INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY
ENVIRONMENTAL HEALTH CRITERIA 58
SELENIUM
This report contains the collective views of an international group of
experts and does not necessarily represent the decisions or the stated
policy of the United Nations Environment Programme, the International
Labour Organisation, or the World Health Organization.
Published under the joint sponsorship of
the United Nations Environment Programme,
the International Labour Organisation,
and the World Health Organization
World Health Orgnization
Geneva, 1987
The International Programme on Chemical Safety (IPCS) is a
joint venture of the United Nations Environment Programme, the
International Labour Organisation, and the World Health
Organization. The main objective of the IPCS is to carry out and
disseminate evaluations of the effects of chemicals on human health
and the quality of the environment. Supporting activities include
the development of epidemiological, experimental laboratory, and
risk-assessment methods that could produce internationally
comparable results, and the development of manpower in the field of
toxicology. Other activities carried out by the IPCS include the
development of know-how for coping with chemical accidents,
coordination of laboratory testing and epidemiological studies, and
promotion of research on the mechanisms of the biological action of
chemicals.
ISBN 92 4 154258 6
The World Health Organization welcomes requests for permission
to reproduce or translate its publications, in part or in full.
Applications and enquiries should be addressed to the Office of
Publications, World Health Organization, Geneva, Switzerland, which
will be glad to provide the latest information on any changes made
to the text, plans for new editions, and reprints and translations
already available.
(c) World Health Organization 1987
Publications of the World Health Organization enjoy copyright
protection in accordance with the provisions of Protocol 2 of the
Universal Copyright Convention. All rights reserved.
The designations employed and the presentation of the material
in this publication do not imply the expression of any opinion
whatsoever on the part of the Secretariat of the World Health
Organization concerning the legal status of any country, territory,
city or area or of its authorities, or concerning the delimitation
of its frontiers or boundaries.
The mention of specific companies or of certain manufacturers'
products does not imply that they are endorsed or recommended by the
World Health Organization in preference to others of a similar
nature that are not mentioned. Errors and omissions excepted, the
names of proprietary products are distinguished by initial capital
letters.
CONTENTS
ENVIRONMENTAL HEATH CRITERIA FOR SELENIUM
1. SUMMARY AND RECOMMENDATIONS FOR FURTHER RESEARCH
1.1. Summary
1.1.1. Properties and analytical methods
1.1.2. Sources, environmental transport, and distribution
1.1.3. Environmental levels and exposures
1.1.4. Selenium metabolism
1.1.4.1 Absorption
1.1.4.2 Total human body selenium content
1.1.4.3 Distribution
1.1.4.4 Metabolic pools of selenium in the body
1.1.4.5 Metabolic conversion
1.1.4.6 Effect of chemical form of selenium on its
metabolism
1.1.4.7 Selenium excretion
1.1.5. Effects on animals
1.1.5.1 Selenium toxicity
1.1.5.2 Selenium deficiency
1.1.6. Effects on man
1.1.6.1 General population exposure
1.1.6.2 Occupational exposure
1.2. Recommendations for further activities
2. CHEMICAL AND PHYSICAL PROPERTIES; ANALYTICAL METHODS
2.1. Properties
2.2. Analytical methods
2.2.1. Sample collection, processing, and storage
2.2.2. Sample decomposition or other preliminary treatment
2.2.2.1 Wet digestion
2.2.2.2 Predigestion
2.2.2.3 Combustion
2.2.2.4 Fusion
2.2.2.5 Concentration
2.2.3. Removal from interfering substances
2.2.4. Detection and identification of selenium
2.2.5. Measurement of selenium
2.2.5.1 Fluorometric analysis
2.2.5.2 Neutron activation analysis
2.2.5.3 Atomic absorption spectrometry
2.2.5.4 Other methods
3. SOURCES, TRANSPORT, AND CYCLING OF SELENIUM IN THE ENVIRONMENT
3.1. Natural sources
3.1.1. Rocks and soils
3.1.2. Natural selenium in the food chain
3.l.3 Water and air
3.2. Man-made sources
3.2.1. Agriculture
3.2.2. Industry
3.3. Environmental transport
3.4. Biological selenium cycle
4. LEVELS IN ENVIRONMENTAL MEDIA
4.1. Levels and chemical forms of selenium in food
4.1.1. Levels in food
4.1.1.1 Natural differences among food commodities
4.1.1.2 Effects of natural differences in the
availability of selenium in the
environment on levels in food
4.1.1.3 Man-induced changes in selenium levels in
food
4.1.2. Chemical forms of selenium in food
4.2. Drinking-water
4.3. Air
5. HUMAN EXPOSURE
5.1. Estimate of general population exposure
5.1.1. Food
5.1.1.1 Geographical variation
5.1.1.2 Food habits (consumption patterns)
5.1.1.3 Elderly people
5.1.1.4 Infants and children
5.1.1.5 Special medical diets
5.2. Occupational exposure
5.2.1. Levels in the workplace air
5.2.2. Biological monitoring
6. METABOLISM OF SELENIUM
6.1. Absorption
6.1.1. Gastrointestinal absorption
6.1.1.1 Animal studies
6.1.1.2 Human studies
6.1.2. Absorption by inhalation
6.1.3. Absorption through the skin
6.2. Distribution in the organism
6.2.1. Transport
6.2.2. Organs
6.2.2.1 Animal studies
6.2.2.2 Human studies
6.2.3. Blood
6.2.4. Total-body selenium content
6.3. Excretion in urine, faeces, and expired air
6.3.1. Animal studies
6.3.2. Human studies
6.3.2.1 Excretion of selenium
6.3.2.2 Balance studies
6.4. Retention and turnover
6.4.1. Animal studies
6.4.2. Controlled human studies
6.5. Metabolic transformation
6.5.1. Animal studies
6.5.1.1 Reduction and methylation
6.5.1.2 Form in proteins
6.5.1.3 Conversion of selenium compounds to
nutritionally-active forms of selenium
6.5.2. Human studies
7. EFFECTS OF SELENIUM ON ANIMALS
7.1. Selenium toxicity
7.1.1. Farm animal diseases associated with a high
selenium intake
7.1.2. Toxicity in experimental animals
7.1.2.1 Acute and subacute toxicity - single or
repeated exposure studies with oral,
intraperitoneal, or cutaneous
administration
7.1.2.2 Effects of long-term oral exposure
7.1.2.3 Inhalation toxicity
7.1.3. Blood levels in toxicity
7.1.4. Effects on reproduction
7.1.5. Effects on dental caries
7.1.6. Factors influencing toxicity
7.1.6.1 Form of selenium
7.1.6.2 Nutritional factors
7.1.6.3 Arsenic
7.1.6.4 Sulfate
7.1.6.5 Adaptation
7.1.7. Mechanism of toxicity
7.2. Selenium deficiency
7.2.1. Animal diseases
7.2.2. Intakes needed to prevent deficiency
7.2.2.1 Quantitative dietary levels
7.2.2.2 Bioavailability
7.2.3. Blood and tissue levels in deficiency
7.2.4. Physiological role: glutathione peroxidase
7.2.4.1 Function of selenium and relationship to
vitamin E
7.2.4.2 Effect of selenium intake on tissue-
glutathione peroxidase, activity
7.2.4.3 Relationship between blood-selenium levels
and erythrocyte-glutathione peroxidase
activity
7.2.4.4 Gluthathione peroxidase as an indicator of
selenium status
7.2.5. Other possible physiological roles
7.2.5.1 Homeostasis of hepatic haem
7.2.5.2 Microsomal and mitochondrial electron
transport
7.2.5.3 The immune response
7.2.5.4 Selenium and vision
7.2.6. Effects on reproduction
7.2.7. Factors influencing deficiency
7.2.7.1 Form of selenium
7.2.7.2 Vitamin E
7.2.7.3 Heavy metals and other minerals
7.2.7.4 Xenobiotics
7.2.7.5 Exercise stress
7.3. Ratio between toxic and sufficient exposures
7.4. Protection against heavy metal toxicity
7.4.1. Mercury
7.4.2. Cadmium
7.4.3. Other heavy metals
7.5. Cytotoxicity, mutagenicity, and anti-mutagenicity
7.5.1. Cytotoxicity and mutagenicity
7.5.2. Anti-mutagenicity
7.6. Teratogenicity
7.7. Carcinogenicity and anti-carcinogenicity
7.7.1. Selenium as a possible carcinogen
7.7.2. Selenium as a possible anti-carcinogen
8. EFFECTS OF SELENIUM ON MAN
8.1. High selenium intake
8.1.1. General population
8.1.1.1 Signs and symptoms
8.1.1.2 Attempts to associate high selenium intake
with human diseases
8.1.2. Reports on health effects associated with
occupational exposure
8.1.2.1 Fumes and dust of selenium and its
compounds
8.1.2.1.1 Selenium dioxide
8.1.2.1.2 Hydrogen selenide
8.1.2.1.3 Selenium oxychloride
8.2. Low selenium intake
8.2.1. Evidence supporting the possible essentiality of
selenium in man
8.2.2. Signs and symptoms of low intake
8.2.3. Dietary levels consistent with good nutrition
8.2.3.1 Quantitative estimates
8.2.3.2 Nutritional bioavailability
8.2.4. Blood and urine levels typical of low intake
8.2.5. Relationship between blood-selenium levels and
erythrocyte-glutathione peroxidase activity
8.2.6. Attempts to associate low selenium intake with
human diseases
8.2.6.1 Keshan disease
8.2.6.2 Kashin-Beck disease
8.2.6.3 Cancer
8.2.6.4 Heart disease
9. EVALUATION OF THE HEALTH RISKS ASSOCIATED WITH EXCESSIVE OR
DEFICIENT SELENIUM EXPOSURE
9.1. The need to consider the essentiality of selenium in
health risk evaluation
9.2. Pathway of selenium exposure for the general population
9.3. Quantitative assessment of human selenium exposure
9.3.1. Analytical methods for selenium
9.3.2. Food intake data
9.3.3. Blood-selenium
9.3.4. Hair-selenium
9.3.5. Urine-selenium
9.3.6. Blood-glutathione peroxidase
9.4. Levels of dietary selenium exposure in the general
population
9.5. Evaluation of health risks - general population
9.5.1. Predictive value of animal studies
9.5.2. Studies on high-exposure effects in the general
population
9.5.3. Studies on low-exposure effects in the general
population
9.5.4. Evaluation of the involvement of selenium in human
diseases of multiple etiopathogenesis
9.5.4.1 Keshan disease
9.5.4.2 Kashin-Beck disease
9.5.4.3 Ischaemic heart disease
9.5.4.4 Studies on the involvement of selenium in
cancer
9.5.4.5 Caries
9.5.4.6 Health effects related to reproduction
9.6. Occupational exposure
REFERENCES
WHO TASK GROUP ON SELENIUM
Members
Dr R.F. Burk, Jr, Department of Medicine, University of Texas
Health Science Center, San Antonio, Texas, USA
Professor A.I. Diplock, Department of Biochemistry, Guy's
Hospital Medical School, London, United Kingdom (Chairman)
Dr H.N.B. Gopalan, University of Nairobi, Department of
Botany, Nairobi, Kenya
Dr J. Chen, Department of Nutrition and Food Hygiene,
Institute of Health, China National Centre for Preventive
Medicine, Beijing, People's Republic of China
Dr G.N. Krasovskij, Sysir Institute of General and Community
Hygiene, Academy of Medical Sciences of the USSR, Moscow,
USSRa
Professor C.R. Krishna Murti, Integrated Environmental
Programme on Heavy Metals, Centre for Environmental
Studies, Anna University, College of Engineering, Guindy,
Madras, Indiaa
Dr O.A. Levander, Vitamin and Mineral Nutrition Institute, US
Department of Agriculture, Beltsville, Maryland, USA
(Rapporteur)
Professor A. Massoud, Department of Community, Environmental
and Occupational Medicine, Faculty of Medicine, Ain Shams
University, Cairo, Egypta
Professor M.F. Robinson, Nutrition Department, University of
Otago, Dunedin, New Zealand
Secretariat
Dr J. Parizek, International Programme on Chemical Safety,
World Health Organization, Geneva, Switzerland (Secretary)
Dr E.S. Johnson, International Agency for Research on Cancer,
Lyons, France
-------------------------------------------------------------------
a Invited but unable to attend.
NOTE TO READERS OF THE CRITERIA DOCUMENTS
Every effort has been made to present information in the
criteria documents as accurately as possible without unduly
delaying their publication. In the interest of all users of the
environmental health criteria documents, readers are kindly
requested to communicate any errors that may have occurred to the
Manager of the International Programme on Chemical Safety, World
Health Organization, Geneva, Switzerland, in order that they may be
included in corrigenda, which will appear in subsequent volumes.
* * *
A detailed data profile and a legal file can be obtained from
the International Register of Potentially Toxic Chemicals, Palais
des Nations, 1211 Geneva 10, Switzerland (Telephone no. 988400 -
985850).
ENVIRONMENTAL HEALTH CRITERIA FOR SELENIUM
A WHO Task Group on Environmental Health Criteria for Selenium
was held in Geneva on 2 - 6 December 1985. Dr J. Parizek opened
the meeting on behalf of the Director-General. The Task Group
reviewed and revised the draft criteria document and made an
evaluation of the health risks of exposure to selenium.
DR O.A. LEVANDER of the US Department of Agriculture was
responsible for the preparation of the drafts of the document.
The efforts of all who helped in the preparation and
finalization of the document are gratefully acknowledged.
* * *
Partial financial support for the publication of this criteria
document was kindly provided by the United States Department of
Health and Human Services, through a contract from the National
Institute of Environmental Health Sciences, Research Triangle Park,
North Carolina, USA - a WHO Collaborating Centre for Environmental
Health Effects.
1. SUMMARY AND RECOMMENDATIONS FOR FURTHER RESEARCH
1.1. Summary
1.1.1. Properties and analytical methods
Selenium exists naturally in several oxidation states and some
of its chemical forms are volatile. Many selenium analogues of
organic sulfur compounds exist in nature.
A number of procedures exist for the determination of selenium.
However, as selenium exists in volatile and unstable forms and
because of the inhomogeniety of many types of materials, the
methods of sampling, preparation, and storage are equally as
important as the analytical methods. Great care is necessary to
avoid contamination or loss of the element.
The most commonly used analytical methods depend on wet
digestion for destroying organic matter and freeing the element.
Some procedures depend on the formation of the piazoselenol; this
is extracted in an organic solvent and the fluorescence determined.
The fluorometric method with a wide variety of modifications is
very sensitive and reliable and can be adapted for most materials.
It is also inexpensive. Atomic absorption spectrometry, especially
with the atomization of the selenium as the hydride, has been
useful, and atomic absorption methods based on the Zeeman effect
background correction show promise for the determination of
selenium in biological matrices, without prior sample digestion.
Neutron activation analysis, particularly when combined with the
chemical separation of selenium, is an excellent method, limited
only by the cost and availability of equipment.
1.1.2. Sources, environmental transport, and distribution
Selenium appears to be ubiquitous. However, its uneven
distribution over the face of the earth results in regions with
very low or very high natural levels of selenium in the
environment. Geophysical, biological, and industrial processes are
involved in the distribution and transport of the element and its
cycling, but the relative importance of these processes has not
been established. However, the natural geophysical and biological
processes are probably almost entirely responsible for the present
status of selenium in the general environment. This must be given
primary consideration in any evaluation of the superimposed effects
of man's activity on selenium in the environment and food chains.
Some human activities are responsible for the redistribution of
selenium in the environment. Industrial sources of selenium stem
initially from copper refining. During this refining and the
purification of the selenium, there can be some loss of the element
into the environment. In addition, industries concerned with the
production of glass, electronic equipment, or certain metals may
emit selenium into the environment in the immediate vicinity of the
factories involved. The inclusion of the element in manufactured
products provides another avenue for its distribution. There is
concern in several countries with regard to the possible health
effects of low and/or decreasing levels of selenium in the soil
(sections 1.1.3, 1.1.5.2 and 1.1.6.1); the use of fertilizers
containing selenium compounds in some Nordic countries is a
remarkable example of intentional human intervention in the
environmental distribution of selenium.
1.1.3. Environmental levels and exposures
The range of selenium levels in different foods can vary
widely, depending on the natural availability of selenium in the
environment and on certain activities of man, such as the direct
addition of selenium to the food supply.
No data are available on the chemical forms of selenium in
foods produced under normal conditions.
The limited analytical data available show that the levels of
selenium typically found in foods (on a wet-weight basis) range
from: 0.4 - 1.5 mg/kg in liver, kidney, and seafood; 0.1 - 0.4
mg/kg in muscle meat; < 0.1 - 0.8 mg/kg or more in cereals or
cereal products; < 0.1 - 0.3 mg/kg in dairy products; and < 0.1
mg/kg in most fruits and vegetables. However, in countries with
low selenium levels in soil, lower selenium levels than the above
were reported, in particular in meats, cereals, and dairy products.
The levels of selenium in baby foods tend to follow the same trends
as in adult foods, i.e., meat and cereal products contain the
highest levels, and fruit and vegetable products the lowest. Meat-
based infant formulae contain more selenium than formulae based on
milk or soy protein. Average selenium concentrations in human milk
range from 0.013 to 0.018 mg/litre. However, in countries with low
selenium levels in soil and food, lower selenium levels in breast
milk were reported.
Special medical diets based on egg albumen contain more
selenium than diets based on casein hydrolysate. Chemically-
defined diets or total parenteral nutrition solutions based on
amino acid mixtures contain very low levels of selenium.
Levels of selenium in air and water are usually very low, i.e.,
less than 10 ng/m3 in air and only a few µg/litre water.
Food constitutes the main route of exposure to selenium for the
general population. Because of geochemical differences, the
estimates of adult human exposure to selenium via the diet range
from 11 to 5000 µg/day, in different parts of the world; however,
dietary intake more usually falls within the range of 20 - 300
µg/day. Food consumption patterns also affect dietary-selenium
intake, and the extreme values observed have occurred in
populations consuming a monotonous diet comprising a limited range
of locally-grown staple foods. The estimated selenium intake of
infants in different parts of the world during the first month of
life ranges from 5 to 55 µg/day, because of the variation in levels
of selenium in milk. Children consuming synthetic diets, as part
of long-term diet therapy for certain metabolic diseases, such as
phenylketonuria and maple syrup urine disease, ingest only 5 - 11
µg/day. Adults, maintained on chemically-defined diets or total
parenteral nutrition solutions based on amino acid mixtures,
receive only 1 or 2 µg/day. There is growing evidence of the
importance of the bioavailability of selenium in different foods
for human beings.
Human occupational exposure to selenium is primarily via the
air. Exposure via direct contact is rarely of importance, unless
local irritation or skin damage caused by vesicant selenium
compounds facilitates cutaneous absorption. Few quantitative data
are available on the actual levels of occupational exposure to
selenium. However, it is likely that analyses carried out several
decades ago yielded higher values (several mg/m3) than those
carried out more recently.
1.1.4. Selenium metabolism
1.1.4.1. Absorption
Potential sites of selenium absorption from the environment
are the gastrointestinal tract, the respiratory tract, and the
skin. Limited animal and no quantitative human data are available
on the pulmonary absorption of selenium. Nevertheless, high
urinary-selenium levels in workers exposed to selenium in air have
been reported and could indicate pulmonary absorption. Selenite
and selenium oxychloride have been shown to be absorbed through the
skin of experimental animals. The assessment of the possible
pulmonary or dermal absorption rates for various selenium compounds
must take into account differences in their physical and chemical
properties.
The results of several experimental animal studies as well as
direct measurements of selenium absorption in man indicate that
selenium compounds can be readily absorbed in the intestinal tract
and there appears to be no physiological control over this
absorption. This statement is based largely on studies of 75Se-
labelled selenite absorption in rats fed widely varying amounts of
selenium in the diet. All groups absorbed more than 95% of the
75Se administered by stomach tube. High absorption of selenium was
reported in women given 1-mg doses of selenium as selenomethionine
or selenite dissolved in water.
1.1.4.2. Total human body selenium contenta
The reported estimates of the amount of selenium in the adult
human body range from 3 to 14.6 mg. The lower estimates, 3.0 or
6.1 mg, were obtained in New Zealand by indirect techniques
following the administration of radiolabelled selenite or
selenomethionine, respectively, to healthy human volunteers. The
higher estimate, 14.6 mg, was obtained in the USA and was based on
-------------------------------------------------------------------
a The Task Group felt that the term "body burden" might be
misleading when applied to an essential trace element.
direct analysis of autopsy material accounting for 91.7% of the
body mass. The difference between these 2 estimates may be due
either to differences in the techniques employed or to differences
in dietary-selenium intake in New Zealand and the USA.
1.1.4.3. Distribution
Under normal conditions, levels of selenium are higher in the
kidney and liver than in the other major body tissues. Although
muscle-selenium levels are lower, muscle is the tissue present in
the greatest amount in the body, and thus accounts for the highest
proportion of the total body selenium. It is generally assumed
that the levels of selenium in the above tissues and in red blood
cells and plasma are related to the total body content of selenium,
but additional studies are needed to establish this point under
various conditions of selenium exposure, and also in man.
1.1.4.4. Metabolic pools of selenium in the body
There is no evidence of a regulated or specific storage form of
selenium, analogous to ferritin-bound iron. However, various data
suggest that several body pools of selenium could be considered as
serving a similar purpose. Evidence for a carry-over effect of
selenium exists. Sheep fed plants containing adequate levels of
selenium for 5 months and then fed low-selenium plants for 10
months produced lambs with higher tissue-selenium contents and
fewer signs of selenium deficiency than sheep fed only low-selenium
plants. North Americans moving to the low-selenium area of New
Zealand experienced a slow drop in blood-selenium contents for
approximately 1 year, before reaching New Zealand levels.
These observations could be explained by sequestration of
selenium in the form of selenomethionine and/or other selenoamino
acids incorporated into the primary structure of proteins
throughout the body. Selenium sequestered in this form would be
released and made available at a rate corresponding to the turnover
of proteins and the catabolism of selenomethionine. This mechanism
could provide selenium for animals temporarily unable to obtain it
in sufficient amounts through their diets.
The Task Group concluded that present knowledge about selenium
biochemistry was consistent with the concept of different selenium
pools in the body, even if direct measurements of the quantitative
aspects regarding the compartmentalization of selenium in the human
body and the related turnover rates were not available.
1.1.4.5. Metabolic conversion
Biological utilization of selenium from different chemical
forms has usually been assessed by determining the ability of such
compounds to prevent selenium-deficiency diseases or to increase
glutathione peroxidase activity in selenium-deficient animals.
Selenite and selenomethionine readily fulfil both these criteria
and can thus be considered convertible to metabolically-active
forms. In addition, it should be recognized that, in several of
the early metabolic studies on selenium, many other selenium-
containing compounds were shown to prevent dietary liver necrosis
and thus to undergo metabolic conversion to nutritionally-active
forms.
Some of the excretory metabolites of selenium have been
identified. Dimethylselenide is excreted in the breath at high
levels of selenium exposure. Trimethylselenonium ion and several
unidentified selenium metabolites are excreted in the urine.
1.1.4.6. Effect of chemical form of selenium on its metabolism
Animal studies indicate that selenite is not converted to
selenomethionine in the body. Nevertheless, both selenite and
selenomethionine are able to satisfy the nutritional requirement
for selenium.
The results of several studies indicate differences in the
metabolism of these two forms of selenium. On the basis of limited
human studies, in which 100 µg selenium/day was given, selenium in
the form of selenomethionine was retained in larger amounts than
selenium given as selenite; it also resulted in higher blood-
selenium levels. Comparison of selenite and selenomethionine in
chick studies demonstrated that retention of selenium was greater
when given as selenomethionine than when given as selenite. When
mice were given toxic levels of selenium compounds, selenomethionine
administration resulted in higher tissue-selenium levels than
administration of the same level of selenium as selenite.
Furthermore, manifestations of toxicity disappeared rapidly when
selenite was withdrawn from the diet but subsided only slowly
following withdrawal of selenomethionine. The results of limited
studies suggest that the metabolism of both selenium-
methylselenocysteine and selenocystine resembles that of selenite
rather than that of selenomethionine. These findings can be
explained by the incorporation of a portion of the selenium
supplied as selenomethionine into the tissue proteins as the amino
acid.
1.1.4.7. Selenium excretion
In human volunteers given tracer doses of inorganic or organic
selenium compounds orally, excretion was mainly by the urinary
route. However, when people consumed naturally-occurring selenium
in foods, approximately equal proportions were excreted in the
urine and faeces. Very little was excreted in the sweat, and
animal studies have shown that significant respiratory excretion of
volatile selenium compounds only occurs in cases of very high
selenium exposure. Selenium loss from the body, as judged by
whole-body counting studies following administration of single
doses of 75Se-labelled compounds, consisted of an initial phase of
rapid decrease of whole body radioactivity followed, after several
days, by a phase of more gradual 75Se excretion. The results of
animal studies indicate that the rate of selenium excretion in the
initial phase is affected by the dose of selenium administered as
well as by selenium status, whereas the second phase of excretion
is mainly affected by selenium status.
1.1.5. Effects on animals
1.1.5.1. Selenium toxicity
A characteristic sign of acute selenium poisoning in animals is
the odour due to the pulmonary excretion of dimethyl selenide.
Other signs of acute selenium poisoning in dogs and rats include:
vomiting, dyspnoea, tetanic spasms, and death from respiratory
failure. Pathological changes include congestion of the liver with
areas of focal necrosis, congestion of the kidney, endocarditis,
myocarditis, and peticheal haemorrhages of the epicardium.
The oral LD50 reported for sodium selenite varies from 2.3 to
13 mg selenium/kg body weight because of species differences and
other variables. Inhalation exposure to various selenium
compounds, including selenium dioxide, hydrogen selenide, and
selenium dust, proved to be toxic, causing damage to the
respiratory tract, liver, and other organs, lethality being
dependent on the level and duration of exposure. Because of the
limited number of studies and differences in species and other
experimental conditions used, evaluation of the dose-response
relationships for the different compounds is difficult.
It has been shown that the amount of dietary selenium needed to
cause chronic toxicity in animals is influenced by many variables
including the form of selenium and the type of diet. When fed in
the diet, elemental selenium has a low order of toxicity because of
its insolubility. Sodium selenite or seleniferous wheat fed at a
level of 6.4 mg selenium/kg diet caused growth inhibition, liver
cirrhosis, and splenomegaly in young growing rats fed the diet for
6 weeks. Growth inhibition at 4.8 mg/kg diet was not statistically
significant. Levels of 8 mg/kg diet or more caused additional
effects such as pancreatic enlargement, anaemia, elevated serum-
bilirubin levels and, in 4 weeks, death. In another study when
rats were fed 16 mg selenium/kg diet as sodium selenate in a
commercial "laboratory chow" type ration, their median survival age
was reported to be 96 days, and the predominant histopathological
lesion was acute toxic hepatitis. The predominant lesion at 8
mg/kg in the chow diet was chronic toxic hepatitis and the median
survival age was 429 days. Only 4 mg/kg were needed to cause acute
toxic hepatitis in rats consuming a semi-purified diet containing
12% casein. Increased activities of serum alkaline phosphatase (EC
3.1.3.1) and glutamic-pyruvic transaminase (EC 2.6.1.2) were
observed in young growing rats fed a semi-purified diet containing
4.5 mg selenium/kg as seleniferous sesame meal.
By using other criteria of toxicity, some research workers have
claimed deleterious effects of selenium at lower levels of intake,
however, there is a need to develop and validate more sensitive and
specific indicators of selenium poisoning. Farm animals raised in
regions where there are high levels of available selenium in the
soil develop toxicity diseases as a result of consuming plants
containing excess selenium. The levels of dietary selenium needed
to cause chronic toxicity are 5 mg/kg or more in cattle and 2 mg/kg
or more in sheep. Blood-selenium levels higher than 2 mg/litre are
generally associated with frank chronic selenium poisoning in
cattle, but borderline toxicity problems may be observed at levels
between 1 and 2 mg/litre. Blood-selenium levels of 0.6 - 0.7
mg/litre are associated with chronic selenosis in sheep.
On the basis of anecdotal reports, it has been suggested that
excess selenium may cause various practical reproduction problems
in farm animals, such as decreased reproductive performance in
livestock or decreased hatchability in chickens because of
deformities, at levels that do not cause obvious manifestations of
toxicity. A marked deterioration in reproductive performance was
noted in a multigeneration study in which mice were given sodium
selenate in the drinking-water at 3 mg selenium/litre, a level that
had no effect on growth or survival in a typical single-generation
toxicity study. There was a considerable decrease in the number of
litters produced by the third-generation mice and a considerable
increase in the number of runts among the mice that were born.
When monkeys were fed a cariogenic diet and given 2 mg
selenium/litre as sodium selenite for 15 months followed by 1
mg/litre for 45 months, the incidence of caries increased when the
selenium was given during tooth development but not when the teeth
were exposed post-eruptively. On the other hand, the effect of a
cariogenic diet given for two months after weaning was decreased in
rats whose mothers had been given, during the second half of
pregnancy and during lactation, drinking water containing selenium
at the level of 0.8 mg/litre in the form of sodium selenite or
selenomethionine.
In two independent studies, it was shown that, under certain
circumstances, selenium compounds were shown to be less toxic to
animals kept on a high dietary intake of selenium, thereby
suggesting possible adaptation to high selenium exposures.
1.1.5.2. Selenium deficiency
Animals raised in regions where there are low levels of
available selenium in the soil develop deficiency diseases through
consuming plants lacking adequate selenium. These diseases can be
prevented by feeding inorganic selenium compounds to the animals.
Deficiency diseases in which both selenium and vitamin E may
play a role include nutritional muscular dystrophy in sheep and
cattle, exudative diathesis in chickens, and liver necrosis in
swine and rats. Signs specific for selenium deficiency in the
absence of vitamin E deficiency include pancreatic degeneration in
chicks, and poor growth, reproductive failure, vascular changes,
and cataracts in rats.
The level of dietary selenium needed to prevent deficiency
diseases in animals depends on the vitamin E content of the diet.
For example, chicks receiving a diet deficient in vitamin E need
0.05 mg selenium/kg diet to prevent exudative diathesis whereas
0.01 mg/kg will be sufficient to prevent the disease if the diet
contains 100 mg vitamin E/kg.
Under normal vitamin E intake, the level of dietary selenium
needed to prevent deficiency is about 0.02 mg/kg for ruminants and
0.03 - 0.05 mg/kg for poultry.
Blood-selenium levels of less than 0.05 mg/litre are usually
associated with signs of selenium deficiency in sheep. Hepatic
selenium levels of less than 0.21 mg/kg (dry basis) are associated
with a high incidence of white muscle disease in lambs. However,
levels indicative of marginal deficiency can be influenced by the
vitamin E status of the animal.
The physiological function of selenium and its nutritional
relationship to the biological antioxidant vitamin E can be largely
explained on the basis of its role as a component of the enzyme
glutathione peroxidase (EC 1.11.1.9), which is responsible for the
destruction of hydrogen peroxide and lipid peroxides.
There is a close association between the level of dietary
intake of selenium and the glutathione peroxidase activity in
several organs. Also, the glutathione peroxidase activity of
erythrocytes is closely associated with the selenium content of the
blood. Because of these close relationships, measurement of
glutathione peroxidase offers a convenient method for assessing
selenium intake. However, the activity of the enzyme is influenced
by several nutritional and environmental factors that must be
considered when using it as an index of selenium status.
The ratio of the toxic level of selenium in the diet to the
nutritional level of selenium in the diet is approximately 100, but
this ratio can be decreased by nutritional or environmental
factors. For example, deficiency of vitamin E increases the
susceptibility of animals to selenium toxicity but also increases
the nutritional need for the element. Furthermore, inorganic
mercury potentiates the toxicity of methylated selenium
metabolites, whereas methyl-mercury potentiates selenium
deficiency.
Dietary selenium can protect against the toxicity of several
heavy metals, such as mercury or cadmium, and certain xenobiotics,
such as paraquat, but the mechanism of these protective effects is
not known.
Selenium has been suspected of being a carcinogen in the past,
but more recent research suggests that it may be able to protect
against certain types of cancers in experimental animals.
1.1.6. Effects on man
1.1.6.1. General population exposure
As stated above, the main environmental pathway of selenium
exposure in the general population is through food. Nutritional
surveys have shown that extreme dietary intakes range from 11 -
5000 µg/day, but on most diets intakes between 20 and 300 µg/day
can be considered as more typical. The extremes in intake are
reflected in extreme levels of selenium in blood, mean reported
values ranging from 0.021 to 3.2 mg/litre. The highest blood-
selenium levels ever observed in the general population were found
in an area of the People's Republic of China in which an episode of
intoxication reported as selenosis had occurred some years earlier.
In this respect, as well as in at least two other studies in over-
exposed populations, hair loss and nail pathology were the most
marked and readily documented toxic signs. The Task Group, being
aware of the hepatotoxicity of selenium compounds observed in
animal studies, noted that no clinical signs of hepatotoxicity were
observed in the studies of people exposed to high levels of dietary
selenium, but concluded that there is a need for more thorough
evaluation of hepatic function in persons with high selenium
exposure. Tooth decay was also observed in several studies on
over-exposed populations, but in evaluating its significance the
Task Group was unable to exclude interference by other
environmental factors. The Task Group recognized the difficulty in
establishing an exact dose-response with respect to selenium in the
above studies. The range of blood values noted was 0.44 - 3.2
mg/litre; in these studies no adverse effects were reported at the
lower level, whereas clear effects on the hair and nails were
observed at and above a level of 0.813 mg/litre.
The lowest blood-selenium levels ever reported in a general
population were seen in regions of the People's Republic of China
where Keshan disease and Kashin-Beck disease are known to be
endemic. The Task Group recognized the intensive research being
carried out concerning the involvement of selenium in the
multifactorial etiology of these diseases and the use of selenium
compounds in their prevention.
The Task Group considered a large number of studies on the
possible relationship between low levels of selenium intake and a
high incidence of cancer. In evaluating the available data, the
Task Group noted both consistencies and inconsistencies, which
made it difficult to draw a firm conclusion. However, some recent
case-control studies within prospective studies suggested the
importance of this approach for a firmer evaluation of the
involvement of the level of selenium intake in cancer prevention.
The Task Group was aware of an intervention trial being carried out
in China which could provide additional information on the
association between selenium exposure and human cancer risk.
1.1.6.2. Occupational exposure
As discussed above, the main environmental pathway of
occupational exposure to selenium is through the air, or in some
cases, by direct dermal contact. The selenium compounds likely to
be encountered include selenium dust, selenium dioxide, and
hydrogen selenide. The toxicological potential of selenium for
human beings employed in industry, can be inferred from respiratory
exposure studies carried out on laboratory animals and from ad hoc
case studies of industrial accidents.
Caution must be exercised when extrapolating the results of
animal studies or industrial accidents to the industrial health
aspects of long-term selenium exposure, because data available from
animal studies dealing with respiratory exposure are limited, the
exposure periods are brief, and industrial accidents involve
situations in which the exposure was brief and at an undetermined
level.
Extrapolation from general-population exposure is also quite
difficult, since selenium ingested in the diet may have
characteristics quite distinct from selenium encountered under
occupational exposure conditions: the chemical and physical forms
are likely to be different and these forms may change after contact
with the moist mucous membranes or with sweat. Knowledge about the
health effects of industrial selenium exposure is rudimentary,
since acute exposures are accidental and must be described on an ad
hoc case study basis. The Task Group did not know of any
epidemiological investigations of the effects of long-term
industrial exposure to selenium that included unexposed control
groups, nor were long-term follow-up studies with appropriate
control groups available. Exposure levels in short- and long-term
exposure studies are often ill-defined and the form of selenium is
not characterized. Confounding factors such as simultaneous
exposure to other toxic materials may exist.
1.2. Recommendations for Further Activities
1. The Task Group recommended that further studies were needed on
the well-identified population segments over-exposed to
selenium to clarify the signs and symptoms of selenium
overexposure in man and establish the relevant dose-response
relationships. The exclusion of hepatotoxic effects seems to
be of particular importance.
2. Further research on Keshan disease and Kashin-Beck disease
should be undertaken, not only to improve prevention of these
diseases, but also to provide basic information on the effects
of low selenium intake in man.
3. Further research is needed on the relationship between the
level of selenium intake and the incidence of cancer.
2. CHEMICAL AND PHYSICAL PROPERTIES; ANALYTICAL METHODS
2.1. Properties
Selenium belongs to group VI of the periodic table, between
sulfur and tellurium. It has similar chemical and physical
properties because the structure of its outer electron shells is
similar. Some chemical and physical properties of selenium are
listed in Table 1.
Table 1. Some chemical and physical properties of seleniuma
------------------------------------------------------------------
Properties Values
------------------------------------------------------------------
Relative atomic mass 78.96
Atomic number 34
Atomic radius 0.14 nm
Covalent radius 0.116 nm
Electronegativity (Pauling's) 2.55
Electron structure [Ar]3d104s24p4
Oxidation states -2, 0, +2, +4, +6
Stable isotopes
Mass 74 76 77 78 80 82
Natural abundance (%) 0.87 9.02 7.85 23.52 49.82 9.19
------------------------------------------------------------------
a From: Rosenfeld & Beath (1964) and Cooper et al. (1974).
All of the oxidation states of the element listed in Table 1
are commonly found in nature except the +2 state. However,
selenium compounds containing the divalent positive ion are known.
The natural isotopic pattern has been useful in identifying
selenium-containing molecular fragments produced in mass
spectrometry. While there are no naturally occurring radioisotopes
of selenium, several can be produced by neutron activation. Of
these, 75Se has the longest half-life (120 days) and is used as a
tracer in experiments as well as in the determination of selenium
by neutron activation analysis. Two relatively short-lived
radioisotopes, 77mSe (17.5-second half-life) and 81Se (18.6 min
half-life) have also been used in neutron activation analysis
(Heath, 1969-70; Alcino & Kowald, 1973).
Selenium in the +6 or selenate state is stable under both
alkaline and oxidizing conditions. Thus, it occurs in alkaline
soils, where it is soluble and easily available to plants. It is
also the most common form of the element found in alkaline waters.
Because of its stability and solubility, it may be potentially the
most environmentally dangerous form of the element.
Selenium in the +4 state occurs naturally as selenite. In
alkaline solution, it tends to oxidize slowly to the +6 state, if
oxygen is present, but not in an acid medium. It is readily
reduced to elemental selenium by a number of reducing reagents,
ascorbic acid or sulfur dioxide being commonly used for this
purpose. It readily reacts with certain o-diamines, and this is
used as the basis for some analytical methods. Selenium dioxide,
the anhydride of selenious acid, sublimes at 317°C. This is
important with regard to air pollution through the combustion of
materials containing the element (Heath, 1969-70; US NAS/NRC,
1976), and also to air sampling procedures.
Selenite binds tightly to iron and aluminium oxides. Thus, it
is quite insoluble in soils and generally not present in waters in
any appreciable amount. The nature of this binding has been
suggested to be (Howard, 1971):
(a) hydroxylation of fracture surfaces of oxides in an
aqueous environment;
(b) development of a pH-dependent charge on the surface
by amphoteric dissociation of the hydroxyl groups;
(c) electrostatic attraction of ions of negative charge
(biselenite, for instance) to the surface when it is
positively charged; and
(d) specific adsorption of the ion through exchange with
surface hydroxyl groups.
Elemental selenium (Seo), like elemental sulfur, exists in
several allotropic forms. At the molecular level, while several Se
aggregates can form, only rings containing 8 selenium atoms (Se8)
and Sen polymeric chains exist at room temperature. Se8 is soluble
in carbon disulfide, but Sen is not. Both forms have a very low
order of solubility in water and dilute acids or bases. Seo exists
in both amorphous and crystalline forms. Colloidal Seo, prepared
by reducing aqueous solutions of selenite, is an amorphous form
with a reddish colour, used in some early methods for determining
selenium at low concentrations. On heating at 60 °C, for a short
time, colloidal selenium crystallizes and its colour changes to
black. Elemental selenium boils at 684 °C, and since it, as well
as hydrogen selenide, may form during the pyrolysis of organic
materials, the use of dry ashing in preparing samples for analysis
has serious limitations. This property also contributes to the
problem of atmospheric contamination with selenium in certain
industrial processes (Crystal, 1973). Elemental selenium is very
stable and highly insoluble. It is formed on the reduction of
selenate as well as of selenite. The stability and insolubility of
elemental selenium render it unavailable to plants. Its formation
by natural processes might, thus, be considered one means by which
the element is removed from active cycling in the environment.
In its -2 state, selenium exists as hydrogen selenide and in a
number of metallic selenides. Hydrogen selenide is a strong
reducing agent and a relatively strong acid with a pKa of 3.73. It
is a gas at room temperature and very toxic. Exposure to hydrogen
selenide results in olfactory fatigue, and individuals exposed to
it may soon be unaware of its presence. In air, it decomposes
rapidly to form Seo and water. Selenides of heavy metals occur
naturally in many minerals, and iron selenide may be one of the
insoluble forms of the element in soils (Sindeeva, 1959; Nazarenko
& Ermakov, 1971; Johnson, 1976; US NAS/NRC, 1976).
A large number of selenium analogues of organic sulfur
compounds are known. Many have been identified in plants, animals,
and microorganisms. However, although some aspects of the
metabolism of selenium resemble those of sulfur, in many cases,
their metabolic pathways diverge considerably (Levander, 1976a).
As in the case of sulfur, many of the selenium compounds are
odoriferous, volatile, and relatively unstable. Because of this
volatility, precautions are necessary in handling some samples for
analysis (Klayman & Gunther, 1973; Irgolic & Kudchadker, 1974).
2.2. Analytical Methods
2.2.1. Sample collection, processing, and storage
Initially, attention must be given to the adequacy of the
method used to analyse for selenium. It is strongly recommended
that the analytical method under consideration be verified against
samples of known certified selenium content such as the standard
reference materials available commercially from the US National
Bureau of Standards. Moreover, adequate analytical quality control
measures should be standard operating practice in any laboratory
conducting analyses for selenium. Equally important is the proper
collection and treatment of samples before analysis. The sample
collected must be representative of the material being studied, and
must be protected from either contamination or loss of selenium
during analysis. These considerations are especially important in
the processing of organic matter containing selenium and in the
determination of trace amounts of selenium in materials of
environmental interest such as soils, air, and water.
In the case of soils, it should be recognized that sampling to
a depth of 1 m is more meaningful than shallower sampling,
especially where soil-selenium levels are excessively high (Olson
et al., 1942). Also, considerable variations will occur in soils
lying only a few metres from each other. In air samples, both
volatile and particulate selenium may be present and a dry filter
with an added liquid filter has been suggested for measuring these
two forms and the total selenium in air (Diplock et al., 1973;
Olson, 1976). Water samples may contain suspended solids and, in
sampling for analysis, it must be decided whether these solids,
which may contribute significant amounts of selenium, should be
included in, or excluded from, the sample. Animals are known to
synthesize volatile selenium compounds, and so it is preferable to
analyse animal specimens without drying.
2.2.2. Sample decomposition or other preliminary treatment
Some methods of analysis can be performed without destroying
the sample. In others, the sample must be treated in some way to
remove organic matter, release the selenium, and bring it to the
proper oxidation state. Several procedures for accomplishing this
have been developed and are discussed below.
2.2.2.1. Wet digestion
Wet digestion is most commonly used for freeing the selenium
and destroying the organic matter. It can be used for a wide
variety of materials, wet or dry, and with many procedures for
measuring the selenium. Different mixtures of nitric, sulfuric,
and perchloric acids, with or without such additives as hydrogen
peroxide, mercury, molybdenum, vanadium, or persulfate, have all
been used with equal success (Olson et al., 1973; Olson, 1976).
Some analysts find it convenient to add nitric acid alone to their
samples and allow them to stand overnight at room temperature.
This procedure facilitates the later steps in the wet digestion and
reduces the likelihood of foaming and/or clarring of the sample.
However, reproducible recoveries of selenium are obtained only if
all traces of nitric acid are removed from the sample digests by
heating until the appearance of perchloric acid fumes for 15-20 min
(Haddad & Smythe, 1974). The trimethylselenonium ion of urine is
somewhat resistant to decomposition by wet digestion, so an
extended period of digestion is recommended for both urine and
certain plant materials that may contain the trimethylselenonium
moiety (Olson et al., 1975). While its presence in significant
amounts has not been demonstrated in most tissues, the
trimethylselenonium moiety may occur in kidney in such an amount
that extended digestion of this tissue may also be wise.
Wet digestion readily adapts to the predigestion procedure
mentioned below and it can be adapted for the handling of large
numbers of samples (Chan, 1976; Whetter & Ullrey, 1978). Special
care is needed to avoid explosions when using perchloric acid,
including the use of a digestion rack with a fume manifold and a
special fume hood. However, a wet digestion technique with
phosphoric acid, nitric acid, and hydrogen peroxide has been
described that avoids the use of perchloric acid (Reamer & Veillon,
1983a). This procedure gives results for plasma and urine samples
similar to those obtained with the traditional fluorometric method
using perchloric acid digestion. However, the method needs to be
validated for other sample matrices.
2.2.2.2. Predigestion
A predigestion process can often reduce sampling error, when
solubilized samples or wet digestion are used in determining
selenium. A sample much larger than that required for the actual
determination is heated with about 10 volumes of concentrated
nitric acid at a slow boil for about 20 min. After cooling and
making to volume with water, an appropriate aliquot can be removed
for analysis. With samples of high fat content, the fat will rise
to the surface on cooling and it can be avoided in sampling for
analysis. The fat contains essentially no selenium at this point.
Feed samples fortified by the addition of sodium selenite or
selenate, as well as premixes containing these compounds, can often
be handled much more successfully with this procedure than by
relying on fine grinding to assure a representative sample (Olson,
1976).
2.2.2.3. Combustion
Open combustion, with or without added fixatives, and low
temperature ashing with excited oxygen have not been successfully
used in selenium determination. Using the Schoeniger flask for
destroying organic matter and oxidizing the selenium gives
excellent results, but the method is somewhat inconvenient, and
this has limited its use (Olson et al., 1973).
2.2.2.4. Fusion
Sodium carbonate or sodium peroxide fusion of soils or rocks
and the Parr bomb fusion of some organic materials have given
satisfactory results in selenium determination. However, these
methods are inconvenient in many respects and are little used
(Olson et al., 1973).
2.2.2.5. Concentration
Materials containing very small amounts of selenium may require
a step for concentrating the element. With water, this can usually
be accomplished by evaporation under alkaline conditions. Other
methods of concentrating selenium include coprecipitation with iron
hydroxide (selenite) and some of the techniques used for separating
the element from interfering substances, which will be discussed in
section 2.2.3.
2.2.3. Removal from interfering substances
In many methods, it is necessary to isolate the selenium or to
remove certain interfering substances before the selenium is
measured. Procedures used include: coprecipitation with arsenic,
tellurium, or ferric hydroxide; ion exchange column chromatography;
solvent extraction of selenium halides, of organic selenium
complexes, or of certain interfering metals; distillation of the
tetrabromide; volatilization as hydrogen selenide; precipitation
as elemental selenium; paper, thin-layer, or gas chromatography;
and the ring oven technique (Alcino & Kowald, 1973; Cooper, 1974;
Olson, 1976).
Interference by certain metal ions has been overcome, to a
large extent, by the addition of complexing reagents such as
oxalate or ethylenediaminetetraacetic acid.
2.2.4. Detection and identification of selenium
Selenium in the form of selenite or selenate can be detected
and identified in a number of ways (Alcino & Kowald, 1973). Such
qualitative tests may be useful in some industrial situations, but
normally quantitative analysis is required for confidence in any
interpretation, and in the end may only be slightly more time-
consuming.
2.2.5. Measurement of selenium
Several reviews of methods for measuring selenium in a wide
variety of materials (Nazarenko & Ermakov, 1971; Alcino & Kowald,
1973; Olson et al., 1973; Cooper, 1974; Olson 1976) have been used
as a basis for the following discussion.
The method used will depend on the sample to be analysed, the
sensitivity required, the accuracy required, other analyses to be
made, the number of samples, and the equipment and expertise
available.
For samples containing high concentrations of the element,
gravimetric or titrimetric methods may be the best. The
gravimetric methods have been based on precipitation of selenium as
the element, the sulfide, the salts of certain heavy metals, an
organic complex, or electrogravimetrically as Cu2Se. Titrations
have been based on iodometry, argentometry, potassium permanganate
reductions by selenite, or back-titrations using thiourea or sodium
thiosulfate.
Other methods, many of them very sensitive, can be classified
as: colorimetric, spectrophotometric, fluorometric, atomic
absorption spectrometric, X-ray fluorescent, gas chromatographic,
neutron activation, proton activation, polarographic, coulometric,
potentiometric, spark source mass spectrometric, gas
chromatograpic, mass spectrometric, anodic stripping voltammetric,
and catalytic.
At one time, the colorimetric determination was the most used
for the analysis of samples containing small amounts of selenium.
The method most commonly used (Robinson, 1933) was specific for
selenium and, for its time, quite sensitive. The titrimetric
method of Klein (1943) superceded it, being somewhat more
reproducible and sensitive, until selenium was found to have a role
as an animal nutrient, when considerably greater sensitivity was
required. Very sensitive methods, most commonly used, include
fluorometry, neutron activation, and atomic absorption
spectrophotometry.
2.2.5.1. Fluorometric analysis
The methods most widely used for the determination of selenium
in natural materials are based on fluorometry. Selenious acid
reacts with 1-diamines to give a piazselenol, which is fluorescent.
The 1-diamine of choice is 2,3-diaminonaphthalene. The piazselenol
is extracted from an acid solution (pH 1 - 2) with either
decahydronaphthalene or cyclohexane, in either of which the
fluorescence yield is good. When the organic matter is destroyed,
the reaction can be specific for selenium. Interference from
copper, iron, and vanadium, and some oxidizing agents can usually
be avoided.
Fluorometric methods have been applied to a number of
materials; those for foods and plants have been subjected to
collaborative studies (Olson, 1969; Ihnat, 1974) and have been
accepted as official first, and final action, by the Association
of Official Analytical Chemists (AOAC, 1975). Critical reviews of
fluorometric methods (Haddad & Smythe, 1974; Michie et al., 1978)
have shown them to be reliable providing that the precautions
described for various adaptations are followed. The methods are
sensitive to about 0.01 µg of selenium in a sample, are adaptable
to large numbers of samples, can be automated (Brown & Watkinson,
1977; Szydlowski & Dunmire, 1979; Watkinson, 1979; Watkinson &
Brown, 1979) and can be simplified (Spallholz et al., 1978; Whetter
& Ullrey, 1978). While isotope dilution procedures can be used
with fluorometric methods (Cukor et al., 1964), they are not
essential for reliable results. Fluorometric analyses rely on
perchloric acid digestion of the samples, and, thus, require a
perchloric acid fume hood. Furthermore, 2,3-diaminonaphthalene of
adequate purity is sometimes difficult to obtain. Improved
sensitivity of the fluorometric analysis of water and biological
samples was achieved by redistilling cyclohexane and extracting
2,3-diaminonaphthalene solution in 0.1 M hydrochloric acid with two
portions of redistilled cyclohexane (Nazarenko et al., 1975;
Nazarenko & Kislova, 1977). Wilkie & Young (1970) have also
published detailed experimental procedures for purifying various
reagents used in fluorometric analysis. The analysis can be
carried out with a simple fluorometer, is adaptable to a wide
variety of materials, is reasonably rapid and reliable, and is
relatively inexpensive to perform.
2.2.5.2. Neutron activation analysis
Thermal neutron activation is the most commonly used procedure
for irradiating samples containing selenium. Of the radionuclides
of selenium that it produces, 75Se (half-life 120 days), 81Se
(half-life 18.6 min), and 77mSe (half-life 17.5 seconds) have been
useful in analysis. 81mSe (half-life 57 min) has been used to
measure carrier selenium after separation and reirradiation
(Kronborg & Steinnes, 1975). Proton-induced X-ray fluorescence
methods have also been reported (Bearse et al., 1974; Barrette et
al., 1976).
Two methods of converting radioactivity measurements to
selenium content are:
(a) the direct method using calculations based on certain
known values and constants; and
(b) the comparator method, based on the irradiation and
counting of a known amount of selenium, as a
standard, together with the samples. The comparator
method is the most accurate and most often used, but
the direct method is often used when short-lived
isotopes are measured. Ruthenium has been
recommended as a multi-isotopic comparator when
multi-element analysis is being performed (Van der
Linden et al., 1974).
Neutron activation analysis can be used without and with sample
destruction. When 75Se is measured by non-destructive analysis, a
period of several weeks or months may be allowed for decay of some
interfering radioisotopes. The 77mSe isotope has been used without
this type of delay with some success. Non-destructive methods have
been used for multi-element analysis, but they have been subject to
more errors, because of interference, than destructive methods
followed by chemical separation. The use of high resolution gamma-
ray spectrometry substantially increases the specificity of the
non-destructive methods and the complete processing of 300 - 400
samples per month is possible (Pelekis et al., 1975). A method has
been developed using irradiation with epithermal neutrons and
multiparameter analysis which eliminates the effect of spurious
interferences in the determination of low selenium concentrations
by instrumental neutron activation analysis. High specificity has
been achieved by the selective response of the system to gamma-
gamma coincidences of 75Se (Vobecky et al., 1977, 1979; Pavlik et
al., 1979).
Destructive neutron activation analysis is used with the
chemical separation of the selenium. In most methods, 75Se is
measured, but 77mSe measurement is not uncommon, and 81Se
measurement has also been used. Carrier selenium is usually added
following the irradiation. Destruction of the sample has been
accomplished by wet digestion, sodium peroxide fusion, or
combustion. Combustion has been accomplished in a closed system,
which provides for the dry distillation of selenium and certain
other elements into a trapping system. Wet digestion is normally
followed by separation of the element by distillation as the
tetrabromide, and/or its precipitation, adsorption on an ion
exchange material, solvent extraction of an organic selenium
complex, or reversed phase extraction chromatography. A semi-
automated method based on wet digestion, separation by
distillation, and the use of ruthenium as a comparator has been
reported (D'Hondt et al., 1977).
Neutron activation analysis can be very accurate, sensitive,
and specific, especially when used with sample destruction and
chemical separation of the selenium. It adapts well to multi-
element analysis. However, it requires sophisticated equipment
that most laboratories do not have. Furthermore, it is time-
consuming, especially when used with chemical separation. Its most
important use may be as a reference method against which other
methods may be evaluated or where good reagents are not available.
2.2.5.3. Atomic absorption spectrometry
Methods for selenium analysis now being most actively studied
are those based on atomic absorption spectrometry (AAS). A wide
variety of techniques has been described. Most require some type
of wet digestion, when organic matter is present. Flame
atomization methods are useful for materials with a high selenium
content. However, other more sensitive methods have been
developed, the most used being based on hydrogen selenide
generation. This has the advantages of separating the selenium
from many other elements and measuring it using techniques that
provide excellent sensitivity. Ihnat (1976) compared the
performance of hydride generation with that of a carbon furnace
atomization-AAS method, finding hydrogen selenide generation to be
superior. In further studies, Ihnat & Miller (1977a,b) found that
the sensitivity of this method was excellent and its accuracy,
fairly good, but that its precision in a collaborative study was
not. McDaniel et al. (1976) found that procedures for hydrogen
selenide generation might liberate as little as 10% of the element
from solution. They suggested means for optimizing hydride
generation to give a simple and sensitive selenium determination,
free from interferences, and using a heated graphite atomizer for
the measurement of the element. The hydride generation technique
has been automated (Goulden & Brooksbank, 1974; Pierce & Brown,
1977; Pyen & Fishman, 1978), and improved equipment for hydride
generation has recently been described (Brodie, 1979). Encouraging
results with graphite furnace atomization in the presence of nickel
(Shum et al., 1977) indicate that this method may also find
considerable use.
Atomic absorption methods based on Zeeman effect background
correction to remove spectral interferences (Pleban et al., 1982;
Carnrick et al., 1983) are currently receiving attention. Such
procedures offer the promise of carrying out selenium analyses in
certain biological matrices directly, i.e., without the need for
sample digestion. Development of such techniques would greatly
increase the speed and convenience of selenium analysis.
2.2.5.4. Other methods
Examples of other methods that have received recent attention
and may find some more use in the future include: energy
dispersive X-ray fluorescence (Holynska & Markowicz, 1977);
differential pulse polarography (Bound & Forbes, 1978); anodic
(Andrews & Johnson, 1976) or cathodic (Blades et al., 1976)
stripping voltammetry; atomic fluorescence spectrometry (Thompson,
1975); gas-liquid chromatography (Ermakov, 1975; Poole et al.,
1977); and spark source mass spectrometric isotope dilution
(Paulsen, 1977). A gas chromatographic, mass spectrometric double
isotope dilution technique based on a volatile chelate of selenium
has recently been described, which not only determines total
selenium in biological materials but also allows stable selenium
isotopes to be followed as tracers in metabolic studies (Reamer &
Veillon, 1983b).
A variety of methods has been proposed and used for identifying
and measuring a number of chemical forms of selenium in different
materials including: paper (Hamilton, 1975) or ion exchange (Martin
& Gerlach, 1969) chromatography for selenium compounds in plants;
gas chromatography for volatile selenium compounds (Doran, 1976);
and ion exchange (Shrift & Virupaksha, 1965) or solvent extraction
separation (Kamada et al., 1978) for measuring selenite and
selenate in solutions.
3. SOURCES, TRANSPORT, AND CYCLING OF SELENIUM IN THE ENVIRONMENT
3.1. Natural Sources
Because selenium is present in natural materials, its
occurrence in any substance cannot be assumed to be the result of
human activity and some knowledge of its natural distribution is
required before evaluating its role as a pollutant.
3.1.1. Rocks and soils
Several reviews of selenium geochemistry have been published
(Sindeeva, 1959; Rosenfeld & Beath, 1964; Lakin & Davidson, 1967;
Cooper et al., 1974). The concentration of selenium in igneous
rocks is low, usually much less than 1 mg/kg, and similar levels
probably occur in metamorphic rocks. Sedimentary rocks, such as
sandstone, limestone, phosphorite, and shales may contain from < 1
to > 100 mg/kg.
The selenium content of a soil reflects, to some extent, that
of the parent material from which the soil has been formed. Thus,
in arid and semi-arid areas, soils of high selenium content have
been derived from sedimentary rocks, usually shales and chalks
(Moxon et al., 1950). These soils are alkaline in reaction,
favouring the formation of selenate (Geering et al., 1968), which
is readily available to plants (Moxon et al., 1950). The selenate
is easily leached from the surface soil, but, with limited
rainfall, it is redeposited in the subsurface soil, where it is
still available to plants (Olson et al., 1942). Thus, surface soil
analysis has not been found to be a reliable measure of the
potential of a soil to produce vegetation containing toxic levels
of selenium.
Coal with an unusually high selenium content (> 80 000 mg/kg;
average 300 mg/kg) was recently identified as the ultimate
environmental source of selenium contaminating soils in a
seleniferous region of Enshi county, Hubei province, the People's
Republic of China (Yang et al., 1983). It was thought that
selenium passed from the coal to the soil through weathering,
leaching, and possibly biological action, thus making it available
to the crops. Lime fertilizers, traditionally used in the area,
would also render the selenium accumulated in the soil more readily
available to plants.
Some soils produce crops nutritionally deficient in selenium
(US NAS/NRC, 1971). The most obvious factor in the formation of
these soils is probably the parent material (Hodder & Watkinson,
1976), but other factors such as rainfall, climate, pH, and soil
composition contribute significantly (Gissel-Nielsen, 1976, 1986).
As with producing plants of high selenium content, soils producing
plants low in the element cannot be identified by analysis of the
surface soil alone (Shacklette et al., 1974). Because of the many
factors affecting availability, plant analysis is also necessary
(Kubota et al., 1967).
Sun et al. (1985) reported that the average total selenium
content (112 µg/kg, range 59 - 190 µg/kg) in the soil of 6 low-
selenium Keshan disease areas in China was significantly lower than
that of the 5 corresponding non-endemic areas (234 µg/kg, range 142
- 318 µg/kg). The average water-soluble selenium content and the
percentage of water-soluble selenium in the total soil-selenium of
the endemic areas (4.0 µg/kg, range 2.2 - 8.7 µg/kg and 39 g/kg,
respectively) were also significantly lower than those of the non-
endemic areas (19.9 µg/kg, range 11.4 - 38.8 µg/kg and 9.2 g/kg,
respectively). The total soil-selenium content of a high-selenium
area in China was 7865 µg/kg (range 6390 - 10 660 µg/kg), but the
percentage of water-soluble selenium was not very high (30 - 40
g/kg).
3.1.2. Natural selenium in the food chain
All plants absorb selenium from the soil, the amount depending
mainly on the species, the stage of growth, and the availability of
the selenium in the soil. Biogeochemical factors influencing the
availability of selenium in the soil, such as pH, iron content,
etc., have been reviewed by Ermakov & Kovalskij (1968), Allaway
(1973), and Kovalskij (1974, 1978). Various plant tissues contain
different selenium levels, generally following protein content.
Certain plants, known as selenium accumulator plants, take up
quantities of selenium great enough to be toxic for animals. The
selenium content of animal tissues reflects that of the feeds
consumed (Allaway, 1973; Kovalskij & Ermakov, 1975). Thus,
naturally-occurring selenium readily passes up through the food
chain via animals to human beings, and the amount of selenium in
the human diet is largely determined by the amount of selenium in
the soil available for absorption by plants. However, as discussed
in section 7, the increased retention of selenium by animals, under
conditions of low selenium intake, and the decreased retention
during high selenium intake ensures that, in different geographical
zones, the range of selenium levels in animal products is less than
that in plant tissues. This "buffering effect" of animals in the
food chain tends to moderate the extremes of selenium intake to
which human beings are exposed, whereas herbivores are subject to
much greater fluctuations in selenium intake.
3.1.3. Water and air
Under natural conditions, the concentration of selenium in
water usually ranges from a few tenths to 2 or 3 µg/litre (Ermakov
& Kovalskij, 1974; US NAS/NRC, 1976). The highest natural
concentration reported to date is 9000 µg/litre, almost all other
values falling below 500 µg/litre (US NAS/NRC, 1976). A garlicky
odour has been noted in waters containing 10 - 25 µg selenium/litre
and an astringent taste can be detected in water samples containing
100 - 200 µg/litre (Pletnikova, 1970). In a study carried out in
Nebraska, USA, which was biased towards finding waters with
elevated selenium contents, it was reported that about one-third of
161 well samples contained over 10 µg selenium/litre and about 4%,
over 100 µg/litre (Engberg, 1973). Surface waters seem much less
likely to contain excessive levels of selenium than ground waters.
Soils, plants, microorganisms, animals, and volcanoes all
contribute selenium to the atmosphere. All of these sources,
and possibly also certain sediments, produce volatile forms,
and soils and volcanoes probably contribute particulate matter
containing the element. Some man-made activities also yield
atmospheric selenium (section 3.2.2), and it is difficult to
establish the proportion of selenium that comes from these
activities and the proportion that comes from natural sources
(US NAS/NRC, 1976). However, data concerning atmospheric
selenium over the poles and over the Atlantic Ocean, as well
as over areas of minimal human activity (Zoller et al., 1974;
Duce et al., 1975; Dams & De Jonge, 1976) suggest that the
average concentration from natural sources should be less than
0.04 ng/m3, except close to centres of volcanic activity.
3.2. Man-Made Sources
3.2.1. Agriculture
Early agricultural uses of selenium compounds as pesticides
were very limited and short-lived. The recent use of selenium
compounds as feed additives or injectables for the prevention of
selenium deficiency diseases in farm animals represents a source
for environmental contamination, but, compared with the levels
already present in most feeds and the amounts found in soils, even
in the deficient areas, this source seems insignificant (US
NAS/NRC, 1976). In New Zealand, little increase in the selenium
content of human foods was observed after the introduction of
selenium supplementation for farm animals (Thomson & Robinson,
1980). The use of fertilizers (Koivistoinen & Huttunen, 1985) and
foliar sprays (Gissel-Nielsen, 1973, 1986) to rectify selenium
deficiency in feeds is being introduced, but they are unlikely to
become a hazard in the environment. Other fertilizers do not
contain enough selenium to contribute significantly to the
environment (Gissel-Nielsen, 1971). Irrigation of seleniferous
lands may result in an increased concentration of selenium in
drainage waters (US NAS/NRC, 1976), but until recently there was no
evidence to suggest that this constitutes a hazard. However, in
the San Joaquin Valley of California, reproductive problems have
now been observed in aquatic birds that used irrigation drain-water
ponds as waterfowl habitat (Ohlendorf et al., 1986). Although the
toxic signs reported in these birds resembled avian selenosis, the
possible role of other water-borne toxins such as high levels of
arsenic or boron has been pointed out.
3.2.2. Industry
The main industrial sources emitting selenium into the
environment include the mining, milling, smelting, and refining of
copper, lead, zinc, phosphate, and uranium, the recovery and
purification of selenium itself, the use of selenium in the
manufacture of various products, and the burning of fossil fuels.
Some problems from selenium emissions might arise in regions near
selenium-producing or coal-burning industries. For instance, in
areas where copper sulfide ores were mined and processed,
atmospheric selenium levels ranging from 0.15 to 6.5 µg/m3 were
reported within 0.5 - 10 km of the ore-processing plant (Seljankina
et al., 1974). Selenium pollution of the air may be important with
regard to possible contamination of open-air water reservoirs;
waterways can be contaminated directly by selenium from mining or
industrial effluents. For example, waste-waters of ore mines and
effluents from a number of non-ferrous industries were reported to
contain selenium in the concentration range of 14 - 56 µg/litre.
Residual selenium impurities in conditionally cleaned effluents
released into open water bodies can contribute some contamination
(Shumaev et al., 1976). Effluents from sewage plants can add
selenium to waters; raw sewage has been reported to contain up to
280 µg selenium/litre, and levels of 45 - 50 µg/litre have been
reported in primary and secondary sewage effluents (Baird et al.,
1972). Emission factors for atmospheric selenium contamination and
solid waste generation for various industrial processes in the USA
are shown in Table 2. The pattern of such emissions will vary
depending on the industrial and mining characteristics of
individual countries. In Canada, for example, over half of the air
pollution due to selenium resulted from primary copper and nickel
production, the combustion of coal contributing less than 25%.
When coal from deposits throughout the USA was analysed (138
analyses), the highest selenium content was 10.65 mg/kg and the
average, was 2.8 mg/kg, a hundred times less than the high-selenium
coals reported in China (Yang et al., 1983) (section 3.1.1). The
chemical form of the element in coal is not known (US NAS/NRC,
1976). The selenium from the coal was released into the
surrounding soils by natural processes, thereby contaminating the
food chain with toxic levels of selenium.
Few data on the selenium content of fuel oils are available.
Most reported values have been below 1 mg/kg (US NAS/NRC, 1976).
Oil shale has been reported to contain selenium (Shendrikar &
Faudel, 1978), but apparently there are no reports of its
occurrence in natural gas.
3.3. Environmental Transport
Because of the lack of quantitative data, it is impossible to
evaluate the relative importance of various processes in the
environmental transport of selenium. However, Bertine & Goldberg
(1971) have estimated that rivers move over 7 x 106 kg of selenium
every year, even though surface waters generally contain little of
the element. The quantity of selenium passing via grain into the
world feed and food supplies would be at least 105 kg, if it is
assumed that the annual global production of cereal grains and corn
is 109 tonnes and that these grains contain an average of 0.1 mg
selenium/kg. Although this simple calculation neglects all other
plants and the large ocean flora, the quantity of selenium moved in
grains alone is equivalent to about 7% of the average world
production of the element for industrial purposes. Thus, natural
geophysical and biological processes probably play dominant roles
in shaping the present status of selenium in the human environment
and must be taken into account in any evaluation of the effects of
man's activities on selenium in the environment.
Table 2. Estimated selenium materials balance for the
USA for 1970a,b
---------------------------------------------------------
Source Selenium Atmospheric Solid
input emissions waste
---------------------------------------------------------
(---------tonne----------)
Industrial production
Mining and milling 2900 5 1633
Smelting and refining 1270 227 181
Selenium refining 877 59 362
Industrial consumption
Glass and ceramics 59 59
Electronics and duplicating 227 c c
Pigments 59 c c
Iron and steel alloys 23 11 c
Other 38 c c
Other
Coal consumption 1315 680 635
Fuel oil consumption 59 59 0
Incineration 36 c 36
Other disposal 322 c 318
Total emissions: 1100 3165
---------------------------------------------------------
a From: US NAS/NRC (1976).
b The input not accounted for as atmospheric emissions or
solid waste would appear in intermediate or commercial
products.
c Less than 1 tonne.
3.4. Biological Selenium Cycle
Several diagrammatic schemes have been presented for the
cycling of selenium in nature (Moxon et al., 1950; Lakin &
Davidson, 1967; Olson, 1967; Frost & Ingvoldstad, 1975; US NAS/NRC,
1976). None of the schemes is complete and none provides good
quantification for the various steps in the cycle. Indeed, a
complete scheme is perhaps too complex to be presented
diagrammatically with clarity, and reasonably accurate
quantification of the various steps would demand much more
background information than is available at present. It appears
that selenium has many pathways for distribution among geological
media, biological media, the atmosphere, and waters by geophysical,
biological, and man-made activities. It also appears that there
are "sinks" into which selenium may fall (elemental selenium,
ferric hydroxy complexes, and metal selenides) from which it may be
only very slowly recycled. These may represent a natural
detoxification process. However, the fact that there are areas of
deficiency and areas of excess of the element over the earth's
surface is evidence that its depletion or concentration can occur
locally. How quickly changes in the selenium status of a region
occur is not known, and there is a distinct need for data from
which an evaluation could be made of the rate and impact of these
changes as well as the influence of man's activities on them. For
example, Allaway (1973) concluded that any soil-plant-animal chain
of food production on neutral or acid soils would eventually become
depleted in biologically effective selenium, though it is not known
how rapidly this might occur. Shrift (1973) suggested the
existence of a biological selenium cycle from the standpoint of
transformations between several oxidation-reduction states, as
shown in Fig. 1. Reduction of the element is common and well
established, but its biological oxidation is less well documented.
However, a missing link on the oxidative side of the biological
selenium cycle was found by Sarathchandra & Watkinson (1981), when
they discovered a strain of Bacillus megaterium, isolated from
soil, that oxidized elemental selenium to selenite and traces of
selenate. The role of non-biological oxidation and reduction of
the element has not yet been elucidated. While a balance may be
maintained in the redox state of selenium throughout the world,
there are regions where oxidation prevails and others where
reduction prevails. Knowledge of the conditions and mechanisms
leading to each situation is essential to understand the
development of regions of selenium excess or deficiency.
4. LEVELS IN ENVIRONMENTAL MEDIA
4.1. Levels and Chemical Forms of Selenium in Food
4.1.1. Levels in food
In spite of the relatively few data available, some
generalizations concerning the selenium content of foodstuffs can
be made. For example, the level of selenium in food depends on
natural differences among food commodities and the natural
availability of selenium in the environment. Moreover, certain of
man's activities can influence the selenium content of human foods.
4.1.1.1. Natural differences among food commodities
A wide range of values has been reported in the selenium
content of foods (mg/kg wet weight): liver, kidney, and seafood,
0.4 - 1.5; muscle meats, 0.1 - 0.4; cereal and cereal products, <
0.1 - > 0.8; diary products, < 0.1 - 0.3; and fruits and
vegetables, < 0.1) (Oelschlager & Menke, 1969; Morris & Levander,
1970; Schroeder et al., 1970; Suchkov, 1971; Arthur, 1972; Millar &
Sheppard, 1972; Ferretti & Levander, 1974; Sakurai & Tsuchiya,
1975; Abutalybov et al., 1976; Bieri & Ahmad, 1976; Kasimov et al.,
1976; Olson et al., 1978). It should be pointed out that these
values are given for raw foods (food as purchased) rather than
cooked foods (food as eaten). The effects of cooking on the
selenium content of foods are described in section 4.1.1.3.
Moreover, the values presented in these food composition studies
should not be compared from one country to another, because of the
variations in the analytical methods and sampling procedures used
(section 2).
Organ meats, such as kidneys or liver, contain the highest
levels of selenium, but some seafood products contain almost as
much. Muscle meats are significant sources of selenium, though
they do not contain as much as organ meats or seafoods. Certain
semolina, grain, and cereal products can contribute appreciably to
the dietary-selenium intake, but wide variations in selenium
content have been found in different samples of the same foodstuff.
Such variation is typical of many plant foods and the reasons for
this are discussed below. Milk, cheese, and egg samples from
several countries showed low to moderate values for selenium, but
again the results were quite variable. Fruits and vegetables
generally contained very low levels of selenium, though garlic and
mushrooms contained moderate levels of the element.
The selenium contents of baby foods tended to show the same
general pattern as those in adult foods (Morris & Levander, 1970;
Arthur, 1972), i.e., meat and cereal products contained the highest
levels and fruit and vegetable products, the lowest. Shearer &
Hadjimarkos (1975) analysed samples of mature human milk from 241
subjects living in the USA and found that the overall mean selenium
content was 0.018 mg/litre. Over 98% of the samples contained
between 0.007 and 0.033 mg/litre. Grimanis et al. (1978) reported
an average of 0.015 mg/litre in 5 samples of mature human milk
collected in Greece. These authors found slightly higher values in
15 samples of human transitional milk (0.016 mg/litre) and
colostrum (0.048 mg/litre). A slightly lower average selenium
concentration of 0.013 mg/litre, was reported for samples of human
transitional milk in New Zealand (Millar & Sheppard, 1972). The
most extreme values for the selenium content of human milk were
reported from China and ranged from 0.0026 mg/litre in areas where
Keshan disease was prevalent to 0.283 mg/litre in high-selenium
areas (Yang et al., 1986).
It can be seen from Table 3 that meat-based infant formulae
have a higher selenium content than formulae based on milk or soy
protein, and that casein-based powdered formulae for special
medical purposes (diet therapy for errors in amino acid metabolism
or malabsorptive states) contain low levels of selenium. Blended
food tube-feeding formulae containing meat tend to have higher
selenium contents than comparable products containing only milk or
casein and soy protein (Table 3). Chemically-defined diets having
egg albumen as the protein source contained more selenium than
diets based on casein hydrolysate, which, in turn, contained more
selenium than diets based on purified amino acids. Total
parenteral nutrition solutions based on casein hydrolysate also
contain more selenium than solutions based on amino acid mixtures,
which contain very low levels of selenium.
Table 3. Selenium content of commercial formula dietsa
-----------------------------------------------------------------
Product Selenium content
(mg/kg wet (mg/kg dry
weight) weight)
-----------------------------------------------------------------
Infant formulae:
milk-based 0.004 - 0.027
soy-based 0.004 - 0.030
meat-based 0.046 - 0.070
casein-based formulae for special 0.048 - 0.120
medical purposes (powders)
Food supplements and tube-feeding
formulae:
milk-based 0.005 - 0.045
soy-casein-based 0.013 - 0.020
blended foods 0.037 - 0.056
Chemically defined diets:
(powders)
egg albumen low residue 0.224 - 0.351
egg albumen moderate nitrogen 0.388 - 0.886
egg albumen high nitrogen 0.503 - 0.570
casein hydrolysate 0.052 - 0.071
amino acid mixture 0.001 - 0.011
-----------------------------------------------------------------
Table 3. (contd.)
-----------------------------------------------------------------
Product Selenium content
(mg/kg wet (mg/kg dry
weight) weight)
-----------------------------------------------------------------
Total parenteral nutrition
solutions:
casein hydrolysate 0.037 0.324
(0.032 - 0.041)
diluted 1:1 with 50% dextrose 0.019 0.093
solution (0.017 - 0.020)
amino acid mixture 0.001 0.010
-----------------------------------------------------------------
a Adapted from: Zabel et al. (1978).
One of the factors that can influence the amount of selenium in
plant foodstuffs is the nature of the plant itself. Plants have
been divided into 3 groups depending on their tendency to take up
selenium from seleniferous soils (Rosenfeld & Beath, 1964):
Group 1 Primary selenium accumulators - can contain very
high amounts of the element (often over 1000
mg/kg dry weight).
Group 2 Secondary selenium accumulators - rarely contain
more than a few hundred mg/kg.
Group 3 Many weeds and most crop plants, grains, and
grasses - rarely contain more than 30 mg/kg, even
when grown on seleniferous soils; when grown on
normal soils generally contain less than 1 mg/kg.
Plants from Groups 1 and 2 do not usually contribute to the
selenium intake of human beings, since they are not consumed
directly by people and are consumed by animals only when other
feeds are not available. However, plants from Group 3 can
contribute large amounts of selenium, if they are grown in
seleniferous areas (see below).
The concentrations of many nutritionally essential trace
elements are known to be decreased by the milling of grains into
cereals (Czerniejewski et al., 1964) and preliminary analysis of
random supermarket samples of cereal foods suggested that refined
products such as white flour or white bread contained less selenium
than whole grain foods such as whole wheat flour or whole wheat
bread (Morris & Levander, 1970). However, a subsequent more
controlled analytical study in which various grain fractions were
taken from the same production batch showed that milling a variety
of grains decreased the concentration of selenium in the consumer
product by only 10 - 30% (Ferretti & Levander, 1974). Thus, the
selenium content of cereal products is somewhat less than that of
the parent grains but the decreases in concentration are not nearly
as great as those observed with other essential trace minerals.
Selenium tends to be localized mainly in the protein fraction
of plant and animal tissues; thus, the protein content of a food
also influences its selenium content. For example, the
concentration of selenium in a series of soybean products prepared
from a given lot of soybeans increased as the protein content of
the product increased (Ferretti & Levander, 1976). However,
protein content is only an expression of the potential of a food to
contain selenium, and a food high in protein will not necessarily
also be high in selenium. Moreover, non-protein seleno amino acids
occur in some foods and thus, the protein content may not always
reflect the selenium content.
4.1.1.2. Effects of natural differences in the availability of
selenium in the environment on levels in food
The most important factor in determining the selenium content
of plant foods and feeds is the amount of selenium in the soil that
is available for uptake by the plant. Sun et al. (1985) reported
that the selenium content of local crops was significantly
correlated with the selenium content of the local soil, the
correlation coefficients of soybean, corn, and rice being 0.9456,
0.9953, and 0.9954, respectively. Since the water-soluble selenium
in the soil was directly correlated with the pH and inversely
correlated with humin and total iron in the soil, the selenium
content of local crops was also influenced by the amount of water-
soluble selenium in the soil. An analytical survey carried out in
the USA demonstrated that the level of selenium in alfalfa plants
varied from less than 0.01 to more than 5.0 mg/kg, and it was
assumed that these levels reflected the amount of available
selenium in the soil (Kubota et al., 1967). A variation in the
selenium content of wheat from 0.04 to 21.4 mg/kg, depending on
where the plant was grown, was reported by Schroeder et al. (1970).
Samples of several foods bought in local markets in Caracas,
Venezuela contained much more selenium than similar foods purchased
in supermarkets in the eastern USA (Table 4). The likely source of
selenium in the milk, eggs, and meat from Caracas is sesame cake,
since sesame is produced mostly in the seleniferous area of
Venezuela and the pressed cake is widely used as an ingredient in
animal feed. Selenium levels as high as 14 mg/kg in corn and 18
mg/kg in rice have also been observed in certain food samples taken
from high-selenium regions in Venezuela (Jaffe, 1976). These
concentrations of selenium are as high as those reported in foods
from seleniferous zones of the USA (Smith & Westfall, 1937).
Great extremes in the selenium content of staple foods have
also been reported recently from China (Yang et al., 1983). For
example, samples of corn, rice, and soybeans, taken from a high-
selenium area with a history of human intoxication reported as
chronic selenosis (section 8.1.1.1) contained average selenium
levels of 8.1, 4.0, and 11.9 mg/kg, respectively, whereas samples
of the same staples collected in a low-selenium area where Keshan
disease was prevalent (a human selenium-deficiency disease)
(section 8.2.2) contained average selenium levels of only 0.005,
0.007, and 0.010 mg/kg, respectively (Table 5). A low selenium
content in food has also been reported in other countries with low
selenium soils, such as New Zealand and Finland. Typical ranges
for selenium contents in food in such countries include (mg/kg wet
weight): liver, kidney, and seafood, 0.09 - 0.92; muscle meats,
0.01 - 0.06; cereal and cereal products, 0.01 - 0.07; milk, <
0.01; and fruits and vegetables, < 0.01 - 0.02 (Koivistoinen,
1980; Thomson & Robinson, 1980). Samples of staple foods collected
in areas where soils contained intermediate levels of selenium also
contained intermediate levels of selenium.
Table 4. Comparison of selenium contents
of selected foods available in Caracas,
Venezuela, and Beltsville, Maryland, USA
(mg selenium/kg wet weight)a
-----------------------------------------
Food Caracas Beltsville
-----------------------------------------
Powdered milk 0.417 0.169
Whole milk 0.115 0.012
American-type cheese 0.425 0.090
Swiss-type cheese 0.382 0.104
Pork 0.833 0.209
Chicken 0.702 0.106
Egg 1.520 0.116
-----------------------------------------
a From: Mondragon & Jaffe (1976).
As might be expected, the selenium contents of food products of
animal origin depend heavily on the amount of naturally-occurring
selenium in the feed given to the animal. In the USA, it was shown
that the selenium content of swine muscle was highest in areas
known to have a high level of available selenium in the soil and
lowest in areas in which the available soil-selenium was low (Ku et
al., 1972). This indicates that selenium is readily passed up the
soil-plant-animal food chain to human beings.
4.1.1.3. Man-induced changes in selenium levels in food
(a) Human activities that increase selenium levels
Perhaps the most direct way that man's activities can increase
the selenium content of the food supply is the deliberate addition
of selenium to the feeds of poultry and certain livestock. In some
countries, this is now accepted practice to prevent the occurrence
of selenium-deficiency diseases, many of which cause significant
economic losses to farmers throughout the world. In the USA, for
example, farmers are permitted to add 0.1 mg selenium/kg (as sodium
selenite or selenate) complete feed for beef and dairy cattle,
sheep, chickens, ducks, swine (0.3 mg/kg in starter and prestarter
rations), and 0.2 mg/kg for turkeys (US Department of Health and
Human Services, Food and Drug Administration, 1984; Subcommittee on
Selenium - Committee on animal Nutrition, 1983). It has been shown
that the edible tissues of poultry, swine, and sheep, fed diets
fortified with inorganic selenium salts at the regulated levels,
did not contain any more selenium than the tissues of animals fed
diets, naturally adequate in selenium (Allaway, 1973; Ullrey et
al., 1977). The homeostatic mechanisms that appear to limit the
concentrations of selenium in the edible tissues of animals fed
certain levels of sodium selenite are discussed further in section
6.
Table 5. Selenium contents of staple foods grown on soils in areas of China with excess,
moderate, and deficient levels of seleniuma
-----------------------------------------------------------------------------------------
Place Corn Rice Soybean
Number Se content Number Se content Number Se content
of (mg/kg) of (mg/kg) of (mg/kg)
samples samples samples
-----------------------------------------------------------------------------------------
High-selenium area 44 8.1 22 4.0 17 11.9
with a history of (0.5-28.5)b (0.3-20.2)b (5.0-22.2)b
intoxication reported
as chronic selenosis
High-selenium area 2 0.57 2 0.97 2 0.34
reported to be without
selenosis
Moderate-selenium- 82 0.036 76 0.035 31 0.069
adequate area (± 0.056) (± 0.027) (± 0.076)
(Beijing)
Low-selenium area 10 0.009 32 0.022 - -
(± 0.009) (± 0.009)
Low-selenium area 195 0.005 49 0.007 150 0.010
with Keshan disease (± 0.003) (± 0.003) (± 0.008)
-----------------------------------------------------------------------------------------
a Adapted from: Yang et al. (1983).
b Mean ± SD or range shown in parenthesis.
Finland is the first country to decide to increase the selenium
content of Finnish feed and food by the addition of sodium selenate
to fertilizers, to be used in the whole country at a concentration
of 16 or 6 mg/kg for cereal and grassland crops, respectively
(Koivistoinen & Huttunen, 1985). The manufacture of these
selenized fertilizers began in the summer of 1984 and they will be
used at application rates of 10 g selenium/ha per growing season.
It is anticipated that the added selenium in the fertilizers will
be transported to the human food chain, and the selenium level of
the cereal and grassland crops will be raised to 0.1 and 0.15 -
0.20 mg/kg (dry basis), respectively.
Another potential way in which the level of selenium in the
food chain might be increased by man's activities is the proposed
use of selenium-bearing fly ash as a soil supplement. Furr et al.
(1977) analysed cabbages grown on potted soil supplemented with
several different fly ashes and found that the selenium levels in
the cabbages were closely correlated with those in the respective
fly ashes in which the plants were cultured. The ready
bioavailability to animals of the selenium in plants grown on fly
ash was demonstrated in a study (Stoewsand et al., 1978) in which
Japanese quail were fed a complete diet containing 60% winter wheat
that had been grown to maturity on either soil or a deep bed of fly
ash. The soil contained 2.1 mg selenium/kg and the wheat grown
thereon contained 0.02 mg/kg, a level typical of wheat from
selenium-deficient areas. The fly ash contained 21.3 mg
selenium/kg, and the wheat grown on it contained 5.7 mg/kg, a level
sometimes found in wheat from naturally seleniferous areas. The
levels of selenium in the tissues of the quail fed the wheat grown
on fly ash were much higher than those of quail fed the wheat grown
on soil (Table 6). Moreover, the selenium contents of eggs from
quail fed the wheat grown on fly ash were 3.5 mg/kg in the yolk and
almost 10 mg/kg in the white compared with 0.5 and 0.2 mg
selenium/kg, respectively, in eggs from quail fed the control
wheat. Obviously, the use of selenium-bearing fly ash as a soil
supplement to provide nutritionally desirable levels of selenium in
plants should be carried out with care, since inappropriate use
could lead to an unwanted build-up of selenium residues in the food
chain. Also, the application of fly ash to soil at rates
sufficient to correct selenium deficiency in animals may damage the
soil.
Table 6. Selenium in tissues of male Japanese quail fed winter
wheat grown on soil or fly asha
-------------------------------------------------------------------
Wheat grown Tissue selenium
on: brain heart kidney liver muscle
-------------------------------------------------------------------
(mg/kg dry weight)
Soil 0.8 ± 0.1 0.7 ± 0.2 3.6 ± 0.4 1.6 ± 0.0 0.3 ± 0.2
Fly ash 3.4 ± 1.6 4.4 ± 0.7 9.5 ± 1.1 12.7 ± 2.4 4.1 ± 0.6
-------------------------------------------------------------------
a From: Stoewsand et al. (1978).
The possibility has been raised that some selenium may be added
to the food supply as a result of atmospheric contaminants from
the burning of fossil fuel or industrial emissions settling out on
the leaves of plants or on the surface of the soil. The
geographical pattern of selenium levels in rain-water from Denmark
or the USA implicates airborne selenium from the industrial and
domestic uses of fossil fuel as sources (Kubota et al., 1975).
Total industrial emissions of selenium in the USA for the year 1970
were estimated to be about 1.1 million kg, 62% of which was derived
from the burning of coal (US NAS/NRC, 1976). This level of
industrial emission of selenium may be compared with a projected
annual use of selenium as a feed additive for chickens and swine in
the "low-selenium" areas of the USA of 6000 kg (US FDA, 1974). But
there is no evidence that the selenium deposited in rainwater has
any influence on the concentration of selenium in forage plants
grown in Denmark or in the heavily industrialized northeastern USA
(Kubota et al., 1975). Furthermore, selenium deficiency has been
observed in farm livestock grazing near coal-burning electric power
stations (Anonymous, 1975). This suggests that airborne selenium
exists in the form of inert elemental selenium, insoluble selenide
salts, or as selenium dioxide, which would be tightly bound by
acidic, iron-bearing soils. In any case, the selenium would not be
available for uptake by plants; thus, there would appear to be
little prospect of adding appreciable selenium to the food supply
via atmospheric contamination.
(b) Human activities that decrease selenium levels in the food
chain
The metabolic antagonism between sulfur and selenium (Levander,
1976a) led early workers to suggest that borderline selenium
deficiency might be exacerbated by sulfate fertilization (Schubert
et al., 1961). More recent research has demonstrated that sulfate
fertilization can result in decreased selenium concentrations in
forage plants (Pratley & McFarlane, 1974; Westermann & Robbins,
1974). In many cases, this decrease can be largely explained by a
dilution effect caused by a growth response of the plant to the
sulfate fertilization, but decreases in selenium concentration
independent of crop yield have also been reported. Such decreases
could be the result of competitive interference by sulfate with the
uptake of selenate by plants. The use of phosphate fertilizers in
parts of Australia and New Zealand has resulted in an apparent
increase in the incidence of selenium deficiency in livestock
(Judson & Obst, 1975), but others have reported increases in the
selenium content of forages after fertilization with phosphorus
(Robbins & Carter, 1970; Carter et al., 1972). These conflicting
results may be due to differences in the selenium content of the
phosphatic rock from which the fertilizers were prepared.
Another possible way in which man's activities might decrease
selenium in the food chain or render it less available, is through
heavy metal pollution. For example, silver has been demonstrated
to antagonize selenium in a wide variety of situations (Diplock,
1976). In an area where there has been a large amount of silver
mining activity, muscular dystrophy of the type usually associated
with selenium deficiency has been reported in calves fed a ration
based on milk powder, even though the selenium content of the milk
seemed to be adequate. Apparently, the milk powder was derived
from cows that grazed land containing high levels of silver
residues from the old mining industry. Further research is needed
to establish fully the role of silver in potentiating these field
cases of apparent selenium deficiency. Because of the many
interactions of selenium with other environmental pollutants such
as mercury, cadmium, or arsenic (section 7), it is possible that
other cases of heavy metal-induced selenium deficiency will be
discovered.
The influence of cooking on the selenium content of foods was
determined because of the well-known instability and volatility of
many selenium compounds. Certain vegetables that normally contain
high levels of selenium such as asparagus or mushrooms, lost up to
40% of their selenium content as a result of boiling (Higgs et al.,
1972). Majstruk & Suchkov (1978) reported that up to 50% of the
original selenium was lost from vegetables and dairy products
during cooking; the addition of salt and acid pH or extended
cooking, particularly promoted such losses. But, other typical
cooking procedures such as boiling cereals, baking poultry or fish,
or broiling meats had little effect on selenium levels in the food
(Higgs et al., 1972). Similar results were observed by others for
baking fish or broiling meats, though some decrease in selenium
concentration was caused by frying organ meats or fish (Ganapathy
et al., 1975). Thompson et al. (1975) determined the selenium
contents of cooked food composites and found a rough agreement with
the calculated values based on unprepared foods from which they
were derived. On the basis of all these studies, it can be
concluded that usual cooking procedures cause little decrease in
the selenium content of most foods.
4.1.2. Chemical forms of selenium in food
There is little information on the chemical forms of selenium
that occur in human food (Levander, 1986). Cappon & Smith (1982)
reported that canned tuna fish contained variable percentages of
hexavalent selenium, ranging from 7.6 to 44.8%, which was
independent of the total selenium content. The other forms of
selenium in the tuna were di- and quadrivalent selenium. On an
average percentage basis, hexavalent selenium was more easily
extractable by water than di- or quadrivalent selenium. In
recently canned samples, an average of 55.6% of the total selenium
content was water-extractable. However, for the older samples, the
corresponding average extractable level was 48.2%. The results
suggest that sample storage may influence the chemical form of
selenium in canned tuna. Olson et al. (1970) found that, in a
certain portion of high selenium wheat (the pronase hydrolysate of
gluten), about half of the selenium was in the form of
selenomethionine. About 15% of the selenium in water extracts of
seleniferous cabbage leaves was in the form of selenomethionine,
but appreciable amounts of other selenium compounds were also
present (Hamilton, 1975). The chemical forms of selenium in food
are likely to affect the bioavailability of selenium (section 7.2.2
and 8.2.3).
4.2. Drinking-Water
Analysis of samples taken from various public water-supply
systems in the USA showed that less than 0.5% contained selenium
levels that exceeded the US PHS limit of 10 µg/litre (Taylor, 1963;
Lakin & Davidson, 1967; McCabe et al., 1970). Samples from 1280
central water sources providing water for 6.5 million Bulgarians
contained less than 2 µg selenium/litre (Gitsova, 1973). Tap water
from Stockholm, Sweden contained only 0.06 µg/litre (Lindberg,
1968) and tap and mineral waters from Stuttgart, Federal Republic
of Germany contained 1.6 and 5.3 µg/litre, respectively
(Oelschlager & Menke, 1969). Surface water sources in different
subregions of the Cernovici region of the Ukrainian SSR contained
selenium levels ranging from 0.09 ± 0.01 to 3.00 ± 0.40 µg/litre;
ground-water levels ranged from 0.07 ± 0.0 to 4.00 ± 0.85 µg/litre
(Suchkov & Kacap, 1971). A few tenths to several µg per litre were
reported in non-seleniferous regions of the USSR, the maximum
reported value being 5.1 µg/litre. In Argentina, the selenium
content of 22 surface waters varied from < 2 to 19 µg/litre with a
median value of 3 µg/litre (WHO, 1984).
Elevated levels of selenium (> 100 µg/litre) can be found in
seeps, springs, and shallow wells, though waters from deep wells
contain only a few µg/litre (US NAS/NRC, 1976). Highly variable
amounts of selenium were reported in wells from seleniferous areas
in the USA ranging from non-detectable levels to 330 µg/litre
(Smith & Westfall, 1937). Some well waters contained enough
selenium to be considered as poisonous for man or livestock and
loss of hair and nails in children was attributed to selenosis, but
the evidence for this was not considered convincing (US NAS/NRC,
1976).
The average selenium content of 11 samples of drinking-water
taken from a high-selenium area in China with a history of disease
reported as chronic selenosis was 54 µg/litre (Yang et al., 1983).
Four of these samples were surface water from a village with
previous heavy intoxication and these averaged 139 (117 - 159)
µg/litre. Such water would contribute substantial amounts of
selenium, but would still comprise a small fraction of the total
selenium intake in this area, since the contribution from food was
estimated to be about 5000 µg/day (section 5.1.1.1). The other 7
samples of drinking-water originated from several sources and
averaged only 5 µg/litre.
Rosenfeld & Beath (1964) concluded that selenium does not occur
in water in sufficient amounts to produce selenium toxicity in man
or animals, except in isolated cases. This was confirmed and
expanded in the statement by US NAS/NRC (1976, 1980) that waters
are rarely a significant source of selenium from either a
nutritional or a toxicity point of view. The low concentrations of
selenium in drinking-water are probably the result of several
mechanisms that act to decrease the selenium content of waters
(section 2.1).
4.3. Air
Selenium levels in the air breathed by the general population
are probably well below 10 ng/m3, on average (US NAS/NRC, 1976).
Hashimoto & Winchester (1967) found 0.3 - 1.6 ng airborne
selenium/m3 in an urbanized area. Levels of 3.6 - 9.7 ng
selenium/m3 were detected by Pillay et al. (1971), who showed that
over half of the selenium was not retained on filters able to trap
particulates greater than 0.1 µg in diameter. Zoller & Reamer
(1976) concluded that atmospheric levels of selenium in most urban
regions vary from 0.1 to 10 ng/m3. Selenium levels in air around
industries that use selenium can be higher (section 3.2.2), perhaps
of the order of a few µg/m3. Work-place air in selenium industries
apparently contained mg levels of selenium/m3 in the past, but more
recent measurements have indicated lower levels (section 5.2).
5. HUMAN EXPOSURE
5.1. Estimate of General Population Exposure
As discussed in section 4, selenium levels in air are low in
areas without selenium-emitting industries and are not likely to
exceed 10 ng/m3. Assuming that a person respires 20 m3 air daily,
the contribution to daily selenium intake via this route would be
0.2 µg or less. Since this intake is much lower than that through
food (see below), airborne selenium makes a negligible contribution
to the average daily selenium intake of the general population.
However, levels of selenium may be somewhat higher in the
atmosphere around some industrial plants using selenium (section
3.2.2). If it is assumed that 3 µg selenium/m3 is found in the air
near these industries, then exposure to selenium via air would be
60 µg/day.
Other possible exposures of the general population to selenium
via the air include contributions from house dust and tobacco
smoke. Lakin & Davidson (1967) reported that house or office dust
could contain up to 10 mg selenium/kg. Although no data are
available to provide an estimate of human exposure to selenium from
this source, the possibility of such exposure should be considered
because of the increased interest in the quality of indoor air.
Olsen & Fros