FAO Meeting Report No. PL/1965/10/2
EVALUATION OF THE HAZARDS TO CONSUMERS RESULTING FROM THE USE OF
FUMIGANTS IN THE PROTECTION OF FOOD
The content of this document is the result of the deliberations of the
Joint Meeting of the FAO Committee on Pesticides in Agriculture and
the WHO Expert Committee on Pesticide Residues, which met 15-22 March
Food and Agriculture Organization of the United Nations
World Health Organization
1 Report of the second joint meeting of the FAO Committee on
Pesticides in Agriculture and the WHO Expert Committee on Pesticide
Residues, FAO Meeting Report No. PL/1965/10; WHO/Food Add./26.65.
(including hydrogen cyanide evolved from calcium cyanide)
Hydrogen cyanide, prussic acid
H - C - N
Relevant physical and chemical properties
Physical state (atmospheric pressure, 20°C); colourless liquid
Flammability limits in air: 6-41% by volume
Water: soluble in all proportions
Organic solvents: infinitely soluble in alcohol and ether
Specific gravity (liquid): 0.688
Specific gravity (gas): 0.9
Hydrogen cyanide has been widely employed for fumigating dry
foodstuffs including cereals and milled cereal products, seeds,
pulses, nuts and dried fruit and also tobacco. It has also been used
for the disinfestation of buildings, such as flour mills, warehouses,
and domestic houses, and ships (the latter usually directed against
rats). For all these purposes hydrogen cyanide has been largely
superceded by other fumigants which are more convenient or more
efficient (in particular by methyl bromide) or other methods of
control have taken the place of fumigation. However its use continues
on a limited scale.
It is not generally recommended for moist materials such as fresh
fruit and vegetables many of which suffer damage by burning, wilting
or discoloration (Monro, 1961).
For fumigation the usual source of the gas is the liquid hydrogen
cyanide either in cylinders or absorbed on a solid material.
Various crude forms of calcium cyanide are also used as a vehicle
for the generation of hydrogen cyanide gas by the action of water or
moisture and some of these, in granular form, have been used as grain
fumigants. These products generally contain not more than 50% calcium
cyanide, Ca(CN)2 and produce about half of this weight as available
HCN or approximately 20-25% of the weight of the crude material. The
main impurities are lime, cyanamide, carbon and calcium carbide. The
material is usually added continuously to the grain stream as a bin is
filled using a dosage rate of 10 lb or 20 lb per 1000 bushels. It is
recommended that the grain should not be moved for at least 72 hours
and should, if possible, be allowed to remain for a week or 10 days.
Because of its extreme solubility in water, hydrogen cyanide is
most firmly retained by moist commodities. Generally the bulk of the
gas escapes from drier products fairly readily and without reaction
with the constituents, but small amounts of gas may be retained for
long periods. Monier-Williams (1930) gives data collected from the
literature up to 1929 for residual hydrogen cyanide found in a large
number of treated commodities under the headings: milk and milk
products; oils and fats, meat, fish, etc.; cereals, flour, etc.; fresh
fruit; dried fruit; fresh vegetables; tea, coffee, cocoa; and
miscellaneous foods, together with details of treatments.
Cereal grains take up hydrogen cyanide during fumigation and
small amounts of gas remain associated with the grains for long
periods, but with moving, cleaning and milling this is progressively
It has been suggested that a proportion of the sorbed cyanide may
be combined with a constituent of the bran and that this compound
slowly undergoes decomposition (Turtle, 1941).
After fumigation with hydrogen cyanide at a measured
concentration-time product of 60 mg h/l and subsequent aeration for
seven days, whole wheat of 11.5% moisture content showed hydrogen
cyanide residues of 10 ppm, bran 33 ppm and the flour 5 ppm. Wheat of
up to 18% moisture content fumigated at a measured concentration-time
product of 100 mg h/l and then milled to 65% extraction without
washing or cleaning, showed residues in the flour of 10 ppm hydrogen
cyanide, and 3 ppm in bread baked from this flour (Pest Infestation
One hundred and ninety days after fumigating wheat and maize with
hydrogen cyanide 0.3-0.5 ppm was found, and in flour 0.2 ppm after 45
days (Desbaumes and Deshusses, 1956).
A bag of wheatmeal (85% extraction) fumigated with hydrogen
cyanide at a measured concentration-time product of 185 mg h/l showed
a residue of 104 ppm at the centre of the bag after airing for two
days, 18 ppm after seven days, 6 ppm after 14 days and 5.4 ppm after
30 days (Pest Infestation Laboratory, 1943a).
Much of the evidence on the retention of hydrogen cyanide by
wheat and its milled products relates to treatments with the granular
form of calcium cyanide, particularly with the proprietary material
Cyanogas G which yields about 25% by weight of hydrogen cyanide.
After application of Cyanogas to 155 tons of Manitoba wheat of
about 12% moisture content at a rate of 20 lb per 1000 bushels the bin
remained closed for 17 days. Residual hydrogen cyanide determined in
31 samples collected as the bin was emptied (and therefore before
cleaning) varied between 26 and 62 ppm (Pest Infestation Laboratory,
Hydrogen cyanide reacts with laevulose in dried fruit to form
laevulose cyanhydrin (Monier-Williams, 1930; Turtle, 1941; Page and
Lubatti, 1948). This compound may be retained after prolonged aeration
since the slight acidity in dried fruit favours its stability. Seven
days after a normal treatment of dried fruit with hydrogen cyanide an
average residue of 60 ppm could be expected, of which about 75% would
exist as laevulose cyanhydrin and 25% as free hydrogen cyanide. Only
if wet fruit is treated would residues up to 250 ppm be expected
Effect of fumigant on treated crop
(a) HCN naturally occurring in food
Some foodstuffs of vegetable origin contain HCN, generally as
glucoside. From glucoside, free HCN is liberated by enzymatic action
in plants or in the digestive tract. The best characterized
cyanogenetic glucoside is perhaps amygdalin, which is present
especially in the seeds and leaves of the cherry, almond, peach, etc.
Cherry kernels yield about 170 mg per 100 g and bitter almond pulps
about 250 mg per 100 g (Sollman, 1944).
Feeding amygdalin to a small group of rats at a level of 1000 ppm
(equivalent of 60 ppm HCN in the diet) for 12 weeks was without
effect. Since amygdalin in the digestive tract is only partly
hydrolysed, the level of cyanogenetic glucosides up to 500 ppm in
foods is considered to be of no health hazard (Lehman, 1959).
Lima beans contain linamarin. After enzymatic hydrolysis 42 ppm
HCN was found in lima beans; some specimens of lima beans yield as
much as 180 ppm HCN (Lehman, 1959; Malkus, 1957). HCN in canned whole
apricots, cherries and prunes was found to be 0.13, 0.048 and 0.012
ppm respectively (Luh and Pinochet, 1959).
In some samples of sec wine as high as 0.140 ppm free HCN was
detected; combined HCN amounted to 0.230 ppm. In alcoholic
fermentation a soluble substance, possibly vitamin B12, is formed,
which readily eliminates HCN at normal temperatures as proved in
experiments with yeast (Mestres, 1961).
(b) HCN added for fumigation purposes
Wheatmeal (85% extraction) fumigated with hydrogen cyanide at
measured concentration-time products of 54-185 mg h/l showed damage to
baking quality in the form of decrease in loaf volume, coarsening of
crumb and decreased spring figure and increased extensibility in
extensometer tests (Pest Infestation Laboratory, 1943a).
In wheat fumigated at a wide range of dosage levels and at
moisture contents of 11, 15 and 18% no damage to milling quality was
noted but damage to the baking quality of 65% extraction flour
prepared from this wheat was observed at all levels of treatment
especially at the higher dosage and moisture contents. After thorough
aeration of the flour for one month no damage to baking quality was
observed, showing that the previous damage was due to unaired hydrogen
cyanide (Pest Infestation Laboratory, 1940).
The damage to baking can also be largely reversed by treatments
with certain of the chemicals used as "improvers" including nitrogen
trichloride (Agene) (now no longer permitted), and potassium bromide
(Turtle, 1941; Desbaumes and Deshusses, 1956). Wholemeal flour treated
at a measured concentration-time product of 80 mg h/l showed no
destruction of vitamin B1 (Pest Infestation Laboratory, 1940).
Hydrogen cyanide is extremely toxic and the intoxication can be
caused not only by ingestion and inhalation, but also by percutaneous
resorption of liquid HCN and its vapours. Death of the organism
results from inhibition of the iron (ferric) containing cell
respiratory enzymes. The cytochromoxydase is the most sensitive. The
inhibition is reversible.
Cyanides in the organism are in their greatest part metabolized
to thiocyanate and excreted in this form in urine (Lang, 1894). In
rabbits 80% is excreted in 24-48 hours; in dogs the excretion is
slower; in sheep, 60% is excreted within three days (Baumann et al.,
1933; Mukerji and Smith, 1943; Blakley and Coop, 1949).
There are other metabolites as well as thiocyanate. To a slight
extent cyanide can be oxidized to carbon dioxide and formate (Boxer
and Richards, 1952). From the urine of rats
2-iminothiazolidine-4-carboxylic acid was isolated and formed 15% of
the injected dose of KCN, thiocyanate accounting for 80%. The above
acid is formed in vivo from cystine and HCN and, from the metabolic
of view, it is inert (Wood and Cooley, 1956).
Thiocyanate is present normally in human saliva in a
concentration of about 0.01% (Shohl, 1939). In serum and urine, the
average values for thiocyanate, as KCNS, are reported as follows: in
non-smokers 0.54 mg % and 0.65 mg/24 hours, in smokers 1.52 mg % and
10 mg/24 hours (Lawton et al., 1943).
The conversion of cyanide to thiocyanate occurs by means of the
specific enzyme rhodanase, which catalyses the formation of
thiocyanate from cyanide in the presence of sodium thiosulfate or
colloidal sulfur. Rhodanase activity in the liver decreases in the
order rat>rabbit>man>dog. In vitro, the whole liver from one dog
is capable of detoxicating 4015 g of cyanide in 15 minutes (Lang,
1933; Himwich and Saunders, 1948). Rhodanase is present in large
amounts in all tissues but not in blood. In the detoxication mechanism
of the organism an important role is played by the availability of
sulfur. With high concentrations of thiocyanate in the organism,
cyanide can be liberated. This accounts for some of the toxic symptoms
observed after the injection of large doses of thiocyanate. The
formation of cyanide from sodium thiocyanate was seen both in dogs and
men when NaCNS was injected in doses of 300 or 700 mg/kg per man,
respectively (Goldstein and Rieders, 1951). Partial conversion of
thiocyanate to cyanide in the presence of erythrocytes was confirmed
by in vitro experiments (Pines and Crymble, 1952). This conversion
is evidently dependent on the presence of an enzyme found only in
erythrocytes and called thiocyanate oxidase (Goldstein and Rieders,
Compound Animal Route LD50 mg/kg Reference
Potassium Mouse subcutaneous 6.02 + 0.33 Spector, 1956
cyanide intravenous 2.5 (LD) "
Rat oral 10-15 (MLD) "
intravenous 2.5 (MLD) "
Acute toxicity (continued)
Compound Animal Route LD50 mg/kg Reference
Potassium Dog oral 5.3 (LD) Gettler &
cyanide Baine, 1938
Sodium Rabbit subcutaneous 2.2 (MLD) Spector, 1956
cyanide Guinea-pig subcutaneous 5.8 Ghiringhelli,
Dog intravenous 2.8 (LD) Spector, 1956
Sodium Mouse oral 598.4 + 18.3 Spector, 1956
thiocyanate intravenous 483.5 + 9.3 "
Rat oral 764.7 + 50.9 "
intra-peritoneal 540 + 42.5 "
The minimum lethal absorbed dose of HCN after administration of
cyanide to the dog was 1.1-1.5 mg/kg by inhalation and 1.06-1.4 mg/kg
by mouth. The same figures in man obtained from cases of suicide are
0.5-1.4 mg/kg by mouth and in one case 3.6 mg/kg (Gettler and Baine,
For man, the acute toxic oral dose of HCN is usually given as
50-90 mg, for potassium or sodium cyanide 200 mg, representing 81 and
110 mg HCN, respectively (Lehman, 1959). Data on the oral lethal dose
of cyanide for man in four cases of suicide, calculated from the total
amount of HCN absorbed in the body at the time of death, and from the
amount of HCN found in the digestive tract, differed considerably
(calculated as mg HCN): 1450 (62.5 kg body-weight), 556.5 (74.5 kg),
296.7 (50.7 kg), and 29.8 (51 kg) (Gettler and Baine, 1938).
By inhalation an HCN concentration of 135 ppm (150 mg/m3) is
given as lethal after 30 minutes, 270 ppm (300 mg/m3) as immediately
lethal (Patty, 1942).
The American TLV (threshold limit value) for HCN (1964) in eight
hours' exposure in industry is 10 ppm (11 mg/m3) (Anon, 1964).
No cases of chronic intoxication in industry have been diagnosed.
The possibility of chronic intoxication with HCN or cyanides is
usually considered to be improbable. Reports of single cases of
"chronic cyanide poisoning" after repeated occupational exposure are
considered to represent thiocyanate intoxication (Hamilton and Hardy,
1949). In one case the symptoms were reproduced by daily intravenous
injection of 1.4 g of sodium thiocyanate (Wütherich, 1954). Two other
cases with thyroid changes following occupational exposure to cyanide
have been described (Hardy et al., 1950).
Dog. Three males and two females were fed for 30 days on a
diet containing 150 ppm of sodium cyanide. One male and one female
served as controls. No unusual signs or symptoms were noted, and
general behaviour or appearance, and food consumption were not
affected. Total and differential leucocyte counts, haemoglobin and
haematocrit were determined prior to the start of the experiment and
four weeks later. The results were similar, organ weights fell within
the normal range. In comparison to the controls and after elimination
of histopathological changes induced by infection, it was concluded
that feeding 150 ppm of sodium cyanide to dogs for 30 days did not
induce any gross or microscopical pathology (American Cyanamid Co.
Three female dogs were given NaCN in gelatin capsules every day
in doses of 0.5, 2 and 2 × 2 mg/kg body-weight for 14-1/2 months,
always in the morning when their stomachs were empty. The fourth bitch
in the group was the control. The two experimental dogs which were
given doses of 2 and 2 × 2 mg/kg body-weight, showed toxic symptoms
immediately after dosing, which did not last more than 30 minutes. In
the dog which was given doses of 0.5 mg/kg, toxic symptoms of
temporary character only began to appear after 53 weeks. After one of
the doses she died suddenly in anoxaemic convulsions. In the course of
the experiment, a complete haematological examination was carried out
at intervals of 1-2 months, as well as determination of plasma
proteins, residual nitrogen, blood sugar, potassium, sodium,
chlorides, calcium, thiocyanate, cholesterol, bicarbonate
concentration, functional liver tests with tetrabromphenolphthalein,
examination of urine for protein, and examination of the sediment.
Only elevated erythrocyte counts up to the eighth month of the
experiment, and a little lowering of the level of albumin towards the
end of the experiment were found. The concentration of thiocyanate in
plasma stabilized towards the end of the experiment at a level of
under 1 mg %. Such a low level of thiocyanate could not have any toxic
effect. Degenerative changes of ganglion cells in the central nervous
system were found post mortem. The Purkinje cell system of the
cerebellum was especially affected, as a consequence of repetitive
attacks of acute hypoxia (Hertting et al., 1960).
Pig. Two pigs remained healthy after being fed for 11 days
upon a diet of wheatfeed treated at a very high dosage with calcium
cyanide (equivalent to 1500 ppm of hydrogen cyanide) and partially
aired before feeding. The amount of recoverable hydrogen cyanide in
the diet fell, during the period of the feeding experiment from 318
ppm to 206 ppm (Pest Infestation Laboratory, 1944).
Rat. Two groups of 20 rats (10 males and 10 females) were fed a
diet fumigated with hydrogen cyanide, containing residual HCN in the
concentration of 100 and 300 ppm, for two years. Another group of 20
rats was fed a control diet. Growth, food consumption and survival in
both groups were comparable. Haematological values determined
initially and at the end of the experiment appeared to be within
normal limits. Organ-body-weight ratios for the liver, kidneys,
spleen, brain, heart, adrenals and testes or ovaries did not show any
substantial differences from controls. Histological examination of
tissues was carried out for the heart, lung, liver, spleen, stomach,
small and large intestines, kidneys, adrenals, thyroid, testes or
uterus and ovary, and the cerebrum and cerebellum of the brain. In the
tissues examined no changes due to hydrogen cyanide feeding were
found. At the end of the experiment the amount of free cyanide and
thiocyanate in blood, liver and kidney was determined. In the group
fed 100 ppm HCN free cyanide was found only in red blood cells with an
average of 5.40 µg per 100 ml, thiocyanate was found in plasma with an
average of 936 µg per 100 ml, in the liver and kidney 719 and 1023 µg
per 100 g of tissue, respectively.
In the group fed 300 ppm HCN, free cyanide was found in the liver
of one rat and in the erythrocytes of less than 50% of animals
(average 1.97 µg per 100 g tissue). Average values for thiocyanate in
plasma and erythrocytes were 1123 and 246 µg per 100 ml, respectively,
in the liver and kidney 665 and 1188 µg per 100 g tissue,
respectively. The average thiocyanate values in the controls were as
follows: plasma 361 µg, red blood cells 73 µg per 100 ml; liver 566
µg, kidney 577 µg per 100 g (Howard and Hanzal, 1955).
Comment on the experimental studies reported
Lethal doses of cyanide are of about the same order of magnitude
for most species of mammals. For attaining critical concentration in
tissues and for inducing acutely toxic effect, the intensity and
rapidity of absorption of the HCN dose is decisive. In the short-term
experiments in dogs, the diet contained 150 ppm of sodium cyanide; in
long-term experiments rats 300 ppm HCN, and in neither case was any
sign of intoxication detected except that there was a more than
three-fold increase in the level of thiocyanate in plasma and
erythrocyte in the rat. The lack of other signs of toxicity can be
explained by the fact that HCN administered in food is diluted and for
this reason it is absorbed only slowly, so that the rapidity of the
enzymatic conversion to thiocyanate does not allow the toxic level of
CN' in tissues to be attained. It cannot be excluded that the low
toxicity observed in these cases was due to the chemical reaction of
HCN with other components of the food, or to its chemical
transformation in the gastrointestinal tract.
Long-term experiment in rats in the course of their whole
life-span can be taken as a basis for determining the acceptable daily
dose for man. This dose (100 ppm) does not increase the level of
thiocyanate in the blood to the same extent as reported in smokers,
which is about three-fold of that of non-smokers, so that effects of
thiocyanate produced in the organism from the consumption of food
treated with HCN are improbable.
Level causing no toxicological effect in the rat
The maximum no-effect level in the rat was 100 ppm as residue in
the diet after fumigation with HCN, equivalent to 5 mg/kg body-weiglat
Estimate of acceptable daily intake for man of cyanide
resulting from the fumigation of food: 0-0.05 mg HCN/kg body-weight.
Further work considered desirable
Reproduction studies on the rat.
Anon, (1964) Threshold limit values for 1964. Arch. environm.
Hlth., 9, 545
American Cyanamid Company (1959) Report No. 59-14, 10 August, p. 199
Baumann, E. J., Sprinson, D. B. & Metzger, N. (1933) J. biol. Chem.,
Blakley, R. L. & Coop, I. E. (1949) N.Z. J. Sci. Tech., 31A, 1
Boxer, G. E. & Rickards, J. C. (1952) Arch. Biochem., 39, 7
Desbaumes, P. & Deshusses, J. (1956) Mitt. Lebensmitt. Hyg., 47,
Gettler, A. O. & Baine, J. O. (1938) Amer. J. med. Sci., 195, 182
Ghiringhelli, L. (1956) Med. d. Lavoro, 47, 192
Goldstein, F. & Rieders, F. (1951) Amer. J. Physiol., 167, 47
Goldstein, F. & Rieders, F. (1953) Amer. J. Physiol., 173, 287
Hamilton, A. & Hardy, H. L. (1949) Industrial toxicology, Second
ed., New York, Paul B. Hoeber. Cited in Wolfsie, J. H. & Shaffer C.
Boyd (1959) J. occup. Med., 1, 282
Hardy, H. L., Jeffries, W. McK., Wasserman, M. M. & Waddell, W. R.
(1950) New Engl. J. Med., 242, 968
Hertting, G., Kraupp, O., Schnitz, E. & Wuketich, S. (1960) Acta
pharmacol., 17, 27
Himwich, W. A. & Saunders, J. P. (1948) Amer. J. Physiol., 153,
Howard, J. W. & Hanzal, R. F. (1955) Agric. Food Chem., 3, 325
Lang, K. (1933) Biochem. Z., 259, 243
Lang, S. (1894) Arch. exp. Pathol. Pharmacol., 34, 247
Lawton, A. H., Sweeney, T. R. & Dudley, H. C. (1943) J. industr.
Hyg., 25, 13
"Lehman, A. J." (1959) Quart. Bull. Ass. Food Drug Offic., 23, 55
(article signed A.J.L.)
Luh, B. S. & Pinochet, M. F. (1959) Food Res., 24, 423
Malkus, Z. (1957) Czech. Hyg., 2, 251
Mestres, R. (1961) Ann. Falsif. Expert. Chim. 54, 284
Monier-Williams, G. W. (1930) Rep. Publ. Hlth Med. Subj. No. 60,
London, H.M. Stationery Office
Monro, H. A. U. (1961) Manual of fumigation for insect control, FAO,
Agric. Studies, 56
Mukerji, B. & Smith, R. G. (1943) Ann. Biochem., 3, 23
Page, A. B. P. & Lubatti, O. F. (1948) Chem. and Ind., November 13,
Patty, F. A. (1942) J. industr. Hyg., 2, 631
Pest Infestation Laboratory (1940) Unpublished Report No. 37
Pest Infestation Laboratory (1943a) Unpublished Report No. 87
Pest Infestation Laboratory (1943b) Unpublished Report No. 100
Pest Infestation Laboratory (1944) Unpublished Report No. 107
Pines, K. L. & Crymble, M. M. (1952) Proc. Soc. exp. Biol., 81,
Shohl, A. T. (1939) Mineral metabolism, Reinhold, New York, p. 62
Spector, W. S. (1956) Handbook of Toxicology, vol. 1, Saunders,
Sollman, T. (1944) A manual of pharmacology, p. 826, Saunders,
Turtle, E. E. (1941) (Ph.D. Thesis, University London)
Wood, J. L. & Cooley, S. L. (1956) J. biol. Chem., 218, 449
Wütherich, F. (1954) Schweiz med. Wschr., 84, 105