampled separately.

0026 The number of incremental samples should be at

least 100.

0027 Each incremental sample should be about 300 g.

0028 The aggregated sample is 30 kg.

0029 The aggregated sample is divided into samples of

10 kg, each of which is homogenized and analyzed

separately.

Significance

0030 The presence of aflatoxins in foods and animal feeds

is quite widespread, and it is possible that, with pre-

sent agricultural practice, it may not be possible to

prevent contamination in some parts of the world.

The aflatoxins are acute toxins, and they have

undoubtedly caused the death of farm animals and

humans on several occasions. They also have chronic

toxicity being immunosuppressive and carcinogenic.

For these reasons, many countries have set legislative

maximum levels in human foods. Some countries also

set maximum acceptable levels in animal feeds in

order to prevent the contamination of milk by afla-

toxin M passing through the food chain. There is as

yet no international agreement on these levels, and

those countries that are importers of commodities

that are susceptible to contamination tend to be

more stringent than those that are producers and

exporters. There is the potential for difficulties in

international trade while this situation remains.

See also: Carcinogens: Carcinogenic Substances in

Food: Mechanisms; Mycotoxins: Toxicology