WHO/EOS/98.1
Guidelines for d g-vvater quality
SECOND EDITION
Addendum to Volume 2 Health criteria and other supporting.information
World Health Organization Geneva Guidelines for drinking-water quality
SECOND EDITION WHO/EOS/98.1 Distribution: General English Only
Guidelines for drinking-water quality
SECOND EDITION
Addendum to Volume 2
Health criteria and other supporting information
World Health Organization Geneva 1998 ©World Health Organization 1998
Reprinted 2002
This document is not a formal publication of the World Health Organization and all rights are reserved by the Organization. The document may, however, be freely reviewed, abstracted, or reproduced or translated in part, but not for sale or for use in conjunction with commercial purposes.
For authorization to reproduce or translate the work in full, and for any use by commercial entities, applications and enquiries should be addressed to the Division of Operational Support in Environmental Health, 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 the reprints, adaptations and translations already available.
The designations employed and the presentation of the material in this document 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. iii
Contents
Preface ...... v
Acknowledgements ...... vii
Acronyms and abbreviations used in the text...... viii
Introduction ...... 1
Inorganic constituents ...... 3 Aluminium ...... 3 Boron ...... I5 Copper ...... 3I Nickel ...... 47 Nitrate and nitrite ...... 63 Uranium ...... 8I
Organic constituents ...... 95 Cyanobacterial toxins: Microcystin-LR ...... 95 Edetic acid (EDTA) ...... III Polynuclear aromatic hydrocarbons ...... I23
Pesticides ...... I 53 Bentazone ...... I 53 Carbofuran ...... I57 Cyanazine ...... I65 I ,2-Dibromoethane ...... I77 2,4-Dichlorophenoxyacetic acid (2,4-D) ...... I9I I ,2-Dichloropropane (I ,2-DCP) ...... 20 I Diquat ...... 209 Glyphosate ...... 2I9 Pentachlorophenol ...... 229 Terbuthylazine (IBA) ...... 245
Disinfectant by-product ...... 255 Chloroform ...... 255
Annexes ...... 277 Annex I. List of participants in preparatory meetings ...... 277 Annex 2. Tables of guideline values ...... 28I
V
Preface
Between I993 and I997, the World Health Organization (WHO) published the second edition of Guidelines for drinking-water quality in three volumes: Volume I, Recommendations, in I993, Volume 2, Health criteria and other supporting information, in I996, and Volume 3, Surveillance and control of community supplies, in I997. As with the first edition, the development of these guidelines was organized and carried out jointly by WHO headquarters and the WHO Regional Office for Europe. At the Final Task Group Meeting (Geneva, Switzerland, 2I-25 September I992), when the second edition of the Guidelines was approved, it was agreed to establish a continuing process of updating of the guidelines, with a number of chemical substances and microbiological agents subject to periodic evaluation. Addenda containing these evaluations will be issued as necessary until the third edition of the Guidelines is published, approximately I 0 years after the second edition. In I995, a Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality agreed on the framework for the updating process and established three working groups to support the development of addenda and monographs on chemical aspects, microbiological aspects, and protection and control of water quality. The Committee selected the chemical substances to be evaluated in the first addendum, designated coordinators for each major group of chemicals, and identified lead institutions for the preparation of health criteria documents evaluating the risks for human health from exposure to the particular chemicals in drinking-water. Institutions from Canada, Finland, France, Germany, the Netherlands, Sweden, the United Kingdom, and the USA, as well as the ILO!UNEP/WHO International Programme on Chemical Safety (IPCS), prepared the requested health criteria documents. Under the responsibility of the designated coordinators for each chemical group, the draft health criteria documents were submitted to a number of scientific institutions and selected experts for peer review. Comments were taken into consideration by the coordinators and authors before the documents were submitted for final evaluation by the 1997 Working Group Meeting on Chemical Substances in Drinking-Water. The Working Group reviewed the health risk assessments and, where appropriate, decided upon guideline values. During the preparation of draft health criteria documents and at the I997 Working Group Meeting, careful consideration was always given to previous risk assessments carried out by IPCS in its Environmental Health Criteria monographs, the International Agency for Research on Cancer, the Joint FAO/WHO Meetings on Pesticide Residues, and the Joint FAO/WHO Expert Committee on Food Additives, which evaluates contaminants such as nitrate and nitrite in addition to food additives. Evaluations of chemical substances given in this Addendum supersede evaluations previously published in Volume 1 and Volume 2 of the Guidelines.
vii
Acknowledgements
The work of the following coordinators was crucial in the development of this first addendum on chemical substances in drinking-water:
P. Chambon, Health Environment Hygiene Laboratory of Lyon, Lyon, France (inorganic constituents) U. Lund, Water Quality Institute, Horsholm, Denmark (organic constituents) H. Galal-Gorchev, Urban Environmental Health, World Health Organization, Geneva, Switzerland (pesticides) E. Ohanian, Environmental Protection Agency, Washington, DC, USA (disinfectants and disinfectant by-products)
The coordinators for the overall administrative and technical aspects of this addendum were, respectively, J. Kenny and H. Galal-Gorchev, Urban Environmental Health, WHO, Geneva, Switzerland. Ms Marla Sheffer of Ottawa, Canada, was responsible for the scientific editing of the addendum. Special thanks are due to the authors and their institutions for the preparation of draft health criteria documents. The following institutions prepared such health criteria documents: Health Canada; Water Research Centre, England; National Public Health Institute, Finland; Health Environment Hygiene Laboratory of Lyon, France; Fraunhofer Institute for Toxicology and Aerosol Research, Germany; National Institute of Public Health and Environmental Protection, Netherlands; National Food Administration, Sweden; International Programme on Chemical Safety, Switzerland; and the Environmental Protection Agency, USA. The preparation of this addendum was made possible by the financial support afforded to WHO by the following countries: Canada, Japan, and the USA. A financial contribution for the convening of the Working Group Meeting on Chemical Substances in Drinking-Water and for printing the addendum was received from the European Commission and is gratefully acknowledged. The preparation of the first addendum to the Guidelines for drinking-water quality involved the participation of numerous institutions and experts. The work of these institutions and scientists, whose names appear in Annex I, was central to the completion of this addendum and is much appreciated. viii
Acronyms and abbreviations used in the text 1
ADI acceptable daily intake ATPase adenosine triphosphatase CAS Chemical Abstracts Service Cl confidence interval DNA deoxyribonucleic acid EEC European Economic Community EPA Environmental Protection Agency (USA) FAO Food and Agriculture Organization of the United Nations IARC International Agency for Research on Cancer ILO International Labour Organisation IPCS International Programme on Chemical Safety IUPAC International Union of Pure and Applied Chemistry JECFA Joint FAO/WHO Expert Committee on Food Additives JMP Joint Meeting on Pesticides JMPR Joint FAO/WHO Meeting on Pesticide Residues
LCso lethal concentration, median LDso lethal dose, median LOA EL lowest-observed-adverse-effect level LOEC lowest-observed-effect concentration MTD maximum tolerated dose NCI National Cancer Institute (USA) NOAEL no-observed-adverse-effect level NOEL no-observed-effect level NTP National Toxicology Program (USA) PMTDI provisional maximum tolerable daily intake RNA ribonucleic acid TDI tolerable daily intake UNEP United Nations Environment Programme USA Umted States of America USP US Pharmacopoeia WHO World Health Organization w/v weight/volume
These are the acronyms and abbreviations that are used without definition in the text. Introduction
Chemical substances evaluated in this addendum were selected by the 1995 Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality for one or more of the following reasons:
• adequate data were not available to allow a guideline value to be derived, or only a provisional guideline value could be derived, in the second edition of the Guidelines; • the substance was recommended for evaluation by the Task Group convened to finalize the second edition of the Guidelines; • new health risk assessments were available from IPCS through its Environmental Health Criteria monographs, JMPR, or JECF A; • a new evaluation of the carcinogenic risk of the chemical' was available from IARC; • requests to evaluate the chemical were made to the WHO Secretariat.
Concepts of guideline value and provisional guideline value, assumptions made, and scientific principles for the assessment of risk to human health from exposure to chemicals in drinking-water used in this addendum are described in Volume 1, Recommendations, of the second edition of the Guidelines. Only a brief summary of the approaches used to derive the guideline values is given here. In developing the guideline values for potentially hazardous chemicals, a daily consumption of 2 litres of drinking-water by a person weighing 60 kg was generally assumed. Where it was judged that infants and children were at a particularly high risk from exposure to certain chemicals, the guideline values were derived on the basis of a 5-kg infant consuming 0.75 litre per day or a 10-kg child consuming I litre per day. For compounds showing a threshold for toxic effects, a tolerable daily intake (TDI) approach was used to derive the guideline value. A portion of the TDI was allocated to drinking-water, based on potential exposure from other sources, such as food and air. Where information on other sources of exposure was not available, an arbitrary (default) value of 10% of the TDI was allocated to drinking-water. For compounds considered to be genotoxic carcinogens, guideline values were determined using a mathematical model. The guideline values presented in the addendum to Volume 1 are the concentrations in drinking-water associated with an estimated excess lifetime cancer risk of 1o-s (one additional cancer case per I 00 000 of the population ingesting drinking-water containing the substance at the guideline value for 70 years). In this addendum to Volume 2, concentrations associated with excess lifetime cancer risks of 10-4, 1o-s, and 1o- 6 are presented to emphasize the fact that each country should select its own appropriate risk levels. It is emphasized that the guideline values recommended are not mandatory limits. Such limits should be set by national or regional authorities, using a risk benefit approach and taking into consideration local environmental, social, economic, and cultural conditions.
ALUMINIUM
First draft 1 prepared by H. Ga/al-Gorchev International Programme on Chemical Safety World Health Organization, Geneva, Switzerland
No health-based guideline value for aluminium was recommended in the second edition of the WHO Guidelines for drinking-water quality. It was concluded that although further studies were needed, the balance of epidemiological and physiological evidence did not support a causal role for aluminium in Alzheimer disease. An aluminium concentration of 0.2 mg/litre in drinking-water provided a compromise between the practical use of aluminium salts in water treatment and discoloration of distributed water. The Coordinating Committee for the updating of the WHO Guidelines recommended that a health criteria document be prepared for aluminium, based on the IPCS Environmental Health Criteria monograph that was finalized in 1995.
1. GENERAL DESCRIPTION l.I Identity
Aluminium is the most abundant metallic element and constitutes about 8% of the Earth's crust. It occurs naturally in the environment as silicates, oxides, and hydroxides, combined with other elements, such as sodium and fluoride, and as complexes with organic matter.
Compound CAS no. Molecular formula Aluminium 7429-90-5 AI Aluminium chloride 7446-70-0 AICI3 Aluminium hydroxide 21645-51-2 AI(OH)3 Aluminium oxide 1344-28-1 AIP3 Aluminium sulfate I 0043-01-3 Al2(S04)3
1.2 Physicochemical properties (Lide, 1993)
Property AI AIC/3 Al(OH)J A/203 Al2(S04h Melting point (°C) 660 190 300 2072 770 (d) Boiling point (°C) 2467 262 (d) 2980 3 Density at 20°C (g/cm ) 2.70 2.44 2.42 3.97 2.71 Water solubility (g/litre) (i) 69.9 (I) (i) 31.3 at 0°C
d, decomposes; i, insoluble
This document is essentially a synopsis of the evaluation of aluminium contained in the IPCS Environmental Health Criteria monograph for aluminium (WHO, 1997). 4 INORGANIC CONSTITUENTS
1.3 Organoleptic properties
Use of aluminium salts as coagulants in water treatment may lead to increased concentrations of alumip.ium in finished water. Where residual concentrations are high, aluminium may be deposited in the distribution system. Disturbance of the deposits by change in flow rate may increase aluminium levels at the tap and lead to undesirable colour and turbidity (WHO, 1996). Concentrations of aluminium at which such problems may occur are highly dependent on a number of water quality parameters and operational factors at the water treatment plant.
1.4 Major uses
Aluminium metal is used as a structural material in the construction, automotive, and aircraft industries, in the production of metal alloys, in the electric industry, in cooking utensils, and in food packaging. Aluminium compounds are used as antacids, antiperspirants, and food additives (ATSDR, 1992). Aluminium salts are also widely used in water treatment as coagulants to reduce organic matter, colour, turbidity, and microorganism levels. The process usually consists of addition of an aluminium salt (often sulfate) at optimum pH and dosage, followed by flocculation, sedimentation, and filtration (Health Canada, 1993 ).
1.5 Environmental/ate
Aluminium is released to the environment mainly by natural processes. Several factors influence aluminium mobility and subsequent transport within the environment. These include chemical speciation, hydrological flow paths, soil-water interactions, and the composition of the underlying geological materials. Acid environments caused by acid mine drainage or acid rain can cause an increase in the dissolved aluminium content of the surrounding waters (ATSDR, 1992; WHO, 1997). Aluminium can occur in a number of different forms in water. It can form monomeric and polymeric hydroxy species, colloidal polymeric solutions and gels, and precipitates, all based on aquated positive ions or hydroxylated aluminates. In addition, it can form complexes with various organic compounds (e.g. humic or fulvic acids) and inorganic ligands (e.g. fluoride, chloride, and sulfate), most but not all of which are soluble. The chemistry of aluminium in water is complex, and many chemical parameters, including pH, determine which aluminium species are present in aqueous solutions. In pure water, aluminium has a minimum solubility in the pH range 5.5-6.0; concentrations of total dissolved aluminium increase at higher and lower pH values (CCME, 1988; ISO, 1994).
2. ANALYTICAL METHODS
Aluminium is reacted with pyrocatechol violet followed by spectrometric measurement of the resulting coloured complex. The method is restricted to the determination of the aquated cations and other forms of aluminium readily converted to that cationic form by acidification. The limit of detection is 2 J.lg/litre (ISO, 1994). ALUMINIUM 5
The limit of detection for the determination of aluminium by inductively coupled plasma atomic emission spectroscopy ranges from 40 to 100 f.tg/litre (ISO, 1996). Flame and graphite furnace atomic absorption spectrometric (AAS) methods are applicable for the determination of aluminium in water at concentrations of 5-100 mg/litre and 0.01-0.1 mg/litre, respectively. The working range of the graphite furnace AAS method can be shifted to higher concentrations either by dilution of the sample or by using a smaller sample volume (ISO, 1997).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Aluminium enters the atmosphere as a major constituent of atmospheric particulates originating from natural soil erosion, mining or agricultural activities, volcanic eruptions. or coal combustion. Atmospheric aluminium concentrations show widespread temporal and spatial variations. Airborne aluminium levels range from 0.0005 flg/m3 over Antarctica to more than 1 'flg/m3 in industrialized areas (WHO. 1997).
3.2 Water
The concentration of aluminium in natural waters can vary significantly depending on various physicochemical and mineralogical factors. Dissolved aluminium concentrations in waters with near-neutral pH values usually range from 0.001 to 0.05 mg/litre but rise to 0.5-1 mg/litre in more acidic waters or water rich in organic matter. At the extreme acidity of waters affected by acid mine drainage, dissolved aluminium concentrations of up to 90 mg/litre have been measured (WHO, 1997). Aluminium levels in drinking-water vary according to the levels found in the source water and whether aluminium coagulants are used during water treatment. In Germany, levels of aluminium in public water supplies averaged 0.01 mg/litre in the western region. whereas levels in 2.7% of public supplies in the eastern region exceeded 0.2 mg/litre (Wilhelm & Idel, 1995). In a 1993-1994 survey of public water supplies in Ontario, Canada, 75% of all average levels were less than 0.1 mg/litre, with a range of 0.04-0.85 mg/litre (OMEE, 1995). In a large monitoring programme in 1991 in the United Kingdom, concentrations in 553 samples (0.7%) exceeded 0.2 mg/litre (MAFF, 1993). In a survey of 186 community water supplies in the USA, median aluminium concentrations for all finished drinking-water samples ranged from 0.03 to 0.1 mg/litre; for facilities using aluminium sulfate coagulation. the median level was 0.1 mg/litre, with a maximum of 2. 7 mg/litre (Miller et al., 1984). In another US survey. the average aluminium concentration in treated water at facilities using aluminium sulfate coagulation ranged from 0.01 to 1.3 mg/litre, with an overall average of 0.16 mg/litre (Letterman & Driscoll, 1988; ATSDR, 1992). 6 INORGANIC CONSTITUENTS
3.3 Food
Aluminium is present in foods naturally or from the use of aluminium containing food additives. The use of aluminium cookware, utensils, and wrappings can increase the amount of aluminium in food; however, the magnitude of this increase is generally not of practical importance. Foods naturally high in aluminium include potatoes, spinach, and tea. Processed dairy products, flour, and infant formula may be high in aluminium if they contain aluminium-based food additives (FAO/WHO, 1989; Pennington & Schoen, 1995; WHO, 1997). Adult dietary intakes of aluminium (mg/day) have been reported in several countries: Australia (1.9-2.4), Finland (6.7), Germany (8-11), Japan (4.5), Netherlands (3.1), Sweden (13), Switzerland (4.4), United Kingdom (3.9), and USA (7.1-8.2). Intake of children 5-8 years old was 0.8 mg/day in Germany !and 6.5 mg/day in the USA. Infant intakes of aluminium in Canada, the United Kingdom, and the USA ranged from 0.03 to 0.7 mg/day (WHO, 1997).
3.4 Estimated total exposure and relative contribution of drinking-water
Aluminium intake from foods, particularly those containing aluminium compounds used as food additives, represents the· major route of aluminium exposure for the general public, excluding persons who regularly ingest aluminium containing antacids and buffered analgesics, for whom intakes may be as high as 5 g/day (WHO, 1997). At an average adult intake of aluminium from food of 5 mg/day and a drinking-water aluminium concentration of 0.1 mg/litre, the contribution of drinking-water to the total oral exposure to aluminium will be about 4%. The contribution of air to the total exposure is generally negligible.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
In experimental animals, absorption of aluminium via the gastrointestinal tract is usually less than 1%. The main factors influencing absorption are solubility, pH, and chemical species. Organic complexing compounds, notably citrate, increase absorption. Aluminium absorption may interact with calcium and iron transport systems. Aluminium, once absorbed, is distributed in most organs within the body, with accumulation occurring mainly in bone at high dose levels. To a limited but as yet undetermined extent, aluminium passes the blood-brain barrier and is also distributed to the fetus. Aluminium is eliminated effectively in the urine (WHO, 1997). In humans, aluminium and its compounds appear to be poorly absorbed, although the rate and extent of absorption have not been adequately studied. The mechanism of gastrointestinal absorption has not yet been fully elucidated. Variability results from the chemical properties of the element and the formation of various chemical species, which is dependent upon the pH, ionic strength, presence of competing elements (e.g. silicon), and presence of complexing agents within the ALUMINIUM 7 gastrointestinal tract (e.g. citrate). The urine is the most important route of aluminium excretion (WHO, 1996, 1997).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
The oral LD50 of aluminium nitrate, chloride, and sulfate in mice and rats ranges from 200 to 1000 mg of aluminium per kg of body weight (WHO, 1997).
5.2 Short-term exposure
Groups of 25 male Sprague-Dawley rats were fed diets contammg basic sodium aluminium phosphate or aluminium hydroxide at 0, 5, 67, 141, or 288/302 mg of aluminium per kg of body weight per day for 28 days. No treatment-related effects on organ and body weights, haematology, clinical chemistry parameters, and histopathology were observed, and there was no evidence of deposition of aluminium in bones. The NOELs were 288 and 302 mg of aluminium per kg of body weight per day for sodium aluminium phosphate and aluminium hydroxide, respectively (Hicks et al., 1987). In a study in which a wide range of end-points was examined, groups of 10 female Sprague-Dawley rats received drinking-water containing aluminium nitrate for 28 days at 0, 1, 26, 52, or 104 mg of aluminium per kg of body weight per day. The only effects noted were mild histopathological changes in the spleen and liver of the high-dose group. Although tissue aluminium concentrations were generally higher in treated animals, the increases were significant only for spleen, heart, and gastrointestinal tract of the high-dose group. The NOAEL was 52 mg of aluminium per kg of body weight per day (Gomez et al., 1986). Groups of 10 female Sprague-Dawley rats received aluminium nitrate in their drinking-water at doses of 0, 26, 52, or 260 mg of aluminium per kg of body weight per day for l 00 days. Organ and body weights, histopathology of the brain, heart, lungs, kidney, liver, and spleen, haematology, and plasma chemistry were examined. The only effect observed was a significant decrease in body-weight gain associated with a decrease in food consumption at 260 mg of aluminium per kg of body weight per day. Aluminium did not accumulate in a dose-dependent manner in the organs and tissues examined. The NOAEL in this study was 52 mg of aluminium per kg of body weight per day (Domingo et al., 1987a). Sodium aluminium phosphate, a leavening acid, was administered to groups of six male and six female beagle dogs at dietary concentrations ofO, 0.3, 1.0, or 3.0% for 6 months. Statistically significant decreases in food consumption occurred sporadically in all treated groups of female dogs, but there was no associated decrease in body weight. No significant absolute or relative organ-weight differences were found between any of the treated groups and controls. Haematological, blood chemistry, and urinalysis data showed no toxicologically significant trend. The NOAEL was the highest dose tested, approximately 70 mg of aluminium per kg of body weight per day (Katz et al., 1984). 8 INORGANIC CONSTITUENTS
Beagle dogs (four per sex per dose) were fed diets containing basic sodium aluminium phosphate at 0, 10, '22-27, or 75-80 mg of aluminium per kg of body weight per day for 26 weeks. The only treatment-related effect was a sharp, transient decrease in food consumption and concomitant decrease in body weight in high-dose males. The LOAEL was 75-80 mg/kg of body weight per day (Pettersen et al., 1990).
5.3 Long-term exposure
No adverse effects on body weight or longevity were observed in Charles River mice (54 males and 54 females per group) receiving 0 or 5 mg of aluminium (as potassium aluminium sulfate) per kg of diet during their lifetime (Schroeder & Mitchener, 1975a; FAO/WHO, 1989). Two groups of Long-Evans rats (52 of each sex) received 0 or 5 mg of aluminium (as potassium aluminium sulfate) per litre of drinking-water during their lifetime. No effects were found on body weight; average heart weight; glucose, cholesteroL and uric acid levels in serum; and protein and glucose content and pH ofurine. The life span was not affected (Schroeder & Mitchener, 1975b; FAO/WHO, 1989).
5.4 Reproductil'e and developmental toxicity
Aluminium nitrate was administered by gavage to groups of pregnant Sprague-Dawley rats on day 14 of gestation through day 21 of lactation at doses of 0, 13, 26, or 52 mg of aluminium per kg of body weight per day. These doses did not produce overt fetotoxicity, but growth of offspring was significantly delayed (body weight, body length, and tail length) from birth to weaning in aluminium-treated groups (Domingo et al., 1987b). Aluminium nitrate was administered by intubation to male Sprague-Dawley rats at 0, 13, 26, or 52 mg of aluminium per kg of body weight per day for 60 days prior to mating and to virgin females for 14 days prior to mating, with treatment continuing throughout mating, gestation, parturition, and weaning of the litters. No reproductive effects on fertility (number of litters produced), litter size, or intrauterine or postnatal offspring mortality were reported. There was a decrease in the numbers of corpus lutea in the highest dose group. However, a dose-dependent delay in the growth of the pups was observed in all treatment groups; female offspring were affected at 13 mg of aluminium per kg of body weight per day and males at 26 and 52 mg of aluminium per kg of body weight per day. Because of the design of this study, it is not clear whether the postnatal growth effects in offspring represented general toxicity to male or female parents or specific effects on reproduction or development. However, the reported LOAEL in females in this study was 13 mg of aluminium per kg of body weight per day (Domingo et al., 1987c). The developmental toxicity of aluminium by the oral route is highly dependent on the form of aluminium and the presence of organic chelators that influence bioavailability. Aluminium hydroxide did not produce either maternal or developmental toxicity when it was administered by gavage during embryogenesis to mice at doses up to 92 mg of aluminium per kg of body weight per day (Domingo ALUMINIUM 9 et al., 1989) or to rats at doses up to 265 mg of aluminium per kg of body weight per day (Gomez et al., 1990). When aluminium hydroxide at a dose of 104 mg of aluminium per kg of body weight per day was administered with ascorbic acid to mice, no maternal or developmental toxicity was seen, in spite of elevated maternal placenta and kidney concentrations of aluminium (Colomina et al., 1994); on the other hand, aluminium hydroxide at a dose of 133 mg of aluminium per kg of body weight per day administered with citric acid produced maternal and fetal toxicity in rats (Gomez et al., 1991). Aluminium hydroxide (57 mg of aluminium per kg of body weight) given with lactic acid (570 mg/kg of body weight) to mice by gavage was not toxic, but aluminium lactate (57 mg of aluminium per kg of body weight) produced developmental toxicity, including poor ossification, skeletal variations, and cleft palate (Colomina et al., 1992).
5.5 Mutagenicity and related end-points
Aluminium can form complexes with DNA and cross-link chromosomal proteins and DNA, but it has not been shown to be mutagenic in bacteria or induce mutation or transformation in mammalian cells in vitro. Chromosomal aberrations have been observed in bone marrow cells of exposed mice and rats (WHO, 1997).
5.6 Carcinogenicity
There is no indication that aluminium is carcinogenic. JECF A evaluated the limited studies of Schroeder and Mitchener (l975a,b; section 5.3) and concluded that there was no evidence of an increase in tumour incidence related to the administration of potassium aluminium sulfate in mice or rats (FAO/WHO, 1989).
5. 7 Neurotoxicity
Behavioural impairment has been reported in laboratory animals exposed to soluble aluminium salts (e.g. lactate, chloride) in the diet or drinking-water in the absence of overt encephalopathy or neurohistopathology. Both rats (Commissaris et al., 1982; Thorne et al., 1987; Connor et al., 1988) and mice (Yen-Koo, 1992) have demonstrated such impairments at doses exceeding 200 mg of aluminium per kg of body weight per day. Although significant alterations in acquisition and retention of learned behaviour were documented, the possible role of organ damage (kidney, liver, immunological) due to alun1iniun1 was incompletely evaluated in these studies (WHO, 1997). In studies on brain development in mice and rats, grip strength was impaired in offspring of dams fed lOO mg of aluminium (as aluminium lactate) per kg of body weight per day in the diet, in the absence of maternal toxicity (WHO, 1997).
6. EFFECTS ON HUMANS
There is little indication that aluminium is acutely toxic by oral exposure despite its widespread occurrence in foods, drinking-water, and many antacid preparations (WHO. 1997). 10 INORGANIC CONSTITUENTS
In 1988, a population of about 20 000 individuals in Camelford, England, was exposed for at least 5 days to unknown but increased levels of aluminium accidentally distributed to the population from a water supply facility using aluminium sulfate for treatment. Symptoms including nausea, vomiting, diarrhoea, mouth ulcers, skin ulcers, skin rashes, and arthritic pain were noted. It was concluded that the symptoms were mostly mild and short-lived. No lasting effects on health could be attributed to the known exposures from aluminium in the drinking water (Clayton, 1989). It has been hypothesized that aluminium exposure is a risk factor for the development or acceleration of onset of Alzheimer disease (AD) in humans. WHO (1997) has evaluated some 20 epidemiological studies that have been carried out to test the hypothesis that aluminium in drinking-water is a risk factor for AD. Study designs ranged from ecological to case-control. Six studies on populations in Norway (Flaten, 1990), Canada (Neri & Hewitt, 1991), France (Michel et al., 1991; Jacqmin et al., 1994), Switzerland (Wettstein et al., 1991), and England (Martyn et al., 1989) were considered of sufficiently high quality to meet the general criteria for exposure and outcome assessment and the adjustment for at least some confounding variables. Of the six studies that examined the relationship between aluminium in drinking-water and dementia or AD, three found a positive relationship, but three did not. However, each of the studies had some deficiencies in the study design (e.g. ecological exposure assessment; failure to consider aluminium exposure from all sources and to control for important confounders, such as education, socioeconomic status, and family history; the use of surrogate outcome measures for AD; and selection bias). In general, the relative risks determined were less than 2, with large confidence intervals, when the total aluminium concentration in drinking-water was 0.1 mg/litre or higher. Based on current knowledge of the pathogenesis of AD and the totality of evidence from these epidemiological studies, it was concluded that the present epidemiological evidence does not support a causal association between AD and aluminium in drinking-water (WHO, 1997). In addition to the epidemiological studies that examined the relationship between AD and aluminium in drinking-water, two studies examined cognitive dysfunction in elderly populations in relation to the levels of aluminium in drinking water. The results were again conflicting. One study of 800 male octogenarians consuming drinking-water with aluminium concentrations up to 98 f.lg/litre found no relationship (Wettstein et al., 1991). The second study used "'any evidence of mental impairment" as an outcome measure and found a relative risk of I. 72 at aluminium drinking-water concentrations above 85 f.lg/litre in 250 males (Forbes et al., 1994). Such data are insufficient to show that aluminium is a cause of cognitive impairment in the elderly.
7. CONCLUSIONS
The Environmental Health Criteria document for aluminium (WHO, 1997) concluded that: ALUMINIUM 11
On the whole, the positive relationship between alumimum in drinking water and AD, which was demonstrated in several epidemiological studies, cannot be totally dismissed. However, strong reservations about inferring a causal relationship are warranted in view of the failure of these studies to account for demonstrated confounding factors and for total aluminium intake from all sources.
Taken together, the relative risks for AD from exposure to aluminium in drinking-water above I 00 mg/litre, as determined in these studies, are low (less than 2.0). But, because the risk estimates are imprecise for a variety of methodological reasons, a population-attributable risk cannot be calculated with precision. Such imprecise predictions may, however, be useful in making decisions about the need to control exposures to aluminium in the general population.
The degree of aluminium absorption depends on a number of parameters, such as the aluminium salt administered, pH (for aluminium speciation and solubility), bioavailability, and dietary factors. These should be taken into consideration during tissue dosimetry and response assessment. Thus, the use of currently available animal studies to develop a guideline value is not appropriate because of these specific toxicokinetic/dynamic considerations. Owing to the uncertainty surrounding the human data and the limitations of the animal data as a model for humans, a health-based guideline value for aluminium cannot be derived at this time. The beneficial effects of the use of aluminium as a coagulant in water treatment are recognized. Taking this into account and considering the potential health concerns (i.e. neurotoxicity) of aluminium, a practicable level is derived based on optimization of the coagulation process in drinking-water plants using aluminium-based coagulants, to minimize aluminium levels in finished water. A number of approaches are available for minimizing residual aluminium concentrations in treated water. These include use of optimum pH in the coagulation process, avoiding excessive aluminium dosage, good mixing at the point of application of the coagulant, optimum paddle speeds for flocculation, and efficient filtration of the aluminium floc (Letterman & Driscoll, 1988; WRc, 1997). Under good operating conditions, concentrations of aluminium of 0.1 mg/litre or less are achievable in large water treatment facilities. Small facilities (e.g. those serving fewer than 10 000 people) might experience some difficulties in attaining this level, because the small size of the plant provides little buffering for fluctuation in operation, and small facilities often have limited resources and access to expertise to solve specific operational problems. For these small facilities, 0.2 mg/litre or less is a practicable level for aluminium in finished water (WRc, 1997).
8. REFERENCES
A TSDR (1992) Toxrcological profile for aluminrum. Atlanta, GA, US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry (TP-91/01). 12 INORGANIC CONSTITUENTS
CCME ( 1988) Canadzan water quality gwdelmes Ottawa, Ontario, Canadian Council of M misters of the Environment. Clayton DB (1989) Water pollutwn at Lowermoore North Cornwall. Report of the Lowermoore mcident health advismy comn11ttee Truro, Cornwall District Health Authority, 22 pp. Colomina MT et al. ( 1992) Concurrent ingestion of lactate and aluminum can result in developmental toxicity in mice. Research commumcatwns in chemtcal pathology and pharmacology, 77:95-106. Colomina MT et al. (1994) Lack of maternal and developmental toxicity in mice given high doses of aluminium hydroxide and ascorbic acid during gestation Pharmacology and toxicology, 74:236-239. Commissaris RL et al. (1982) Behavioral changes m rats after chronic aluminium and parathyroid hormone administratiOn. Neurobehavwral toxicology and teratology, 4:403-410 Connor DJ, Jape RS, Harrell LE (1988) Chronic, oral aluminum administration to rats: cognition and cholinergic parameters. Pharmacology, bwchemistry, and behavwur, 31:467-474. Domingo JL et al. (1987a) NutritiOnal and toxicological effects of short-term mgestion of aluminum by the rat. Research commumcatwns 111 chemical pathology and pharmacology, 56:409-419. Domingo JL et al. (1987b) Effects of oral aluminum administration on perinatal and postnatal development in rats. Research commumcations in chemical pathology and pharmacology, 57:129-132. Domingo JL et al (1987c) The effects of aluminium ingestion on reproduction and postnatal survival in rats. Life sciences, 41:1127-1131. Domingo JL et al. (1989) Lack of teratogenicity of aluminum hydroxide m rats. Life sciences, 45:243-247 FAO/WHO (1989) Aluminium. In: Toxicological evaluatwn ofcertamfood additives and contammants Thirty-thu·d meetmg of the Jomt FAO/WHO Expert Committee on Food Additives Geneva, World Health Organization, pp. 113-154 (WHO Food Additives Series 24). Flaten TP (1990) Geographical associations between aluminium in drinking water and death rates with dementia (including Alzheimer's disease), Parkinson's disease and amyotrophic lateral sclerosis in Norway Environmental geochemistry and health, 12:152-167 Forbes WF et al (1994) Geochemical risk factors for mental functioning, based on the Ontario longitudmal study of aging (LSA). 11. The role of pH Canadwn journal ofagmg, 13:249-267. Gomez M et al. (1986) Short-term oral toxicity study of aluminium in rats. Archives ofpharmacology and toxicology, 12:145-151. Gomez M et al ( 1990) Evaluation of the maternal and developmental toxicity of aluminum from high doses of aluminum hydroxide in rats. Veterinary and human toxicology, 32:545-548. Gomez M, Domingo JL, Llobet JM (1991) Developmental toxicity evaluation of oral aluminum in rats: influence of citrate. Neurotoncology and teratology, 13:323-328. Health Canada (1993) Guidelines for Canadian drmking water quality Water treatment principles and applications. a manual for the productwn of drinking water. Ottawa, Ontario, Health Canada, Environmental Health Directorate. Printed and distributed by Canadian Water and Wastewater Association, Ottawa, Ontario. Hicks JS, Hackett DS, Sprague GL (1987) Toxicity and aluminium concentration in bone following dietary administration of two sodium aluminium phosphate formulations in rats. Food and chemtcal toxicology, 25(7)·533-538. ISO (1994) Water qua!tty- Determinatwn of aluminium· spectrometnc method usmg pyrocatechol violet. Geneva, International Organization for Standardization (ISO I 0566.1994 (E)). ISO (1996) Water quality- Determmatwn of 33 elements by inducflvely coupled plasma atomic emisswn spectroscopy. Geneva, International Organization for Standardization (ISO 11885:1996 (E)). ISO (1997) Water quality - Determmation of aluminium - Atomic absorption spectrometric methods. Geneva, International Organization for Standardization (ISO 12020: 1997 (E)). Jacqmin H et al. (1994) Components of drinking water and risk of cognitive impairment in the elderly. American Journal ofepidemiology, 139:48-57. Katz AC et al. (1984) A 6-month dietary toxicity study of acidic sodium aluminium phosphate in beagle dogs. Food and chemical toxicology, 22.7-9. ALUMINIUM 13
Letterman RD, Driscoll CT ( 1988) Survey of residual aluminum in filtered water. Joumal of the Amencan Water Works Assocwtwn, 80.154-158. Lide DR, ed ( 1993) CRC handbook ofchemist!)' and phys1cs, 73rd ed. Boca Raton, FL, CRC Press. MAFF ( 1993) Alummwm m food- 39th report of the Steenng Group on Chemical Aspects of Food Sur1'edlance London, UK Ministry of Agnculture, Fisheries and Food. Martyn CN et al. ( 1989) Geographical relation between A1zheimer' s d1sease and alummiUm in drinking water. Lancet, I 59-62 Miche1 Pet al. (1991) Study of the relationship between aluminium concentration in drinking water and nsk of Alzheimer's d1sease. In. lqbal Ketal., eds. Al::heuner's disease. Bas1c mechanisms, diagnosis a11d therapeutic strategies New York, NY, John Wiley, pp. 387-391 Miller RG et al (1984) The occurrence of alummum m dnnking water. Journal of the Amencan Water Works Assocwtwn, 76·84-91. Neri LC, Hewitt D (1991) AlummiUm, Alzheimer's disease, and drinking water. Lancet, 338.390. OMEE (1995) Drinkmg water surveillance programme - Province of Ontario. Toronto, Ontario, Ontario Ministry of Environment and Energy. Pennington JA, Schoen SA ( 1995) Estimates of dietary exposure to aluminium. Food additives and contanunants, 12(1)·119-128. Pettersen JC, Hackett DS, Zwicker GM (1990) Twenty-six week toxicity study with KASAL (basic sodium aluminum phosphate) in beagle dogs. Environmental geochemistry and health, 12:121-123. Schroeder MA, Mitchener M (1975a) Life-term effects of mercury, methylmercury and nine other trace metals on mice. Journal ofnutritwn, 105:452-458. Schroeder MA, Mitchener M (1975b) Life-term studies in rats: effects of aluminum, barium, beryllium, and tungsten Journal ofnutrifwn, 105'421-427. Thorne BM et al. (1987) Aluminum toxicity and behavior in the weanling Long-Evans rat. Bulletin of the Psychonomic Socwty, 25: 129-132 Wettstein A et al. ( 1991) Failure to find a relationship between mnestic skills of octogenarians and aluminium in drinking water. International archives of occupational and environmental health, 63:97-103. WHO (1996) Guidelines for drinkmg-water quality, 2nd ed. Vol. 2. Health cntena and other supportmg mformatwn Geneva, World Health Organization. WHO ( 1997) Alumimum Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria 194). Wilhelm M, !del H (1995) [Aluminium in drinking water. a risk factor for Alzheimer's disease?) Forum, Stadte, Hygiene, 46:255-258 (in German). WRc (1997) Treatment technology for aluminium, boron and uramum. Document prepared for WHO by the Water Research Centre, Medmenham, and reviewed by S. Clark, US EPA; A. van Dijk-Looijaard, KIWA, Netherlands; and D. Green, Health Canada. Yen-Koo HC (1992) The effect of aluminum on conditioned avoidance response (CAR) in mice. Toxicology and mdustnal health, 8:1-7.
BORON
First draft prepared by C. Smallwood Environmental Criteria and Assessment Office United States Environmental Protection Agency, Cincinnati, Ohio, USA
In the 1993 WHO Guidelines for drinking-water quality, a health-based guideline value for boron of 0.3 mg/litre was derived from a 2-year study in dogs published in 1972 (testicular atrophy was the critical end-point of toxicity). New scientific data have become available, including an IPCS Environmental Health Criteria monograph for boron which had been finalized by a Task Group m November 1996, and which was therefore available for the present evaluation.
1. GENERAL DESCRIPTION
1.1 Identity
Boron (CAS no. 7440-42-8) is never found in the elemental form in nature. It exists as a mixture of the 10B (19.78%) and 11 B (80.22%) isotopes (Budavari et al., 1989). Boron's chemistry is complex and resembles that of silicon (Cotton & Wilkinson, 1988).
1.2 Physicochemical properties
Elemental boron exists as a solid at room temperature, either as black monoclinic crystals or as a yellow or brown amorphous powder when impure. The amorphous and crystalline forms of boron have specific gravities of 2.3 7 and 2.34, respectively. Boron is a relatively inert metalloid except when in contact with strong oxidizing agents. Sodium perborates are persalts, which are hydrolytically unstable because they contain characteristic boron-oxygen-oxygen bonds that react with water to form hydrogen peroxide and stable sodium metaborate (NaB02·nHP). Boric acid is a very weak acid, with a pK. of 9.15, and therefore boric acid and the sodium borates exist predominantly as undissociated boric acid [B(OH)3] in dilute aqueous solution at pH <7; at pH> 10, the metaborate anion B(OH)4- becomes the main species in solution. Between these two pH values, from about 6 to 11, and at high concentration (>0.025 mol/litre), highly water soluble polyborate ions such as B30 3(0H)4-, B40 5(0H)4-, and B50 6(0H)4- are formed. The chemical and toxicological properties of borax pentahydrate
Na2BP7·5Hp, borax Na2B40 7 ·1 OHp, boric acid, and other borates are expected to be similar on a molar boron equivalent basis when dissolved in water or biological fluids at the same pH and low concentration. 16 INORGANIC CONSTITUENTS
1.3 Major uses
Boric acid and borates are used in glass manufacture (fibreglass, borosilicate glass, enamel, frit, and glaze), soaps and detergents, flame retardants, and neutron absorbers for nuclear installations. Boric acid, borates, and perborates have been used in mild antiseptics, cosmetics, pharmaceuticals (as pH buffers), boron neutron capture therapy (for cancer treatment), pesticides, and agricultural fertilizers.
1.4 Environmental/ate
Waterborne boron may be adsorbed by soils and sediments. Adsorption desorption reactions are expected to be the only significant mechanism influencing the fate of boron in water (Rai et al., 1986). The extent of boron adsorption depends on the pH of the water and the concentration of boron in solution. The greatest adsorption is generally observed at pH 7.5-9.0 (Waggott, 1969; Keren & Mezuman, 1981; Keren et al., 1981). In natural waters, boron exists primarily as undissociated boric acid with some borate ions. As a group, the boron-oxygen compounds are sufficiently soluble in water to achieve the levels that have been observed (Sprague, 1972). Mance et al. (I988) described boron as a significant constituent of seawater, with an average boron concentration of 4.5 mg/kg.
2. ANALYTICAL METHODS
A spectrometric method using azomethine-H is available for the determination of borate in water. The method is applicable to the determination of borate at concentrations between O.OI and I mg/litre. The working range may be extended by dilution (ISO, I990). A widely used method for the analysis of boron in bone, plasma, and food is inductively coupled plasma atomic emission spectroscopy (Hunt, I989). This method is also used for water (ISO, I996) and wastewater (Huber, 1982). Detection limits in water range from 6 to I 0 J.!g of boron per litre. Inductively coupled plasma mass spectroscopy (ICP-MS) is a widely used non-spectrophotometric method for the analysis of boron, as it uses small volumes of sample, is fast, and applies to a wide range of materials (fresh and saline water, sewage, wastewater, soils, plant samples, and biological materials). ICP-MS can detect boron down to 0.15 J.!g/litre (WHO, in press). Using direct nebulization, ICP-MS can give a detection limit of I ng/g in human blood, human serum, orchard leaves, and total diet (Smith et al., I99I).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Boron is not present in the atmosphere at significant levels (Sprague, I972). Because borates exhibit low volatility, boron would not be expected to be present to BORON 17 a significant degree as a vapour in the atmosphere. Atmospheric emissions of borates and boric acid in a particulate (<1--45 f.!m in size) or vapour form occur as a result of volatilization of boric acid from the sea, volcanic activity, mining operations, glass and ceramic manufacturing, the application of agricultural chemicals, and coal-fired power plants.
3.2 Water
The natural borate content of groundwater and surface water is usually small. The borate content of surface water can be significantly increased as a result of wastewater discharges, because borate compounds are ingredients of domestic washing agents (ISO, 1990). Naturally occurring boron is present in groundwater primarily as a result of leaching from rocks and soils containing borates and borosilicates. Concentrations of boron in groundwater throughout the world range widely, from <0.3 to > 100 mg/litre. In general, concentrations of boron in Europe were greatest in southern Europe (Italy, Spain) and least in northern Europe (Denmark, France, Germany, the Netherlands, and the United Kingdom). For Italy and Spain, mean boron concentrations ranged from 0.5 to 1.5 mg/litre. Values ranged up to approximately 0.6 mg/litre in the Netherlands and the United Kingdom, and approximately 90% of samples in Denmark, France, and Germany were found to contain boron at concentrations below 0.3, 0.3, and 0.1 mg/litre, respectively (WHO, in press). Monthly mean values of boron in the Ruhr River, Germany, ranged from 0.31 to 0.37 mg/litre in a survey conducted during 1992-1995 (Haberer, 1996). The majority of the Earth's boron occurs in the oceans, with an average concentration of 4.5 mg/litre (Weast et al., 1985). The amount of boron in fresh water depends on such factors as the geochemical nature of the drainage area, proximity to marine coastal regions, and inputs from industrial and municipal effluents (Butterwick et al., 1989). Boron concentrations in fresh surface water range from <0.001 to 2 mg/litre in Europe, with mean values typically below 0.6 mg/litre. Similar concentration ranges have been reported for water bodies within Pakistan, Russia, and Turkey, from 0.01 to 7 mg/litre, with most values below 0.5 mg/litre. Concentrations ranged up to 0.01 mg/litre in Japan and up to 0.3 mg/litre in South African surface waters. Samples taken in surface waters from two South American rivers (Rio Arenales, Argentina, and Loa River, Chile) contained boron at concentrations ranging between 4 and 26 mg/litre in areas rich in boron-containing soils. In other areas, the Rio Arenales contained less than 0.3 mg of boron per litre. Concentrations of boron in surface waters of North America (Canada, USA) ranged from 0.02 mg/litre to as much as 360 mg/litre, indicative of boron-rich deposits. However, typical boron concentrations were less than 0.1 mg/litre, with a 90th-percentile boron concentration of approximately 0.4 mg/litre. Concentrations of boron found in drinking-water from Chile, Germany, the United Kingdom, and the USA ranged from 0.01 to 15.0 mg/litre, with most values clearly below 0.4 mg/litre. These values are consistent with ranges and means observed for groundwater and surface waters. This consistency is supported by two 18 INORGANIC CONSTITUENTS factors: (i) boron concentrations in water are largely dependent on the leaching of boron from the surrounding geology and wastewater discharges, and (ii) boron is not removed by conventional drinking-water treatment methods.
3.3 Food
The general population obtains the greatest amount of boron through food intake. Concentrations of boron reported in food after 1985 have more validity because of the use of more adequate analytical methods. The richest sources of boron are fruits, vegetables, pulses, legumes, and nuts. Dairy products, fish, meats, and most grains are poor sources of boron. Based on the recent analyses of foods and food products, estimations of daily intakes of various age/sex groups have been made (WHO, in press). The estimated median, mean, and 95th-percentile daily intakes of boron were 0.75, 0.93, and 2.19 mg/day, respectively, for all groups, and 0.79, 0.98 and 2.33 mg/day, respectively, for adults aged 17 and older. Using food included in US Food and Drug Administration Total Diet Studies, lyengar et al. (1988) determined the mean adult male daily intake of boron to be 1.52 mg/day, whereas Anderson et al. (1994) determined the intake to be 1.21 mg/day. Based on the United Kingdom National Food Survey (MAFF, 1991), the dietary intake of boron in the United Kingdom ranges from 0.8 to 1.9 mg/day. It should be noted that increased consumption of specific foods with high boron content will increase boron intake significantly; for example, one serving of wine or avocado provides 0.42 and 1.11 mg, respectively (Anderson et al., 1994).
3.4 Estimated total exposure and relative contribution of drinking-water
The mean daily intake of boron in the diet is judged to be near 1.2 mg/day (Anderson et al., 1994 ). Concentrations of boron in drinking-water have wide ranges, depending on the source of the drinking-water, but for most of the world the range is judged to be between 0.1 and 0.3 mg/litre. Based on usage data, consumer products have been estimated to contribute a geometric mean of 0.1 mg/day to the estimate of total boron exposure (WHO, in press). The contribution of boron intake from air is negligible. The total daily intake can therefore be estimated from mean concentrations and concentration ranges to be between 1.5 and 2 mg.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
Numerous studies have shown that boric acid and borax are absorbed from the gastrointestinal tract and from the respiratory tract, as indicated by increased levels of boron in the blood, tissues, or urine or by systemic toxic effects of exposed individuals or laboratory animals. Clearance of boron compounds is similar in humans and animals. The ratio of mean clearance values as a function of dose in non-pregnant rats versus humans is approximately 3- to 4-fold - i.e. similar to the default value for the toxicokinetic BORON 19 component of the uncertainty factor for interspecies variation1 (WHO, 1994). Elimination of borates from the blood is largely by excretion of >90% of the administered dose via the urine, regardless of the route of administration. Excretion is relatively rapid, occurring over a period of a few to several days, with a half-life of elimination of 24 hours or less. The kinetics of elimination of boron have been evaluated in human volunteers given boric acid via the intravenous and oral routes (Jansen et al., 1984; Schou et al., 1984). Absorption is poor through intact skin but is much greater through damaged skin.
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
The oral LD50 values for boric acid or borax in mice and rats are in the range of about 400-700 mg of boron per kg of body weight (Pfeiffer et al., 1945; Weir &
Fisher, 1972). Oral LD50 values in the range of250-350 mg of boron per kg ofbody weight for boric acid or borax exposure have been reported for guinea-pigs, dogs, rabbits, and cats (Pfeiffer et al., 1945; Verbitskaya, 1975). Signs of acute toxicity for both borax and boric acid in animals given single large doses orally include depression, ataxia, convulsions, and death; kidney degeneration and testicular atrophy are also observed (Larsen, 1988).
5.2 Short-term exposure
In a 13-week study, mice (10 per sex per dose) were fed diets containing boric acid at approximately 0, 34, 70, 141, 281, or 563 mg of boron per kg of body weight per day. At the two highest doses, increased mortality was seen. Degeneration or atrophy of the seminiferous tubules was observed at 141 mg of boron per kg of body weight per day. In all dose groups, extramedullary haematopoiesis of the spleen of minimal to mild severity was seen (NTP, 1987). In a study in which borax was given in the diet to male Sprague-Dawley rats (18 per dose) at concentrations of 0, 500, I 000, or 2000 mg of boron per kg of feed (approximately equal to 0, 30, 60, or 125 mg of boron per kg of body weight per day) for 30 or 60 days, body weights were not consistently affected by treatment. Organ weights were not affected by 500 mg of boron per kg of feed; at 1000 and 2000 mg of boron per kg of feed, absolute liver weights were significantly lower after 60 days, and epididymal weights were significantly lower (37.6% and 34.8%, respectively) after 60 days, but not after 30 days. Weights of prostate, spleen, kidney, heart, and lung were not changed at any dose (Lee et al., 1978). In a 90-day study in rats (10 per sex per dose) receiving 0, 2.6, 8.8, 26, 88, or 260 mg of boron per kg of body weight per day in the diet as boric acid or borax, all
Report of informal discussion to develop recommendations for the WHO Guidelines for drinking water qua!Jty- Boron. Cincinnati, OH, 28-29 September 1997. 20 INORGANIC CONSTITUENTS animals at the highest dose died within 3-6 weeks (Weir & Fisher, 1972). In animals receiving 88 mg of boron per kg of body weight per day, body weights in males and females were reduced; absolute organ weights, including the liver, spleen, kidneys, brain, adrenals, and ovaries, were also significantly decreased in this group. Organ to-body-weight ratios for the adrenals and kidneys were significantly increased, but relative weights of the liver and ovaries were decreased. A pronounced reduction in testicular weights in males in the 88 mg of boron per kg of body weight per day group was also observed. Boric acid or borax was also fed to beagle dogs for 90 days or for 2 years. In the 90-day boric acid study (weight-normalized doses of 0, 0.44, 4.4, or 44 mg of boron per kg of body weight per day; five animals per sex per dose), testis weight was significantly lower than controls in the middle and upper dose groups (reduced by 25% and 40%, respectively). Although testicular microscopic structure was not detectably abnormal in the controls and middle dose group, four of five dogs in the high-dose group had complete atrophy, and the remaining high-dose dog had one third of tubules showing some abnormality. In the borax study, testis weights in the low-, middle-, and high-dose groups were 80%, 85%, and 50% of controls, respectively; only the last was significantly different from controls. No mention was made of the testicular microscopic structure of the controls or low-dose animals; middle-dose animals were not detectably altered (aside from the considerable fixation-induced artifact in the outer third of the tissue), whereas four of five high dose dogs had complete testicular atrophy, and the remaining high-dose dog had "partial" atrophy. No other clinical or microscopic signs of toxicity were reported in any animals (Weir & Fisher, 1972). In the 2-year study, the dogs (four per sex per dose) received the boric acid or borax in the diet at weight-normalized doses ofO, 1.5, 2.9, or 8.8 mg ofboron per kg of body weight per day. An additional group received 29 mg of boron per kg of body weight per day for 38 weeks. Testicular atrophy was observed in two test dogs receiving borax at 26 weeks and in the two and one dogs, respectively, killed after 26 or 38 weeks of boric acid consumption. The authors stated that boric acid caused testicular degeneration in dogs, including spermatogenic arrest and atrophy of the seminiferous epithelium. The study was terminated at 38 weeks. In these studies, the number of dogs was small and variable (one or two dogs at each of three time points) and inadequate to allow statistical analysis. All three treated dogs had widespread and marked atrophy in 25--40% of the seminiferous tubules. A common control group was used for both the borax and boric acid studies. Testicular lesions occurred in the controls (one of four controls had slight to severe seminiferous tubular atrophy, another had moderate to severe atrophy, whereas a third had a detectable but insignificant reduction in spermatogenesis and 5% atrophic seminiferous tubules) (Weir & Fisher, 1972). These studies were conducted before the advent of Good Laboratory Practices (GLPs). Confidence in these studies is low, and they were considered not suitable for inclusion into the risk assessment because of 1) small and variable numbers of dogs, 2) variable background lesions in controls leading to uncertainty of the strength of the response to treatment, 3) lack of GLPs, and 4) other, more recent studies of greater scientific quality with findings at similar intake levels of boron (Ku et al., 1993; Price et al., 1996a). BORON 21
5.3 Long-term exposure
A 2-year study in mice (50 per sex per dose) receiving approximately 0, 275, or 550 mg of boric acid per kg of body weight per day (0, 48, or 96 mg of boron per kg of body weight per day) in the diet (NTP, 1987; Dieter, 1994) demonstrated that body weights were 10-17% lower in high-dose males after 32 weeks and in high dose females after 52 weeks. Increased mortality rates were statistically significant in males, with significant lesions in male mice appearing in the testes and no significant non-neoplastic lesions in female mice. In a 2-year study, rats (35 per sex per dose) were administered weight normalized boron doses ofO, 5.9, 18, or 59 mg/kg of body weight per day in the diet (Weir & Fisher, 1972). High-dose animals had coarse hair coats, scaly tails, hunched posture, swollen and desquamated pads of the paws, abnormally long toenails, shrunken scrotum, inflamed eyelids, and bloody eye discharge. The haematocrit and haemoglobin levels were significantly lower than controls, the absolute and relative weights of the testes were significantly lower, and relative weights of the brain and thyroid gland were higher than in controls. In animals in the mid- and low-dose groups, no significant effects on general appearance, behaviour, growth, food consumption, haematology, serum chemistry, or histop'athology were observed.
5.4 Reproductive and developmental toxicity
Short- and long-term oral exposures to boric acid or borax in laboratory animals have demonstrated that the male reproductive tract is a consistent target of toxicity. Testicular lesions have been observed in rats, mice, and dogs administered boric acid or borax in food or drinking-water (Truhaut et al., 1964; Weir & Fisher, 1972; Green et al., 1973; Lee et al., 1978; NTP, 1987; Ku et al., 1993). The first clinical indication of testicular toxicity in dogs is shrunken scrota observed during treatment; significant decreases in absolute and relative testicular weight are also reported. After subchronic exposure, the histopathological effects range from inhibited spermiation (sperm release) to degeneration of the seminiferous tubules with variable loss of germ cells to complete absence of germ cells, resulting in atrophy and transient or irreversible loss of fertility, but not of mating behaviour. In time-response and dose-response reproductive studies (Linder et al., 1990), adult male Sprague-Dawley rats were administered two doses in one day, with a total dose ofO or 350 mg of boron per kg of body weight in the time-response experiment (animals were sacrificed at 2, 14, 28, or 57 days post-treatment) and a total dose of 0, 44, 87, 175, or 350 mg of boron per kg of body weight in the dose response experiment (animals were sacrificed after 14 days). Adverse effects on spermiation, epididymal sperm morphology, and caput sperm reserves were observed during histopathological examinations of the testes and epididymis. The NOAEL for male reproductive effects in the dose-response study was 87 mg of boron per kg of body weight per day. In a multi-generation study, doses of 0, 117, 350, or 1170 mg of boron per kg of feed (as borax or boric acid) were administered to male and female rats (Weir & Fisher, 1972). At the highest dose, rats were found to be sterile, males showed 22 INORGANIC CONSTITUENTS atrophied testes in which spermatozoa were absent, and females showed decreased ovulation. The NOAEL in this study was 350 mg of boron per kg of feed, equivalent to 17.5 mg of boron per kg of body weight per day. To investigate the development of testicular lesions, boric acid was fed at 61 mg of boron per kg of body weight per day to male F344 rats; sacrifice of six treated and four control rats was conducted at intervals from 4 to 28 days. At 28 days, there was significant loss of spermatocytes and spermatids from all tubules in exposed rats, and basal serum testosterone levels were significantly decreased from 4 days on (Treinen & Chapin, 1991). In another study, the activities of enzymes found primarily in spermatogenic cells were decreased, and enzyme activities associated with premeiotic spermatogenic cells were significantly increased in rats exposed to 60 or 125 mg of boron per kg of body weight per day for 60 days (Lee et al., 1978). Mean plasma follicle-stimulating hormone levels were significantly elevated in a dose-dependent manner in all treatment groups (30, 60, or 125 mg of boron per kg of body weight per day) in this study after 60-day exposures. Reversibility of testicular lesions was evaluated by Ku et al. ( 1993) in an experiment in which F344 rats were dosed at 0, 3000, 4500, 6000, or 9000 mg of boric acid per kg of feed (equivalent to 0, 26, 39; 52, or 78 mg of boron per kg of body weight per day) for 9 weeks and assessed for recovery up to 32 weeks post treatment. Inhibited spermiation was exhibited at 3000 and 4500 mg of boric acid per kg of feed (5.6 J..lg of boron per mg of tissue), whereas inhibited spermiation progressed to atrophy at 6000 and 9000 mg of boric acid per kg of feed (11.9 J..lg of boron per mg of testes); there was no boron accumulation in the testes to levels greater than those found in the blood during the 9-week period. After treatment, serum and testis boron levels in all dose groups fell to background levels. Inhibited spermiation at 4500 mg of boric acid per kg of feed was reversed by 16 weeks post treatment, but focal atrophy, which did not recover up to 32 weeks post-treatment, was detected. Developmental toxicity has been demonstrated experimentally in rats, mice, and rabbits (NTP, 1990; Heindel et al., 1992; Price et al., 1996b). Rats were fed a diet containing 0, 14, 29, or 58 mg of boron per kg of body weight per day as boric acid on gestation days 0-20 (Heindel et al., 1992). An additional group of rats received boric acid at 94 mg of boron per kg of body weight per day on gestation days 6-15 only. Average fetal body weight per litter was significantly reduced in a dose-related manner in all treated groups compared with controls. The percentage of malformed fetuses per litter and the percentage of litters with at least one malformed fetus were significantly increased at :2:29 mg of boron per kg of body weight per day. Malformations consisted primarily of anomalies of the eyes, the central nervous system (CNS), the cardiovascular system, and the axial skeleton. The most common malformations were enlargement of lateral ventricles in the brain and agenesis or shortening of rib XIII. The LOAEL of 14 mg of boron per kg of body weight per day (the lowest dose tested) for rats occurred in the absence of maternal toxicity; a NOAEL was not found in this study. Price et al. (1996a) did a follow-up to the Heindel et al. (1992) study in Sprague-Dawley (CD) rats to determine a NOAEL for fetal body-weight reduction and to determine whether the offspring would recover from prenatally reduced body BORON 23 weight during postnatal development. Boric acid was administered in the diet to CD rats on gestation days 0-20. Dams were terminated and uterine contents examined on gestation day 20. The intake of boric acid was 0, 3.3, 6.3, 9.6, 13, or 25 mg of boron per kg of body weight per day. Fetal body weights were 99, 98, 97, 94, and 88% of controls for the low- to high-dose groups, respectively. Incidences of short rib XIII (a malformation) or wavy rib (a variation) were increased in the 13 and 25 mg of boron per kg of body weight per day dose groups relative to control litters. There was a decreased incidence of rudimentary extra rib on lumbar 1 (a variation) in the high-dose group that was deemed biologically but not statistically significant. The NOAEL in this study was 9.6 mg of boron per kg of body weight per day, based on a decrease in fetal body weight at the next higher dose. Developmental toxicity and teratogenicity of boric acid in mice at 0, 43, 79, or 175 mg of boron per kg of body weight per day in the diet were investigated (Heindel et al., 1992). There was a significant dose-related decrease in average fetal body weight per litter at 79 and 175 mg of boron per kg of body weight per day. In offspring of mice exposed to 79 or 175 mg of boron per kg of body weight per day during gestation days 0-20, there was an increased incidence of skeletal (rib) malformations. These changes occurred at doses for which there were also signs of maternal toxicity (increased kidney weight and pathology); the LOAEL for developmental effects (decreased fetal body weight per litter) was 79 mg of boron per kg of body weight per day, and the NOAEL was 43 mg of boron per kg of body weight per day. Developmental toxicity and teratogenicity of boric acid in rabbits were investigated by Price et al. (1996b) at doses of 0, 11, 22, or 44 mg of boron per kg of body weight per day, given by gavage. Frank developmental effects in rabbits exposed to 44 mg of boron per kg of body weight per day included a high rate of prenatal mortality, an increased number of pregnant females with no live fetuses, and fewer live fetuses per live litter on day 30. At the high dose, malformed live fetuses per litter increased significantly, primarily because of the incidence of fetuses with cardiovascular defects, the most prevalent of which was interventricular septal defect. Skeletal variations observed were extra rib on lumbar 1 and misaligned sternebra. The NOAEL for maternal and developmental effects was 22 mg of boron per kg of body weight per day.
5.5 Mutagenicity and related end-points
The mutagenic activity of boric acid was examined in the Salmonella typhimurium and mouse lymphoma assays, with negative results. No induction of sister chromatid exchange or chromosomal aberrations was observed in Chinese hamster ovary cells (NTP, 1987). Sodium borate did not cause gene mutations in the S. typhimurium preincubation assay (Benson et al., 1984). Borax was not mutagenic in cell transformation assays with Chinese hamster cells, mouse embryo cells, and human fibroblasts (Landolph, 1985). 24 INORGANIC CONSTITUENTS
5.6 Carcinogenicity
Tumour incidence was not enhanced in studies in which B6C3F 1 mice received 0, 2500, or 5000 mg of boric acid per kg of feed for 103 weeks (NTP, 1987) and Sprague-Dawley rats received diets containing 0, 117, 350, or 1170 mg of boron per kg of feed (as borax or boric acid) for 2 years (Weir & Fisher, 1972).
6. EFFECTS ON HUMANS
Available human data on boron compounds for routes other than inhalation focus on boric acid and borax. According to Stoking er (1981 ), the lowest reported lethal doses of boric acid are 640 mg/kg of body weight (oral), 8600 mg/kg of body weight (dermal), and 29 mg/kg of body weight (intravenous injection). Stokinger ( 1981) stated that death has occurred at total doses of between 5 and 20 g of boric acid for adults and <5 g for infants. Litovitz et al. (1988) stated that potential lethal doses are usually cited as 3-6 g total for infants and 15-20 g total for adults. A case series report of seven infants (aged 6-16 weeks) who used pacifiers coated with a borax and honey mixture for 4-10 weeks concluded that exposures ranged from 12 to 90 g. with a very crudely estimated average daily ingestion of 18-56 mg of boron per kg of body weight (O'Sullivan & Taylor, 1983). 2 Toxicity was manifested by generalized or alternating focal seizure disorders, irritability, and gastrointestinal disturbances. Although infants appear to be more sensitive than adults to boron compounds, lethal doses are not well documented in the literature. Goldbloom & Goldbloom (1953) reported four cases of boric acid poisoning and reviewed an additional 109 cases in the literature. The four cases were infants exposed to boric acid by repeated topical applications of baby powder. Toxicity was manifested by cutaneous lesions (erythema over the entire body, excoriation of the buttocks, and desquamation), gastrointestinal disturbances, and seizures. Approximately 35% of the 109 other case reports of boric acid poisoning involved children <1 year of age. The mortality rate was 70.2% for children, compared with 55.0% for all cases combined. Death occurred in 53% of patients exposed by ingestion, 75% of patients subjected to gastric lavage with boric acid, 68% of patients exposed by dermal application for treating burns, wounds, and skin eruptions, and 54% of patients exposed by other routes. Information on signs and symptoms for 80 patients showed that gastrointestinal disturbances were prevalent (73%), followed by CNS effects (67%). Cutaneous lesions were prevalent in 76% ofthe cases and in 88% of cases involving children <2 years of age. Gross and microscopic findings were reported for 45% of fatal cases. In general, boric acid caused chemical irritation primarily at sites of application and excretion and in organs with maximum boron concentrations. The most common CNS findings were oedema and congestion of the brain and meninges. Other common findings
Estimates given here are corrected values, as intakes reported in this publication were underestimated by a factor of 3 (M." Taylor, personal communication to M. Dourson, in a letter dated 28 August 1997). BORON 25 included liver enlargement, vascular congestion, fatty changes, swelling, and granular degeneration. In addition to case reports, poison centres have published case-series reports. Unlike the case reports reviewed by Goldbloom & Goldbloom (1953), more recent reports suggest that the oral toxicity of boron in humans is milder than previously thought. Litovitz et al. (1988) conducted a retrospective review of 784 cases of boric acid ingestion reported to the National Capital Poison Center in Washington, DC, USA, during 1981-1985 and the Maryland Poison Center in Baltimore, MD, USA, during 1984-1985; approximately 88.3% of the cases were asymptomatic. All but two of the cases had acute (single) ingestion, and 80.2% involved children <6 years of age. No severe toxicity or life-threatening effects were noted, although boric acid levels in blood serum ranged from 0 to 340 J.!g/ml. The most frequently occurring symptoms, which involved the gastrointestinal tract, included vomiting, abdominal pain, diarrhoea, and nausea. Other symptoms (primarily CNS and cutaneous) occurred in fewer cases: lethargy, rash, headache, light-headedness, fever, irritability, and muscle cramps. The average dose ingested was estimated at 1.4 g. According to Litovitz et al. ( 1988), 21 of the children <6 years of age, 15 of whom were <2 years of age, ingested the reported potential lethal dose of 3 g; eight adults ingested the reported potential lethal dose of 15 g without clinical evidence of lethal effects. Linden et al. (1986) published a retrospective review of 364 cases of boric acid exposure reported to the Rocky Mountain Poison and Drug Center in Denver, CO, USA, between 1983 and 1984. Vomiting, diarrhoea, and abdominal pain were the most common symptoms given by the 276 cases exposed in 1983. Of the 72 cases reported in 1984 for whom medical records were complete, 79% were asymptomatic, whereas 20% noted mild gastrointestinal symptoms. One 2-year-old child died, presumably from repeated ingestion of an insecticide containing 99% boric acid. Overall, owing to the wide variability of data collected from poisoning centres, the average dose of boric acid to produce clinical symptoms is still unclear, presumably in the range of 100 mg to 55.5 g, reported by Litovitz et al. (1988). Findings from human experiments show that boron is a dynamic trace element that can affect the metabolism or utilization of numerous substances involved in life processes, including calcium, copper, magnesium, nitrogen, glucose, triglycerides, reactive oxygen, and estrogen. Although the first findings involving boron deprivation of humans appeared in 1987 (Nielsen et al., 1987), the most convincing findings have come mainly from two studies in which men over the age of 45, postmenopausal women, and postmenopausal women on estrogen therapy were fed a low-boron diet (0.25 mg/2000 kcal) for 63 days and then fed the same diet supplemented with 3 mg of boron per day for 49 days (Nielsen, 1989, 1994; Nielsen et al., 1990, 1991, 1992; Penland, 1994). These dietary intakes were near the low and high values in the range of usual dietary boron intakes. The major differences between the two studies were the intakes of copper and magnesium: in one experiment, they were marginal or inadequate; in the other, they were adequate. The marginal or inadequate copper and magnesium intakes caused apparent detrimental changes that were more marked during boron deprivation than during boron repletion. Although the function of boron remains undefined, boron is 26 INORGANIC CONSTITUENTS becoming recognized as an element of potential nutritional importance because of the findings from human and animal studies.
7. PROVISIONAL GUllJELINE VALUE
The TDI of boron is derived by dividing the NOAEL (9.6 mg/kg of body weight per day) for the critical effect, which is developmental toxicity (decreased fetal body weight in rats), by an appropriate uncertainty factor, which is judged to be 60. The value of 10 for interspecies variation (animals to humans) was adopted because of lack of toxicokinetic and toxicodynamic data to allow deviation from this default value. Available toxicokinetic data do support, however, reduction of the default uncertainty factor for intraspecies variation from 10 to 6 (WHO, 1994). Interspecies (toxicokinetic) variations for boron relate primarily to clearance. The ratio of mean clearance values in non-pregnant rats versus non-pregnant humans for boron (based on all of the data considered suitable for inclusion) is 3-4. In view of the lack of adequate kinetic studies in rats and hence less than optimum confidence in much of the data that serve as the basis for the ratio, replacement of the default for the toxicokinetic component of the interspecies factor is considered premature at this time. The total uncertainty factor for interspecies variation is 10. Intraspecies variation (toxicokinetics) for boron relates also primarily to variations in clearance. As the critical effect that serves as the basis for the TDI is developmental, pregnant women are the subgroup of interest in this regard. Based on pooled individual data from available studies, the mean glomerular filtration rate (GFR) in 36 healthy women was 145 ± 23 ml/minute in early pregnancy and 144 ± 32 ml/minute in late pregnancy. The standard deviation represented 22% of the mean value in late pregnancy. Based on division of the mean GFR (144 ml!minute) by the GFR at two standard deviations below the mean (80 ml/minute) to address variability for approximately 95% of the population, the ratio for the toxicokinetic component of interspecies variation is 1.8 (compared with the default value for this component of 3.2). As there are no data to serve as a basis for replacement of the default value for the toxicodynamic component of the uncertainty factor for intraspecies variation, the total uncertainty factor for intraspecies variation is 1.8 x 3.2 = 5.7 (rounded to 6). 3 Using an uncertainty factor of 60, the TDI is therefore 0.16 mg/kg of body weight. With an allocation of 10% of the TDI to drinking-water and assuming a 60-kg adult consuming 2 litres of drinking-water per day, the guideline value is 0.5 mg/litre (rounded figure). Conventional water treatment (coagulation, sedimentation, filtration) does not significantly remove boron, and special methods would have to be installed in order to remove boron from waters with high boron concentrations. Ion exchange and reverse osmosis processes may enable substantial reduction but are likely to be
Report of mformal discussions to develop recommendations for the WHO Guidelmes for drinkmg-water quality- Boron. Cincinnati, OH, 28-29 September 1997. Report available from WHO, Division of Operational Support in Environmental Health, Geneva. BORON 27 prohibitively expensive. Blending with low-boron supplies might be the only economical method to reduce boron concentrations in waters where these concentrations are high (WRc, 1997). The guideline value of 0.5 mg/litre is designated as provisional, because it will be difficult to achieve in areas with high natural boron levels with the treatment technology available.
8. REFERENCES
Anderson DL, Cunningham WC, Lindstrom TR (1994) Concentrations and intakes of H, B, S, K, Na, Cl and NaCI in foods. Journal offood composlfton and analys1s, 7:59-82. Ben son WH, Birge W J, Do rough HW ( 1984) Absence of mutagenic activity of sodium borate (borax) and boric acid in the Salmonella preincubation test. Environmental tox1cology and chemistry, 3:209-214. Budavari S et al., eds. (1989) The Merck index, 11th ed. Rahway, NJ, Merck and Co., Inc. Butterwick L, de Oude N, Raymond K (1989) Safety assessment of boron in aquatic and terrestrial environments. Ecotox1cology and enVIronmental safety, 17:339-371. Cotton PA, Wilkinson L (1988) Advanced morgamc chemistry, 5th ed. New York, NY, John Wiley & Sons, pp. 162-165.
Dieter MP (1994) Toxicity and carcinogenicity studies of boric acid in male and female B6C3F 1 mice. Environmental health perspectives, 102 (Suppl. 7):93-97. Goldbloom RB, Goldbloom A (1953) Boric acid poisoning: Report of four cases and a review of 190 cases from the world literature. Journal ofpediatrics, 43:631--643. Green GH, Lott MD, Weeth HJ (1973) Effects of boron water on rats. Proceedings, Western Section, Amencan Society of Animal Sc1ence, 24:254-258. Haberer K (1996) [Boron in drinking water in Germany.] Wasser-Abwasser, 137:364-371 (in German). Heindel JJ et al (1992) Developmental toxicity of boric acid in mice and rats. Fundamental and applied toxicology, 18:266-277. Huber L (1982) ICP-AES, ein neues verfahren zur multielement- bestimmung in wassern, abwassern und schlammen. Wasser, 58: 173-185B. Hunt CD (1989) Dietary boron modified the effects of magnesium and molybdenum on mineral metabolism in the cholecalciferol-deficient chick. Biological trace element research, 22:201-220. ISO (1990) Water quality- Determination of borate- Spectrometric method using azomethine-H Geneva, International Organization for Standardization (ISO 9390: 1990). ISO (1996) Water quality- Determmation of 33 elements by inductively coupled plasma atom1c emission spectroscopy. Geneva, International Organization for Standardization (ISO 11885:1996 (E)). Iyengar GV et al. (!988) Lithium in biological and dietary materials. In. Proceedmgs of the 5th internatwnal workshop on trace element analytical chemistry in medicme and biology, pp. 267-269. Jansen JA, Andersen J, Schou JS (1984) Boric acid single dose pharmacokinetics after intravenous administration to man. Arch1ves of toxicology, 55:64-67. Keren R, Mezuman U (1981) Boron adsorption by clay minerals using a phenomenological equation. Clays and clay minerals, 29:198-204. Keren R, Gast RG, Bar-YosefB (1981) pH-dependent boron adsorption by Na-montmorillonite. Soil Science Society ofAmencajournal, 45:45-48. Ku WW et al. (1993) Testicular toxicity of boric acid (BA): relationship of dose to lesion development and recovery in the F344 rat. Reproductive toxicology, 7:305-319. Landolph JR (1985) Cytotoxicity and negligible genotoxicity of borax and borax ores to cultured mammalian cells. Amencanjournal ofindustnal med1cme, 7:31-43. 28 INORGANIC CONSTITUENTS
Larsen LA ( 1988) Boron. In. Seiler HG, Sigel H, eds. Handbook on tox1city of morgamc compounds. New York, NY, Marcel Dekker, pp. 129-141. Lee lP, Sherins RJ, Dtxon RL (1978) Evidence for mduction of germinal aplasta in male rats by environmental exposure to boron Tox1cology and applied pharmacology, 45.577-590. Linden CH et al. (1986) Acute ingestion of boric acid. Journal of tox1cology. Cluucal toxicology, 24•269-279 Linder RE, Strader LF. Rehnberg GL (1990) Effect of acute exposure to boric acid on the male reproductive system of the rat Journal of toxicology and environmental health, 31:133-146. Litovitz TL et al (1988) Clinical manifestation of toxicity in a series of 784 bone acid ingestions. American JOurnal of emergency med1cme, 31 :209-213 MAFF (1991) Household food consumptwn and expenditure 1991. Annual report of the Nationul Food Survey Comm11tee London, Ministry of Agriculture, Fisheries and Food. Mance G, O'Donnell AR, Smith PR (1988) Proposed environmental qualify standards for List If substances in water Boron Medmenham, Water Research Centre (Report TR 256, March 1988; ISBN 0-902-15663-2) Nielsen FH (1989) Dietary boron affects variables associated with copper metabolism in humans In: Aulse M et al, eds Proceed1ngs of the lOth mternational trace element symposium Vol. 4 Jena, Friedrich-Schtller-Universitiit, pp. 1106-1111. Nielsen FH (1994) Biochemtcal and physiological consequences of boron deprivation in humans. Environmental health perspectives, I 02 (Suppl. 7):59-63 Nielsen FH et al. (1987) Effect of dietary boron on mmeral, oestrogen, and testosterone metabolism in postmenopausal women. The FASEB journal, 1.394-397 Nielsen FH, Mullen LM, Gallagher SK (1990) Effect of boron depletion and repletion on blood indicators of calcium status in humans fed a magnesium-low diet. Journal of trace elements m experimental medicme, 3:45-54. Nielsen FH, Mullen LM, Nielsen EJ (1991) Dietary boron affects blood cell counts and hemoglobin concentrations in humans. Journal of trace elements in expenmentalmed1cine, 4:211-223. Nielsen FH et al. (1992) Boron enhances and mimics some effects of oestrogen therapy in postmenopausal women. Journal of trace elements 111 experimentalmed1cine, 5:237-246. NTP (1987) Toxicology and carcmogenesis stud1es of boric ac1d (CAS no 10043-35-3) in B6C3F1 m1ce (food studies) Research Triangle Park, NC, US Department of Health and Human Services, Public Health Service, National Institutes of Health, National Toxicology Program (NTP Technical Report Series No. 324). NTP (1990) Fmal report on the developmental tox1c1ty of bone ac1d (CAS no 10043-35-3) in Sprague-Dawley rats Research Triangle Park, NC, US Department of Health and Human Services, Public Health Service, National Institutes of Health, NatiOnal Toxicology Program (NTP Report No. 90-1 05). O'Sullivan K, Taylor M (1983) Chronic boric acid poisoning in infants. Arch1ves of diseases in chzldhood, 58:737-739. Penland JG (1994) Dietary boron, brain function and cognitive performance. Environmental health perspectives, 102 (Suppl. 7):65-72. Pfeiffer CC, Hallman LF, Gersh I (1945) Bone actd ointment: A study of possible intoxication in the treatment of burns. Journal of the American Medical Assoczation, 128:266-274. Price CJ et al. (1996a) Developmental toxicity NOAEL and postnatal recovery in rats fed boric acid during gestation. Fundamental and applied tox1cology, 32:179-193. Price CJ et al. (1996b) The developmental toxicity of boric acid in rabbits. Fundamental and app/zed toxicology, 34:176-187. RaiD et al. (1986) Chem1cal attenuation rates, coefficients, and constants inleachate migratwn. Vol. 1· A crit1cal review Report to Electric Power Research Institute, Palo Alto, CA, by Battelle Pacific Northwest Laboratories, Richland, W A (Research Project 2198-1 ). Schou JS, Jansen JA, Aggerbeck B (1984) Human pharmacokinetics and safety of boric acid. Archives of toxicology, Suppl. 7:232-235. Smith FG et al. ( 1991) Measurement of boron concentration and isotope ratios in biological samples by inductively coupled plasma mass spectrometry with direct injection nebulization. Analyt1ca Chunzca Acta, 248:229-234. BORON 29
Sprague R W (1972) The ecological Significance of boron. Anaheim, CA, US Borax Research Corporation, 58 pp. Stokinger HE (1981) Boron. In: Clayton GD, Clayton FE, eds. Patty's mdustrial hyg1ene and toxicology. Vol 2B Toxicology, 3rd ed. New York, NY, John Wiley & Sons, pp. 2978-3005. Treinen KA, Chapin RE (1991) Development of testicular lesions in F344 rats after treatment with boric acid. Tox1cology and applied pharmacology, 107:325-335. Truhaut R, Phu-Lich N, Loisillier F (1964) [Effects of the repeated ingestion of small doses of boron derivatives on the reproductive functions of the rat.] Comptes Rendus de I'Academie des Scrences, 258:5099-5102 (in French with English abstract). Verbitskaya GV (1975) [Experimental and field investigations concerning the hygienic evaluation of boron-containing drinking water.] G1giena i Sanitanya, 7:49-53 (in Russian with English abstract). Waggott A (1969) An mvestigation of the potential problem of increasing boron concentrations m rivers and water courses. Water research, 3:749-765 Weast RC, Astle MJ, Beyer WH, eds. (1985) CRC handbook of chem1stry and physics, 69th ed. Boca Ratan, FL, CRC Press, Inc., pp. B-77, B-129. Weir RJ, Fisher RS (1972) Toxicologic studies on borax and boric acid. Toxicology and applied pharmacology, 23:351-364. WHO (1994) Assessmg human health nsks of chemicals derivatwn of gurdance values for health based exposure ltmlls. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria 170). WHO (in press) Boron. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria monograph). WRc ( 1997) Treatment technology for alumrmum, boron and uranium Document prepared for WHO by the Water Research Centre, Medmenham, and reviewed by S. Clark, US EPA; A. van Dijk-Looijaard, KIW A, Netherlands; and D. Green, Health Canada.
COPPER
First draft prepared by J.M. Donohue Health and Ecological Criteria Division United States Environmental Protection Agency, Washington, DC, USA
In view of uncertainties regarding copper toxicity in humans, a provisional guideline value for copper of 2 mg/litre was established in the 1993 WHO Guidelines for drinking-water quality, based on a 10% allocation of the 1982 JECFA PMTDI to drinking-water. The PMTDI was established as I 0 times the normal copper intake of 0.05 mg/kg of body weight per day, which was considered to be without adverse effect. 1 Copper was selected for re-evaluation by the Coordinating Committee for the updating of the Guidelines because of the provisional guideline value and because the development of an IPCS Environmental Health Criteria monograph for copper was under way.
1. GENERAL DESCRIPTION
1.1 Identity
Copper (CAS no. 7440-5~8) is a transition metal that is stable in its metallic state and forms monovalent (cuprous) and divalent (cupric) cations. Common copper compounds include the following:
Compound CAS no.
Copper(II) acetate monohydrate [Cu(C2H30 2) 2·H20] 6046-93-1 Copper(II) chloride [CuCI2] 7447-39-4 Copper(II) nitrate trihydrate [Cu(N03)2·3H20] 10031-43-3 Copper(II) oxide [CuO] 1317-38-0
Copper(II) sulfate pentahydrate [CuS04·5HzO] 7758-99-8
1.2 Physicochemical properties (Weast, 1983; ATSDR, 1990; Lewis, 1993) Compound Density Water solubility (g/cm3) (gllitre) Copper(II) acetate monohydrate 1.88 72 Copper(II) chloride 3.39 706 Copper(II) nitrate trihydrate 2.32 1378 Copper(II) oxide 6.32 Insoluble Copper(II) sulfate pentahydrate 2.28 316
It was erroneously assumed in the 1993 Guidelines that the PMTDI established by JECFA was based on a study in dogs. 32 INORGANIC CONSTITUENTS
I.3 Organoleptic properties
Dissolved copper imparts a light blue or blue-green colour and an unpleasant, metallic, bitter taste to drinking-water. The taste threshold is between 1 and 5 mg/litre and is influenced by the presence of other solutes (ATSDR, 1990; Olivares & Uauy, 1996a). Blue to green staining of porcelain sinks and plumbing fixtures occurs from copper dissolved in tap-water.
I.4 Major uses
Metallic copper is malleable, ductile, and a good thermal and electrical conductor. It has many commercial uses because of its versatility. Copper is used to make electrical wiring, pipes, valves, fittings, coins, cooking utensils, and building materials. It is present in munitions, alloys, and coatings. Copper compounds are used as or in fungicides, algicides, insecticides, wood preservatives, electroplating, azo dye manufacture, engraving, lithography, petroleum refining, and pyrotechnics. Fertilizers, animal feeds, and pharmaceuticals can contain copper compounds (ATSDR, 1990; Lewis, 1993). Copper compounds are also used as food additives (e.g. nutrient and/or colouring agent) (F AO/IPCS, 1994). Copper sulfate pentahydrate is sometimes added to surface water for the control of algae (NSF, 1996). Copper sulfate was once prescribed as an emetic, but this use has been discontinued owing to adverse health effects (Ellenhorn & Barceloux, 1988).
1.5 Environmentalfate
The fate of elemental copper in water is complex and influenced by pH, dissolved oxygen, and the presence of oxidizing agents and complexing compounds or ions (US EP A, 1995). Surface oxidation of copper produces copper(!) oxide or hydroxide. In most instances, copper(!) ion is subsequently oxidized to copper(II) ion. However, copper(!) ammonium and copper(!) chloride complexes, when they form, are stable in aqueous solution. In pure water, the copper(II) ion is the more common oxidation state (US EPA, 1995). Copper(II) ions will form complexes with hydroxide and carbonate ions. The formation of insoluble malachite [CuiOH)2C03] is a major factor in controlling the level of free copper(II) ion in aqueous solutions. Copper(II) ion is the major species in water at pHs up to 6; at pH 6-9.3, aqueous CuC03 is prevalent; and at pH 9.3-10.7, the aqueous [Cu(C03) 2f- ion predominates (Stumm & Morgan, 1981 ). In one copper-polluted, organic-depleted lake, soluble copper speciated in the 2 order CuOH+ > Cu + > CuC03 based on equilibrium calculations (Lopez & Lee, 1977). Dissolved copper ions are removed from solution by sorption to clays, minerals, and organic solids or by precipitation (Callahan et al., 1979). Copper has been reported to strongly adsorb to clay materials in a pH-dependent fashion; such adsorption is increased by the presence of particulate organic materials (Payne & Pickering, 1975; Huang et al., 1977; Brown et al., 1983). Copper discharged to wastewater is concentrated in sludge during treatment. Sorption can be reversed in a reducing or acidic environment; thus, sediments and sludge can act as a reservoir for COPPER 33 copper ions (Lopez & Lee, 1977). Copper ions form chelates with humic acids and polyvalent organic anions (Cotton & Wilkinson, 1980). The presence of chelating agents increases the solubility of copper in an aqueous medium.
2. ANALYTICAL METHODS
The most important analytical methods for the detection of copper in water are atomic absorption spectrometry (AAS) with flame detection, graphite furnace atomic absorption spectroscopy, inductively coupled plasma atomic emission spectroscopy, inductively coupled plasma mass spectrometry (ICP-MS), and stabilized temperature platform graphite furnace atomic absorption (ISO, 1986, 1996; ASTM, 1992, 1994; US EPA, 1994). The ICP-MS technique has the lowest detection limit (0.02 ~Lg/Jitre) and the AAS technique the highest (10 J..Lgllitre). Detection limits for the other three techniques range from 0.7 to 3 J..Lg/litre. Measurement of dissolved copper requires sample filtration; results from unfiltered samples include dissolved and particulate copper.
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Copper is present in the atmosphere from wind dispersion of particulate geological materials and particulate matter from smokestack emissions. In a nationwide study by the US EPA for the years 1977-1983, the range of copper concentrations in 23 814 air samples was 0.003-7.32 J..Lg/m3 (US EPA, 1987).2 3 Median values for different cities and years ranged from 0.004 to 1.79 J..Lglm , and 3 mean values ranged from 0.0043 to 1.96 J..Lg/m . Concentrations of copper determined in over 3800 samples of ambient air at 29 sites in Canada over the period 3 3 1984-1993 averaged 0.014 J..Lg/m • The maximum value was 0.418 J..Lg/m (WHO, in press). Atmospheric copper is removed by gravitational settling, dry disposition, rain, and snow.
3.2 Water
Because copper is a naturally occurring element, it is found ubiquitously in surface water, groundwater, seawater, and drinking-water (ATSDR, 1990; US EPA, 1991 ). In a 1969 survey of 678 ground water supplies in the USA, the maximum reported copper concentration was 0.47 mg/litre, whereas the mean concentration in samples exceeding the detection limit (0.010 mg/litre) was 0.075 mg/Jitre (US EPA, 1991 ). Comparative values for 109 surface water supplies were 0.304 and 0.066 mg/litre, respectively. Copper levels in surface water samples were reported to range
Additional source Computer printout of frequency distribution hstmg of copper in air, 1977-1983. Research Triangle Park, NC, US Environmental Protection Agency. Environmental Monitoring Systems Laboratory 34 INORGANIC CONSTITUENTS from 0.0005 to I mg/litre, with a median of 0.0 I mg/litre (ATSDR, 1990). In the United Kingdom, the mean copper concentration in the River Stour was 0.006 mg/litre (range 0.003-0.019 mg/litre). Background levels were 0.001 mg/litre, derived from an upper catchment control site. Fourfold increases in copper concentrations were apparent downstream of a sewage treatment plant. In an unpolluted zone of the River Periyar in India, copper concentrations ranged from 0.0008 to 0.010 mg/litre (WHO, in press). Copper concentrations in drinking-water vary widely as a result of variations in pH, hardness, and copper availability in the distribution system. A number of studies indicate that copper levels in drinking-water samples can range from :<::::0.005 to 18 mg/litre, with the primary source most often being the corrosion of interior copper plumbing (ATSDR. 1990; US EPA, 1991). Levels of copper in running or fully flushed water tend to be low, whereas those of standing or partially flushed water samples are more variable and can be substantially higher (frequently ~1 mg/litre) (ATSDR, 1990). In the Netherlands, copper concentrations between 0.2 and 3.8 mg/litre were reported in water standing for 16 hours. Average copper levels in water from municipalities were between 0.04 and 0.69 mg/litre. In two cities in Sweden, the mean standing water copper level was 0.7 mg/litre, with a 90th percentile of 2.1 mg/litre; in water for consumption, the mean copper concentration was 0.6 mg/litre, with a 90th percentile of 1.6 mg/litre. In distributed water from 70 municipalities across Canada, mean concentrations of copper ranged from 0.02 to 0.075 mg/litre; maximum values ranged up to 0.56 mg/litre. In about 20% of the distributed water supplies, the level of copper was significantly higher than in the corresponding treated water samples. Furthermore, the increase was higher in those areas where the water was soft and corrosive (WHO, in press). In another survey in Canada, the median copper concentration in distributed water was 0.27 mg/litre (range >0.01-0.9 mg/litre) for 27 supplies with acid or neutral pH (Health Canada, 1992). Some of the copper in drinking-water may derive from treatment of surface water sources with copper sulfate algicides (US EPA, 1991 ).
3.3 Food
Food is a principal source of copper exposure for humans. Liver and other organ meats, seafood, nuts, and seeds are good sources of dietary copper (NAS, 1989). Vitamin/mineral preparations for children and adults generally contain 2 mg of copper as copper oxide. Infant formula contains 0.6-2 1-lg of copper per kcal (Olivares & Uauy, 1996b). Based on data collected during the US Food and Drug Administration's Total Diet Study (1982-1986), the average dietary intake of copper for adult males was 1.2 mg/day, whereas that for adult females was 0.9 mg/day. The average intake for infants (6 months to 1 year) was 0.45 mg/day, and that for 2-year-olds was 0.57 mg/day (Pennington et al., 1989). In Scandinavian countries, intakes are in the range of 1.0-2.0 mg/day for adults, 2 mg/day for lactovegetarians, and 3.5 mg/day for vegans (Pettersson & Sandstrum, 1995; WHO, in press). The United Kingdom reported intakes of 1.2 and 1.6 mg/day for adult females and males, respectively, and 0.5 mg/day for children COPPER 35
1V;, to 4Yz years old. Australia reported intakes of 2.2 and 1.9 mg/day for adult females and males, respectively, and 0.8 mg/day for the 2-year-old. In Germany, the dietary intake of adults was 0.95 mg/day (WHO, in press). Copper is an essential nutrient. In the USA, the estimated safe and adequate dietary intake for copper is 1.5-3 mg/day for adults, 0.4-0.6 mg/day for infants, and 0.7-2 mg/day for children; although average intakes for adults appear to be suboptimum, there is little evidence of copper deficiency in the population (NAS, 1989). Estimates of average copper requirements are 12.5 j.J.g/kg of body weight per day for adults and about 50 j.J.g/kg of body weight per day for infants (WHO, 1996).
3.4 Estimated total exposure and relative contribution of drinking-water
Food and water are the primary sources of copper exposure in developed countries. In general, dietary copper intakes for adults range from 1 to 2 mg/day (WHO, in press); use of a vitamin/mineral supplement will increase exposure by 2 mg/day. Drinking-water contributes 0.1-1 mg/day in most situations. Thus, daily copper intakes for adults usually range from 1 to 3 mg/day. Consumption of standing or partially flushed water from a system that contains copper pipes or fittings can considerably increase total daily copper exposure, especially for infants fed formula reconstituted with tap-water.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
Absorption of orally administered copper in mammals occurs in the upper gastrointestinal tract and is controlled by a complex homeostatic process involving active and passive transport, plus binding sites on metallothionein and other proteins in the mucosal cells (Linder & Hazegh-Azam, 1996). Mucosal and serosal transport mechanisms differ. Movement of copper across the mucosa occurs by diffusion or facilitated transport. The presence of competing metals, dietary proteins, anions, fructose, and ascorbic acid influences copper uptake from the gastrointestinal tract (Lonnerdal. 1996). Using an in vivo intestinal perfusion system in rats, it was demonstrated that excess concentrations of cations having an electronic configuration similar to that of copper [i.e. iron(ll), zinc, tin, and cobalt] in the presence of a high-affinity ligand such as L-histidine can significantly inhibit the intestinal absorption and/or retention of copper (Wapnir et al., 1993). Within the mucosal cells, 80% of the copper is found in the cytosol bound to proteins, especially metallothionein. Transport from the mucosal cells is mediated by a p-type ATPase active transport system. Serosal transport of copper is inhibited in Menke' s syndrome, one of several genetic diseases related to copper. Copper in the portal blood is bound to albumin or transcuprin; a small amount may be chelated by peptides and amino acids (Linder & Hazegh-Azam, 1996). Copper uptake from the blood by the liver and distribution within the liver are not well understood. They are presumed to involve a transport process that differs from that in the intestines (Linder & Hazegh-Azam, 1996). Within the liver, copper becomes incorporated in ceruloplasmin, as well as the enzymes superoxide 36 INORGANIC CONSTITUENTS dismutase and cytochrome oxidase. Excess copper is bound to hepatic metallothionein. Ceruloplasmin is the primary copper transport protein in systemic circulation and contains about 75% of the plasma copper (Luza & Speisky, 1996). Ceruloplasmin carries copper to the cells for uptake. It also has enzymatic activity as a ferrooxidase and functions in the synthesis of haemoglobin. In humans, the highest concentrations of copper are found in the liver, brain, heart, kidney. and adrenal gland; moderate levels are found in the intestine, lung, and spleen; and low concentrations occur in endocrine glands, bone, muscle, larynx, trachea, aorta, and testes (Schroeder et al., 1966; Evans, 1973). Approximately 50% of the body's copper is found in muscle and bone tissue, whereas about 10% is stored in the liver (Evans, 1973; Luza & Speisky, 1996). The liver of newborn infants contains about 10 times the copper of the adult liver and accounts for 50--60% of the total body copper (Luza & Speisky, 1996). Copper is an essential nutrient and is required for the proper functioning of many important enzyme systems. Copper-containing enzymes include ceruloplasmin, superoxide dismutase, cytochrome oxidase. tyrosinase, monoamine oxidase, lysyl oxidase, and phenylalanine hydroxylase (Linder & Hazegh-Azam, 1996). The activity of the enzyme superoxide dismutase can be used in the assessment of copper status (Olivares & Uauy, 1996b). Ceruloplasmin and serum copper concentrations are also employed as indicators cif copper status. Copper is excreted from the body in bile, faeces, sweat, hair, menses, and urine (Gollan & Deller, 1973; Luza & Speisky, 1996). In humans, the major excretory pathway for absorbed copper is bile, where copper is bound to both low-molecular weight and macromolecular species. Biliary copper travels to the intestine; after minimal reabsorption, it is eliminated in the faeces. In normal humans, <3% of the daily copper intake is excreted in the urine (Luza & Speisky, 1996). Urinary copper may originate from amino acid-bound metal or from dissociated copper-albumin complexes; erythrocyte or ceruloplasmin-bound copper generally will not permeate the glomerulus. thus preventing its excretion (Evans, 1973). There are several genetic disorders that affect copper utilization. The genetic abnormalities associated with Menke syndrome, Wilson's disease, and aceruloplasminaemia are fairly well understood. There is some evidence to suggest a genetic basis for Indian childhood cirrhosis and idiopathic copper toxicosis (Muller et al., 1996; Olivares & Uauy, l996a). In the case ofMenke syndrome and Wilson's disease, the atTected genes have been identified. Data suggest that the predisposing genetic component for idiopathic copper toxicosis may be an autosomal recessive gene (Muller et al., 1996). Individuals with a glucose-6-phosphate dehydrogenase deficiency disorder have an increased susceptibility to copper-induced oxidation reactions and methaemoglobin formation (Moore & Calabrese, 1980), but there is no direct effect of the disorder on copper uptake, distribution, or metabolism. In Menke syndrome, there is minimal copper absorption from the intestines, leading to death during early childhood (Harris & Gitlin, 1996; Olivares & Uauy, 1996a). Children suffering from Menke syndrome (an X chromosome-linked disorder) exhibit mental deterioration, failure to thrive, hypothermia, and connective tissue abnormalities (Harris & Gitlin, 1996 ). The defective protein is a p-type ATPase responsible for intestinal transport of copper; a child with Menke syndrome suffers from a profound COPPER 37 copper deficiency, despite adequate dietary copper. There is no effective treatment for Menke syndrome, although administration of copper as the dihistidine complex delays the development of symptoms (Linder & Hazegh-Azam, 1996). Wilson's disease is an autosomal recessive disorder that affects the hepatic intracellular transport of copper and its subsequent inclusion into ceruloplasmin and bile. A p-type ATPase is again affected, but the enzyme is different from that affected in Menke syndrome. Because copper is not incorporated into ceruloplasmin, its normal systemic distribution is impaired, and copper accumulates in the liver, brain, and eyes (Pennington et al., 1989; Harris & Gitlin, 1996). Wilson disease generally appears in late childhood and is accompanied by hepatic cirrhosis, neurologic degeneration, and copper deposits in the cornea of the eye (Kayser Fleischer rings). Patients with Wilson disease are treated with chelating agents, such as penicillamine, to promote copper excretion (Yarze et al., 1992). Patients that follow their therapeutic regime can expect to live a normal life (Scheinberg & Sternleib, 1996). Restriction of dietary copper alone cannot influence the progression of the disease. Aceruloplasminaemia is an autosomal recessive disorder caused by changes in the ceruloplasmin gene that affect the ability of cs:ruloplasmin to bind copper (Harris & Gitlin, 1996). The symptoms of aceruloplasminaemia do not become apparent until adulthood. They include dementia, diabetes, retinal degeneration, and increased tissue iron stores. The etiologies of Indian childhood cirrhosis and idiopathic copper toxicosis are complex and may involve a combination of genetic, developmental, and environmental factors (Muller et al., 1996; Pandit & Bhave, 1996). Both disorders are characterized by liver enlargement, elevated copper deposits in liver cells, pericellular fibrosis, and necrosis. Both disorders are generally fatal. Exposure to elevated levels of copper in milk or infant formula prepared in copper-containing vessels or with water containing elevated copper concentrations is believed to contribute to the hepatic copper overload. Poor biliary excretion of copper may also play a role in the etiology of the disease.
5. EFFECTS ON LABORATORY ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
Acute responses to copper vary with species and copper compound. Sheep and dogs are more sensitive to copper than rodents, pigs, and poultry (Linder & Hazegh-Azam, 1996). Soluble copper salts are more toxic than insoluble compounds. In a classic study, Wang & Borison (1951) evaluated the acute emetic response of 107 mongrel dogs to a single dose of copper sulfate pentahydrate in aqueous solution. In this group, 20 (19%) responded to 20 mg of copper sulfate [5 mg of copper(II)] with a mean response latency of 19 ± 11 minutes. Ninety-one dogs (85%) responded to 40 mg of copper sulfate [10 mg of copper(II)] with a response latency of 16 ± 9 minutes, and all animals responded to an 80-mg dose of 38 INORGANIC CONSTITUENTS
copper sulfate [20 mg of copper(II)] (latency 19 ± 7 minutes). Some of the animals were then subjected to a vagotomy, sympathectomy, or vagotomy and sympathectomy. After severing of the neural pathways, the dogs were again exposed to copper sulfate. The acute dose required to elicit the emetic response increased in the vagotomized and in the vagotomized/sympathectomized dogs, as did the response latency time. The greatest effects on response threshold and latency time were seen in the vagotomizedlsympathectomized dogs. The authors postulated that the emetic action of copper sulfate in dogs is biphasic. The initial rapid response event is caused by the action of copper sulfate on the peripheral nervous system and is the most sensitive response. The secondary response is elicited by the central nervous system and is responsive to absorbed copper. The requirement of the secondary response for absorption increases its latency time. More recent studies in dogs and ferrets (Bhandari & Andrews, 1991; Makale & King, 1992; Fukui et al., 1994) confirm the importance of gastrointestinal neural pathways and receptors in copper sulfate-induced emesis. In one study, copper sulfate solution was infused into the stomach and duodenum of groups of four or five ferrets with ligated pyloric sphincters (Makale & King, 1992). In one group (four ferrets), the stomach infusion preceded the duodenal infusion; in the other group (five ferrets), the duodenal infusion preceded the stomach infusion. Infusion to the stomach resulted in vomiting in seven of nine ferrets with a mean latency of 4.4 minutes. Infusion to the duodenum resulted in vomiting in one of nine animals. The authors concluded that the primary site of the emetic response to copper sulfate is in the stomach of the ferret.
5.2 Short-term exposure
Several short-term studies of copper toxicity have been conducted in rats and mice. Effects were largely the same in both species, but rats were slightly more sensitive than mice (Hebert et al., 1993). Groups of five male and five female F 344/N rats were administered copper sulfate in their drinking-water for 2 weeks at estimated doses up to 97 mg of copper per kg of body weight per day (Hebert et al., 1993). A LOAEL of 10 mg of copper per kg of body weight per day observed in male rats was based on an increase in the size and number of protein droplets in the epithelial cells of the proximal convoluted tubules of the males. No renal effects were seen in the females receiving the same dose. There was no NOAEL for males in this study; the NOAEL for females was 26 mg of copper per kg of body weight per day. When copper (as copper sulfate) was administered by gavage to 10 male albino rats at a dose of 25 mg/kg of body weight for 20 days, the kidneys displayed necrosis and tubular engorgement. Centrilobular necrosis, perilobular sclerosis, and periportal copper disposition were seen in the liver; haemoglobin and haematocrit values were decreased (Rana & Kumar, 1980). Copper was less toxic to rats when administered in the diet than when administered in drinking-water or by gavage. This was true for 2-week and 13-week exposures. A dietary concentration of 1000 mg/kg of feed (estimated doses of 23 mg/kg of body weight per day for 2 weeks and 16 mg/kg of body weight per day COPPER 39 for 13 weeks) had no adverse effects in male or female F 344/N rats (Hebert et al., 1993 ). Dietary concentrations of 2000 mg/kg of feed ( 44 mg/kg of body weight per day for 2 weeks and 33 mg/kg of body weight per day for 13 weeks) were associated with hyperplasia and hyperkeratosis of the squamous epithelium of the limiting ridge of the rat forestomach. With the 13-week exposure and the 2000 mg/kg of feed (33 mg/kg of body weight per day) dietary concentration, protein droplets were present in the kidneys, and liver inflammation was noted (Hebert et al., 1993). Changes in liver and kidney histopathology were dose-related; males were affected more than females. Staining of the kidney cells for a 2 ~ globulin was negative. Dose-related decreases in haematological parameters at 4000 mg/kg of feed (66 mg/kg of body weight per day) and 8000 mg/kg of feed (134 mg/kg of body weight per day) were indicative of a microcytic anaemia, whereas increases in serum enzymes were indicative of liver damage. There was some evidence of significant biochemical effects with the 2000 mg/kg of feed (33 mg/kg of body weight per day) exposure concentration and definitive effects at the 4000 mg/kg of feed (66 mg/kg of body weight per day) and 8000 mg/kg of feed (134 mg/kg of body weight per day) concentrations. Iron stores in the spleen were depleted, especially for the highest exposure concentration.
5.3 Long-term exposure
Male weanling Wistar rats (four per group) were given either a normal diet containing 10-20 mg of copper per kg of feed (controls) or diets supplemented with 3000, 4000, or 5000 mg of copper per kg of feed for 15 weeks (Haywood & Loughran, 1985). The animals receiving 3000 mg of copper per kg of feed were then allowed to continue the experimental regime for the remainder of the year. Assuming that rats consume 5% of their body weight per day in food, these dietary copper concentrations would correspond to approximate doses of 0.5-1.0, 150, 200, and 250 mg of copper per kg of body weight per day. All copper-supplemented groups exhibited reductions in body-weight gains relative to the control group that persisted until the end of the 15-week exposure period. For the 3000, 4000, and 5000 mg/kg of feed groups, copper concentrations in the liver peaked at 3-4 weeks, declined significantly by 6 weeks, but were still elevated at 15 weeks. Although the timing and duration varied somewhat, all supplemented groups evidenced hepatocellular necrosis during weeks 1-6, followed by a regeneration process that began after 3-5 weeks. The adaptation process noted during the latter part of the first 15 weeks of exposure continued during the 3000 mg/kg of feed group extension period. The average body weight recovered to 80% of that of the control group, and the copper concentration in the liver dropped from 1303 J.lg/g at 15 weeks to 440 J.lg/g at 52 weeks. However, even at 52 weeks, hepatic copper was greater in the exposed animals than in the controls (23 J.lg/g).
5.4 Reproductive and developmental toxicity
Sperm morphology and motility analyses, testis and epididymis weight determination, and estrous cycle characterization were performed in rats and mice as 40 INORGANIC CONSTITUENTS part of a subchronic dietary study (Hebert et al., 1993). No significant differences from control values were found for any of the following reproductive parameters: testis, epididymis. and cauda epididymis weights, spermatid count, spermatid number per testis or per gram testis. spermatozoal motility and concentration, estrous cycle length, or relative length of time spent in the various estrous stages. A reproductive NOAEL of 4000 mg of copper sulfate pentahydrate per kg of diet (66 and 68 mg of copper per kg of body weight per day for male and female rats, respectively) was established for these parameters in this study. There is some evidence that copper is a developmental toxicant. Embryotoxicity and teratogenicity were reported in hamsters when dams were injected intraperitoneally with copper citrate (DiCarlo, 1980) or intravenously with copper sulfate or copper citrate (Ferm & Hanlon. 1974). Such effects were noted at single injection doses as low as 2.13 mg of copper per kg of body weight for copper sulfate and 0.25-1.5 mg of copper per kg of body weight for copper citrate. When mice (7-22 females per group) were fed diets supplemented with 0, 500, 1000. 1500, 2000. 3000, or 4000 mg of copper sulfate per kg of diet for 1 month, fetal mortality and decreased litter size were observed in the 2000-4000 mg/kg of diet groups. Various skeletal and soft-tissue malformations were seen in 2-8% of the surviving fetuses from the two highest dose groups (Lecyk, 1980). The low concentrations of supplemental copper (500 and 1000 mg/kg of diet) had a beneficial effect on development.
5.5 Mutagenicity and related end-points
Copper sulfate has been reported to induce reverse mutations in Escherichia coli (Demerec et al., 1951) but not in TA98, TA100. TA1535. or TA1537 strains of Salmonella typhimurium in the presence or absence of microsomal activation (Moriya et al., 1983: Wong, 1988). Negative microbial mutation results have been reported for copper sulfate and copper chloride in Saccharomyces cerevisiae (Singh, 1983) and Bacillus subtilis (Nishioka, 1975; Kanematsu et al., 1980; Matsui, 1980). Copper chloride or copper acetate concentrations of 20-150 mmol!litre induced errors in viral DNA synthesis from poly(c) templates (Sirover & Loeb, 1976), and copper sulfate concentrations of 1 mmol/litre, but not 0.03-0.3 mmol!litre, caused DNA strand breakage in rat hepatocytes (Sina et al., 1983). Results from in vitro tests are not transferable to the in vivo situation, where copper ions are generally bound to protein or amino acid ligands. Male Wistar rats were implanted with subcutaneous osmotic pumps that continuously administered saline. copper(II) chloride, the copper chelate cupric nitrilotriacetate (Cu-NT A), or NTA for 3 or 5 days. Copper was delivered at a rate of 4 mg/kg of body weight per day, which maintained serum copper levels 30-70% higher than in the untreated controls for copper(II) chloride or 100-120% higher than in controls for Cu-NTA. Hepatic and renal DNA levels of 8-hydroxy guanosine, a deoxyguanosine oxidation product associated with mutagenesis and carcinogenesis, were significantly elevated in the copper-exposed animals (Toyokuni & Sagripanti, 1994 ). Chromosomal aberrations and micronuclei were observed in the bone marrow of inbred Swiss mice exposed to copper (from copper sulfate) COPPER 41 concentrations of 0, 1.3, 2.6, or 5.1 mg/kg of body weight by intraperitoneal or subcutaneous injection (Bhunya & Pati, 1987).
5.6 Carcinogenicity
Copper and its salts do not appear to be animal carcinogens based on limited long-term exposure data. In one study, two strains of mice (18 per sex per group per strain) were exposed to copper hydroxyquinoline (181 mg of copper per kg of body weight per day) by gelatin capsule until they were 28 days old, whereupon the compound was administered for an additional 50 weeks in the feed at 2800 mg/kg of feed (506 mg of copper per kg of body weight per day). No significant increases in tumour incidence were observed in either sex or either strain (BRL, 1968).
6. EFFECTS ON HUMANS
6.1 Acute exposure
The acute lethal dose for adults lies between 4 and 400 mg of copper(II) ion per kg of body weight, based on data from accidental ingestion and suicide cases (Chuttani et al., 1965; Jantsch et al., 1984; Agarwal et al., 1993). Individuals ingesting large doses of copper present with gastrointestinal bleeding, haematuria, intravascular haemolysis, methaemoglobinaemia, hepatocellular toxicity, acute renal failure, and oliguria (Agarwal et al., 1993). At lower doses, copper ions can cause symptoms typical of food poisoning (headache, nausea, vomiting, diarrhoea). Records of published studies of gastrointestinal illness induced by copper from contaminated beverages plus public health department reports for 68 incidents indicate an acute onset of symptoms (Low et al., 1996). Symptoms generally appear after 15-60 minutes of exposure; nausea and vomiting are more common than diarrhoea. Copper concentrations were not available for all incidents. Among 24 outbreaks with quantitative data, the lowest copper concentrations were 3.5 and 3.8 mg/litre; background information on the samples analysed and analytical methods was limited. In 6 of the 24 cases (25%), the copper concentration was less than 10 mg/litre (Low et al., 1996). Data on the amount of beverage consumed were not available for any of these incidents. An analysis of data from the US Centers for Disease Control regarding 155 reported cases of copper intoxication from drinking-water sources during the years 1977-1982 and 1991-1994 indicated that reported levels of copper in drinking-water were in the range of4.0-156 mg/litre (US EPA, 1987; CDC, 1993, 1996).
6.2 Short-term exposure
Recurrent morning episodes of nausea, vomiting, and abdominal pain were reported by three of four family members exposed for over a year to drinking-water containing elevated levels of copper. Symptoms were associated with early-morning consumption of water, juice, or coffee. Copper concentrations measured on three occasions during the period covered by the complaint ranged from 2.8 to 42 INORGANIC CONSTITUENTS
7.8 mg/litre; only the 7.8 mg/litre sample was collected in the early morning (Spitalny et al., 1984). All samples were collected before the complaint was registered with the Department of Health. The median copper concentration for a series of samples collected after the complaint was 3.1 mg/litre, compared with a median value of 1.6 mg/litre for a series of samples collected from another house on the same service line but located closer to the beginning of the line. Recurrent episodes of vomiting, diarrhoea, and abdominal cramps were reported among individuals consuming water from homes with water that exceeded the US EP A maximum contaminant level goal of 1.3 mg/litre in four case-studies (Knobeloch et al., 1994). Data on water use practices and symptoms were collected using a questionnaire. The incidence of symptoms was positively correlated with the copper concentration in the water samples, ingestion of first-draw water, and water intake. There was a negative correlation with consumer age and use of bottled water. The presence of various confounding factors suggests discretion when drawing conclusions based on the data presented. A 26-year-old male presented with symptoms of cirrhosis, liver failure, and Wilson disease (Kayser-Fleischer rings) after more than 2 years of self-prescribed use of copper supplements (O'Donohue et al., 1993). The patient ingested 30 mg of supplemental copper per day for 2 years and 60 mg/day for a poorly defined period of up to a year. Liver damage was extensive, and a transplant was required. The diseased liver had an average copper concentration of 3230 flg/g dry weight (normal 20-50 flg/g); tissue histopathology was similar to that seen in Indian childhood cirrhosis and Wilson disease. The patient's family medical history and evaluation of his parents and sisters for copper excretion suggested that he did not carry the Wilson disease gene. Liver damage apparently resulted from the prolonged daily exposure to over 10 times the US estimated safe and adequate daily intake for copper.
6.3 Long-term exposure
Long-term intake of copper in the diet in the range of 1.5-3 mg/day has no apparent adverse effects. Daily intake of copper below this range can lead to anaemia, neutropenia, and bone demineralization in malnourished children (NAS, 1989). Adults are more resistant than children to the symptoms of a copper deficiency. No studies of adverse effects of long-term exposure of humans to copper at concentrations greater than those that occur in the diet were identified.
7. PROVISIONAL GUIDELINE VALUE
The IPCS Environmental Health Criteria monograph for copper (WHO, in press) concluded that:
The upper limit of the AROI [acceptable range of oral intake] in adults is uncertain but it is most likely in the range of several but not many mg per day 111 adults (several meaning more than 2 or 3 mg/day). This evaluation is based solely on studies of gastrointestinal effects of copper-contaminated drinking-water. A more specific value for the upper AROJ could not be COPPER 43
confirmed for any segment of the general population .... The available data on toxicity in animals were considered unhelpful in establishing the upper limit of the AROI, due to uncertainty about an appropriate model for humans.
A copper level of 2 mg/litre in drinking-water will be protective of adverse effects of copper and provides an adequate margin of safety. Owing to limitations of the epidemiological and clinical studies conducted to date, it is not possible to establish a clear effect level with any precision. Thus, it is recommended that this guideline value for copper of 2 mg/litre remain provisional as a result of uncertainties in the dose-response relationship between copper in drinking-water and acute gastrointestinal effects in humans. It is also noteworthy that copper is an essential element. It is stressed that the outcome of ongoing epidemiological studies in Chile, Sweden, and the USA may, upon publication, shed some light on more accurate quantification of effect levels for copper-induced toxicity in humans, including sensitive subpopulations. Staining of laundry and sanitary ware occurs at copper concentrations above 1 mg/litre. At levels above 5 mg/litre, copper also imparts a colour and an undesirable bitter taste to water.
8. REFERENCES
Agarwal SK, Tiwari SC, Dash SC (1993) Spectrum of poisoning requiring hemodialysis in a tertiary care hospital in India. International ;ourna/ ofartificial organs, 16(1):20-22. ASTM (I 992) Annual book of ASTM standards 11 0 I Philadelphia, PA, American Society for Testing and Materials. ASTM (I 994) Annual book of ASTM standards 11 02 Philadelphia, PA, American Society for Testing and Materials. ATSDR ( 1990) Toxicological profile for copper. Atlanta, GA, US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry (Subcontract No. A TSDR-88-0608-02). Bhandari P, Andrews PL (I 991) Preliminary evidence for the involvement of the putative 5-HT4 receptor in zacopride- and copper sulfate-mduced vomiting in the ferret. European journal of pharmacology, 204:273-280. Bhunya SP, Pati PC (1987) Genotoxicity of an inorganic pesticide, copper sulphate in a mouse in v1vo test system. Cyto/ogia, 52:801-808. BRL ( 1968) Evaluatwn of carcmogemc, teratogenic and mutagenic activities of selected pesllcides and industrial chem1cals. Vo/ I Carcinogenic study prepared by Bionetics Research Labs for the US National Cancer Institute (Publication No. MCI-DCCP-CG-1973-1-1 ). Brown KW, Thomas JC, Slowey JF (1983) The movement of metals applied to soils in sewage effluents. Water, air, and soil pollution, 19:43-54. Callahan MA et al. (1979) Water-related environmental fate of 129 priority pollutants. Vol I. Washington, DC, US Environmental Protection Agency, Office of Water Planning and Standards (Publication No. EPA 440/4-79-029a). CDC (1993) Surveillance for waterborne disease outbreaks- United States, 1991-1992. Morbidity and mortality weekly report, 42 (SS-5): 1-22. US Centers for Disease Control. CDC (1996) Surveillance for waterborne disease outbreaks- United States, 1993-1994. Morbidity and mortality weekly report, 45 (SS-I): 1-33. US Centers for Disease Control. Chuttani HK et al. (1965) Acute copper sulfate poisoning. American journal of medicine, 39:849-854. Cotton FA, Wilkinson G (1980) Advanced inorganic chemistry a comprehensive text. New York, NY, John Wiley & Sons, pp. 798-821. 44 INORGANIC CONSTITUENTS
Demerec M, Bertani G, Flint J (1951) A survey of chemicals for mutagenic action on E coli. American nawraltst, 85:119. DiCarlo FJ Jr ( 1980) Syndromes of cardiovascular malformations induced by copper citrate in hamsters. Teratology, 21:89-10 I. Ellenhorn MJ, Barceloux DG (1988) Medical toxtcology. diagnosts and treatment uf human poisoning. New York, NY, Elsevier Science Publishmg Company, pp. 54, 84 Evans GW (1973) Copper homeostasis and metabolism in the mammalian system. Physwlug1cal rev1ews, 53·535-569 FAO/IPCS (1994) Summarv of evaluatwns performed by the Jotnt FAO/WHO Expert Comm1ttee on Food Addittves Washmgton, DC, ILSI Press. Ferm VH, Hanlon DP ( 1974) Toxicity of copper salts in hamster embryonic development. Bwlogy of reproductiOn, 11 :97-10 I. Fukui H et al. (1994) Possible involvement of penpheral 5-HT4 receptors m copper sulfate-induced vomiting in dogs European Journal ofpharmacology, 257:47-52. Gollan JL, Deller DJ (1973) Studies on the nature and excretion of biimry copper in man. Cltmcal sctence, 44·9-15 Harris ZL, Gitlin JD (1996) GenetiC and molecular basis for copper toxicity. Amencan Journal of clmtcal nutritton, 63: 836S-841S Haywood S, Loughran M (1985) Copper toxicosis and tolerance m the rat. 11. Tolerance - a liver protective adaptation Liw!r, 5:267-275. Health Canada (1992) Copper. In: Gutdeltnes for Canadian dnnking water qualtty. Supportmg documentatiOn Ottawa, Ontario. Hebert CD et al. ( 1993) Subchronic toxicity of cupric sulfate admmistered in dnnking water and feed to rats and mice. Fundamental and applted toxicology, 21:461-475. Huang CP, Elliott HA, Ash mead RM ( 1977) Interfacial reactions and the fate of heavy metals in soil water systems. Journal ufthe Water Pollution Control Federatwn, 49(5).745-756. ISO (1986) Water quality - Determmatwn of cobalt, mckel, copper, zmc, cadmium and lead - Flame atomtc absorptton spectrometric methods Geneva, International Organization for Standardization (ISO 8288-1986 (E)). ISO (1996) Water quality- Determmatwn of 33 elements by mduct1ve~v coupled plasma atomtc em1sswn spectroscopy. Geneva, International Organization for Standardization (ISO 11885: 1996 (E)). Jantsch W, Kulig K, Rumack BH (1984) Massive copper sulfate ingestion resulting in hepatotoxicity. Journal of toxicology Clmtcaltoxtcology, 22(6)·585-588. Kanematsu N, Hara M, Kada T (1980) Rec assay and mutagenicity studies on metal compounds. Mutatwn research, 77:109-116. Knobeloch L et al. (1994) Gastrointestmal upsets associated with ingestion of copper-contaminated water. Environmental health perspecttves, I 02(11 ):958-961. Lecyk M (1980) Toxicity of cupric sulfate m mice embryonic development. Zoologtca Poloniae, 28(2):101-105. Lewis RJ Sr (1993) Hawley's condensed chemtcal dictwnary, 12th ed. (revised). New York, NY, Van Nostrand Remhold, pp. 309-315. Linder MC, Hazegh-Azam M (1996) Copper biochemistry and molecular biology. American journal ofclimcal nutntwn, 63:797S-811 S. Lonnerdal B ( 1996) Bioavaiiability of copper. American journal ofcltnical nutrition, 63:821 S-829S. Lopez JM, Lee GF (1977) Environmental chemistry of copper in Torch Lake, Michigan. Water, atr, and soil pol/utwn, 8:373-385. Low BA, Donohue JM, Bartley CB (1996) Backtlow prevention failures and copper poisonings associated with post-mix soft drink dispensers. Ann Arbor, M!, NSF International. Luza SC, Speisky HC (1996) Liver copper storage and transport dunng development: implications for cytotoxicity. Amencan JOUrnal ofclinical nutrttton, 63 :812S-820S. Makale MT, King GL (1992) Surgical and pharmacological dissociation of cardiovascular and emetic responses to intragastric CuSO,. American JOurnal ofphysiology, 263 :R284-29l. Matsui S (1980) Evaluation of a Bactllus subtilis rec-assay for the detection of mutagens which may occur in water environments. Water research, 14(1 I): 1613-1619. COPPER 45
Moore GS, Calabrese EJ (1980) G6PD-deficiency: a potential high-risk group to copper and chlorite ingestion Journal of environmental pathology and toxicology, 4(2-3):271-279. Moriya M et al. (1983) Further mutagenicity studies on pesticides in bacterial reversion assay systems. Mutatwn research, 116(3-4): 185-216. Muller T et al. ( 1996) Endemic Tyro lean infantile cirrhosis: an ecogenetic disorder. Lancet, 347:877-880. NAS (1989) Recommended d1etary allowances, lOth ed Washington, DC, National Academy of Sciences, Food and Nutrition Board, pp. 224-230. Nishioka H {1975) Mutagenic activities of metal compounds in bacteria. Mutatwn research, 31·185-189. NSF (1996) ANSIINSF Standard 60 Drinkmg water treatment chemicals - health effects. Ann Arbor, M!, NSF International. O'Donohue JW et al. (1993) M1cronodular cirrhosis and acute liver failure due to chronic copper self-intoxication. European journal ofgastroenterology and hepatology, 5(7):561-562. Olivares M, Uauy R (1996a) Limits of metabohc tolerance to copper and biological basis for present recommendations and regulations. AmencanJournal of clinical nutntwn, 63:846S-852S. Olivares M, Uauy R (1996b) Copper as an essenllal nutrient. American journal of clinical nutrition, 63:79IS-796S. Pandit A, Bhave S ( 1996) Present interpretation of the role of copper in Indian childhood cirrhosis Amencan journal of clinical nutntwn, 63:830S-835S. Payne K, Pickering WF {1975) Influence of clay-solute interactions on aqueous copper ion levels. Water, mr, and sod pollutiOn, 5:63-69. Pennington JA, Young BE, Wilson DB (1989) Nutritional elements in U.S. d1ets: results from the Total Diet Study, 1982-1986. Journal of the American D1etet1c AssociatiOn, 89·659-664. Petters son R, Sandstrum BM ( 1995) Copper. In: Oskarsson A, ed. Risk evaluatwn of essential trace elements. Copenhagen, Nordic Council of Ministers (Nord 1995: 18). Rana SV, Kumar A {1980) Biological, haematological and histological observations in copper poisoned rats. Industrial health, 18(1):9-17. Scheinberg IH, Sternleib I {1996) W!lson's disease and idiopathic copper toxicosis. Amertcan JOUrnal of cilmcal nutntion, 63:842S-845S. Schroeder HA et al. {1966) Essential trace metals in man: copper. Journal of chrome diseases, 19: I 007-1034. Sina JF et al. (1983) Evaluation of the alkaline elution/rat hepatocyte assay as a predictor of carcinogenic/mutagenic potential. Mutation research, 113(5):357-391. Singh I {1983) Induction of reverse mutation and mitotic gene conversion by some metal compounds in Saccharomyces cerev1siae. Mutatwn research, 117( 1-2): 149-152. Sirover MA, Loeb LA (1976) Infidelity of DNA synthesis m v1tro: screening for potential metal mutagens or carcinogens. Scwnce, 194:1434-1436. Spitalny KC et al. ( 1984) Dnnking-water-induced copper intoxication in a Vermont family. Pedtatncs, 74(6):1103-1106 Stumm W, M organ JJ {1981) Aquatic chemistry. New York, NY, W1ley Interscience. Toyokuni S, Sagripanti JL (1994) Increased 8-hydroxydeoxyguanosine in kidney and liver of rats contmuously exposed to copper. ToXIcology and applied pharmacology, 126:91-97. US EPA (1987) Drmkmg water criteria document for copper. Cincinnati, OH, US Environmental Protection Agency, Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office. US EPA {1991) Maximum contaminant level goals and nat10nal primary drinking water regulatiOns for lead and copper; final rule. Federal register, 56{110):26460-26564. US Environmental Protection Agency, 7 June. US EPA {1994) Methods for determination of metals 111 environmental samples. Supplement I Washington, DC, US Environmental Protection Agency, Office of Research and Development (EPA-600/R-94-111 ). US EPA {1995) Effect of pH, DJC, orthophosphate and sulfate on dnnkmg water cuprosolvency Washington, DC, US Environmental Protection Agency, Office of Research and Development (EPA/600/R-95/085). Wang SC, Borison HL {1951) Copper sui fate emesis: a study of afferent pathways from the gastrointestmal tract. American JOurnal ofphyswlogy, 164:520-526. 46 INORGANIC CONSTITUENTS
Wapnir RA, Devas G, Solans CV (1993) Inhibition of intestinal copper absorption by divalent cations and low-molecular-weight ligands in the rat. Biological/race element research, 36:291-305. Weast RC, ed. (1983) CRC handbook ofchemistry and physics. Boca Raton, FL, CRC Press. WHO (1996) Trace elements in human nutrition and health. Geneva, World Health Organization. WHO (in press) Copper. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria monograph). Wong PK (1988) Mutagenicity of heavy metals. Bulletin of environmental contamination and toxicology, 40:597-603. Yarze JC et al. (1992) Wilson's disease: current status. American journal ofmedicine, 92:643-655. NICKEL
First draft prepared by K. Petersson Grawe Toxicology Division, National Food Administration Uppsala, Sweden
In the 1993 WHO Guidelines for drinking-water quality, a health-based guideline value for nickel of 0.02 mg/litre was derived from a 2-year study in rats published in 1976 (NOAEL of 5 mg/kg of body weight per day; depressed body weight gain and altered organ-to-body-weight ratios were the critical end-points of toxicity). The Coordinating Committee for the updating of the Guidelines recommended that the guideline value for nickel be re-evaluated in light of more recent experimental data.
1. GENERAL DESCRIPTION
1.1 Identity
Nickel is a lustrous white, hard, ferromagnetic metal. It occurs naturally in five isotopic forms: 58 (67.8%), 60 (26.2%), 61 (1.2%), 62 (3.7%), and 64 (1.2%).
1.2 Physicochemical properties
Nickel usually has two valence electrons, but oxidation states of+ 1, + 3, or +4 may also exist. Nickel is not affected by water but is slowly attacked by dilute hydrochloric or sulfuric acid and is readily attacked by nitric acid. Fused alkali hydroxides do not attack nickel. Several nickel salts, such as the acetate, chloride, nitrate, and sulfate, are soluble in water, whereas hydroxides, oxides, carbonates, sulfides, disulfides, and subsulfides are practically insoluble in water. Alloys of nickel containing more than 13% chromium are to a high degree protected from corrosion in many media by the presence of a surface film consisting mainly of chromium oxide (Morgan & Flint, 1989; Haudrechy et al., 1994).
Property Value Specific density 8.90 g/cm3 at 25°C Melting point 1555°C Boiling point 2837°C
1.3 Major uses
Nickel is used principally in its metallic form combined with other metals and non-metals as alloys. Nickel alloys are characterized by their hardness, strength, and resistance to corrosion and heat. 48 INORGANIC CONSTITUENTS
Nickel is used mainly in the production of stainless steels, non-ferrous alloys, and super alloys. Other uses of nickel and nickel salts are in electroplating, as catalysts, in nickel-cadmium batteries, in coins, in welding products, and in certain pigments and electronic products (IARC, 1990). It is estimated that 8% of nickel is used for household appliances (IPCS, 1991 ).
1.4 Environmentalfate
Nickel occurs predominantly as the ion Ni(H20)/+ in natural waters at pH 5-9 (IPCS, 1991). Complexes with ligands, such as OH", SO/-, HC03-, Cl-, or NH3, are formed to a minor degree in this pH range.
2. ANALYTICAL METHODS
The two most commonly used analytical methods for nickel in water are atomic absorption spectrometry and voltammetry. The reported detection limits are 10 ng/litre and 1 ng/litre, respectively (IPCS, 1991; ATSDR, 1996). Flame atomic absorption spectrometry is suitable in the range of 0.5-100 ~-tg/litre (ISO, 1986), whereas inductively coupled plasma atomic emission spectroscopy can be used for the determination of nickel with a limit of detection of about 10 ~-tg/litre (ISO, 1996).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
3 Nickel concentrations in remote areas are in the range of 1-3 ng/m , whereas 3 concentrations in rural and urban air range from 5 to 35 ng/m . It has been estimated that non-occupational exposure via inhalation is 0.2-1.0 ~-tg/day in urban areas and 0.1-0.4 ~-tg/day in rural areas (Bennett, 1984). The mainstream smoke of one cigarette contains about 0.04-0.58 1-lg of nickel (IARC, 1990).
3.2 Water
Nickel concentrations in groundwater depend on the soil use, pH, and depth of sampling. The average concentration in groundwater in the Netherlands ranges from 7.9 ~-tg/litre (urban areas) to 16.6 ~-tg/litre (rural areas). Acid rain increases the mobility of nickel in the soil and thus might increase nickel concentrations in groundwater (IPCS, 1991 ). In groundwater with a pH <6.2, nickel concentrations up to 980 ~-tg/litre have been measured (RIVM, 1994). In Canada, the median nickel level in drinking-water supplies was below the detection limit of 2 ~-tg/litre; the maximum level observed was 69 ~-tg/litre (Meranger et al., 1981 ). In US drinking-water, 90% of all samples (n = 2503) contained :<;;1 0 ~-tg/litre, and 97% had nickel concentrations of :<;;20 ~-tg/litre (ATSDR, 1996). In Europe, reported nickel concentrations in drinking-water were generally below 10 ~-tg/litre (IPCS, 1991). Nickel levels below 1 ~-tg/litre have been reported from Denmark and Finland (Punsar et al., 1975; Gammelgaard & Andersen, 1985). NICKEL 49
Average dissolved nickel concentrations in surface water in the rivers Rhine and Meuse are below 7 flg/litre (RIWA, 1994). Increased nickel concentrations in groundwater and municipal tap-water (100-2500 flg/litre) in polluted areas have been reported (McNeely et al., 1971; Hopfer et al., 1989). Water left standing overnight in nickel-containing plumbing fittings contained a nickel concentration of 490 flg/litre (Andersen et al., 1983). Certain stainless steel well materials were identified as the source of increased nickel concentrations in groundwater wells in Arizona, USA. Mean nickel levels were 8-395 flg/litre; in some cases, nickel levels were in the range 1-5 mg/litre (Oakley & Korte, 1996). Leaching of nickel from chromium-nickel stainless steel pipework into drinking-water diminished after a few weeks; as chromium was rarely found at any time in the water, this indicates that the leakage of nickel is not of corrosive origin, but rather attributable to passive leaching of nickel ions from the surface of the pipes (Schwenk, 1992). Concentrations of nickel in water boiled in electric kettles may, depending on the material of the heating element, be markedly increased, especially in the case of new or newly decalcified kettles. Nickel concentrations in the range 100-400 flg/litre, with extreme values close to 1000 flg/litre, have been reported (Rasmussen, 1983; Pedersen & Petersen, 1995). Levels of nickel in bottled mineral water were below the detection limit of 25 flg/litre (Alien et al., 1989).
3.3 Food
Nickel levels in food are generally in the range 0.01-0.1 mg/kg, but there are large variations (Booth, 1990; Jorhem & Sundstrom, 1993; Dabeka & McKenzie, 1995; Levnedsmiddelstyrelsen. 1995). Higher median levels of nickel (0.1-0.4 mg/kg) were found in wholemeal products (Smart & Sherlock, 1987; Levnedsmiddelstyrelsen, 1995), whereas markedly higher levels (1-6 mg/kg) were found in beans, seeds, nuts, and wheat bran (Smart & Sherlock, 1987; Jorhem & Sundstrom, 1993). Even higher nickel levels (8-12 mg/kg) were found in cacao (Smart & Sherlock, 1987). Stainless steel cooking utensils (e.g. oven pans, roasting pans) contributed markedly to the levels of nickel in cooked food, sometimes exceeding 1 mglkg in meat (Dabeka & McKenzie. 1995). In contrast, Flint & Packirisamy (1995) found only minor increases in nickel concentrations in acid foodstuffs when new stainless steel pans were used. Daily dietary intakes of nickel were 140-150 flg in the United Kingdom in 1981-1984 (Smart & Sherlock, 1987), 82 flg in Sweden in 1987 (Becker & Kumpulainen, 1991), 150 flg in Denmark (Flyvholm et al., 1984), and 160 flg in the USA (Myron et al., 1978). The dietary intake of nickel in a well controlled Canadian study ranged from 190 flg/day for 1- to 4-year-old children to 406 flg/day for 20- to 39-year-old males. The nickel intake for 20- to 39-year-old women was on average 275 flg/day (Dabeka & McKenzie, 1995). Dietary nickel intake by 0- to 12-month old infants was on average 0.005 mglkg of body weight per day (equal to 50 INORGANIC CONSTITUENTS
0.038 mg/day). Infants fed evaporated milk were exposed to 0.004 mg/kg of body weight per day, whereas infants fed soy-based formula were exposed to 0.010 mglkg of body weight per day (Dabeka, 1989). As nuts and beans are important sources of protein for vegetarians, this population group can be expected to have a markedly higher intake of nickel than reported in the studies cited above. The nickel intake of eight volunteers when they ingested normal diets averaged 0.13 mg/day (range 0.0~.26 mg/day), compared with 0.07 mg/day when diets containing low nickel levels were consumed (range 0.02-0.14 mg/day). When food rich in nickel was ingested, the daily intake was 0.25 mg/day (range 0.07-0.48 mg/day) (Veien & Andersen, 1986).
3.4 Estimated total exposure and relative contribution of drinking-water
Food is the dominant source of nickel exposure in the non-smoking, non occupationally exposed population. According to the 1981 United Kingdom Total Diet Study. the contribution from food is 0.22-0.23 mg/day per person. Water generally contributes 0.005-0.025 mg daily (i.e. 2-11% of the total daily oral intake of nickel) (MAFF. 1985). The value of information on dietary intake of nickel based on market basket studies or on raw foodstuffs is limited because of the high impact that the cooking procedure may have on the nickel content in food. In studies including analyses of prepared food, water contributed, on average, only a minor part (<10%) of the daily dietary intake of nickel (Dabeka & McKenzie, 1995). However, in cases of heavy pollution. use of certain types of kettles, use of non-resistant material in wells, or use of water standing for an extended time in water pipes, the nickel contribution from water may be significant.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
4.1 Laboratory animals
Nickel is poorly absorbed from diets and is eliminated mainly in the faeces. Absorbed nickel is rapidly cleared from serum and excreted in urine (JP CS, 1991). The mechanism for intestinal absorption of nickel is not clear. Iron deficiency increased intestinal nickel absorption in vitro and in vivo, indicating that nickel is partially absorbed by the active transfer system for iron absorption in the intestinal mucosal cells (Tallkvist et al., 1994). In perfused rat jejunum, saturation of nickel uptake was observed at high concentrations of nickel chloride (Foulkes & McMullen, 1986). Iron concentrations in rat tissues were increased by dietary nickel exposure (Whanger, 1973 ). Nickel is bound to a histidine complex, albumin, and alpha-2-macroglobulin in serum (Sarkar, 1984). Absorption of soluble nickel compounds from drinking-water is higher than that from food. After 24 hours, 10-34% of a single oral dose ofwater-soluble nickel compounds (i.e. NiSO~, NiC1 2, Ni(N03) 2) was absorbed, whereas less than 2% of a single oral dose of insoluble or scarcely soluble nickel compounds (i.e. NiO, Ni, NICKEL 51
Ni3S2, NiS) was absorbed. It is not known if the animals were fasted before treatment. The highest nickel concentrations were found in the kidneys and lungs, whereas nickel concentrations in the liver were low (Ishimatsu et al., 1995). Whole-body retention in mice after oral exposure to N?+ was less than 1% of the administered dose 5 days after exposure (Nielsen et al., 1993). Severa et al. (1995) observed an accumulation of nickel in organs in rats orally exposed to nickel in drinking-water at concentrations of 100 mg/litre for 6 months. The nickel concentration in liver was 10 times higher in exposed rats than in unexposed rats; in the kidney, the nickel level was only twice as high in exposed rats as in unexposed rats. Nickel levels in the kidney and blood were similar. There was no increase in nickel levels in organs between 3 and 6 months of exposure. Biliary excretion of nickel subcutaneously administered to rats as nickel chloride was less than 0.5% of the given dose (Marzouk & Sunderman, 1985). Several reports indicate that transplacental transfer of nickel occurs in animals (IPCS, 1991). Elevated concentrations of nickel were detected in fetuses after intramuscular administration of nickel chloride to rats. The fetal organ with the highest nickel concentration was the urinary bladder (Sunderman et al., 1978). A dose-dependent increase in nickel concentrations in rat milk was observed after a single subcutaneous injection of nickel chloride. The milk/plasma ratio was 0.02 (Dostal et al., 1989).
4.2 Humans
Following a 12-hour fast, a volunteer ingested 20 !J.g of 61 Ni-enriched nickel per kg of body weight as Ni(N03) 2 in 1 litre of water. The serum nickel concentration peaked at 2 hours at 34 ~J.g/litre. By 96 hours, 27% of the ingested dose was excreted in the urine (Templeton et al., 1994a). These findings are consistent with the observations made by Sunderman and eo-workers, who reported an absorption of 27 ± 17% of the given nickel dose (as nickel sulfate) added to drinking-water in 10 volunteers after a 12-hour fast. Intestinal absorption was only 1% of the given dose when nickel as sulfate salt was added to scrambled eggs. The half-time for absorbed nickel averaged 28 ± 9 hours (Sunderman et al., 1989). Plasma levels in fasting human subjects did not increase above fasting levels when 5 mg of nickel were added to an American breakfast or a Guatemalan meal rich in phytic acids (Solomons et al., 1982). The same amount of nickel added to water elevated the plasma nickel levels 4- to 7-fold. The absorption of nickel added to milk. tea, coffee, or orange juice was significantly less than the absorption of nickel from water. A fatal case of nickel intoxication indicates that biliary excretion of nickel is of minor importance in humans (Grandjean et al., 1989). According to the above studies, the daily amount of absorbed nickel in humans will be, on average, about 5 !J.g each from water and food. Nickel has been detected in fetal tissues at levels similar to the levels found in adults (McNeely et al., 1972; Casey & Robinson, 1978). Serum levels in the range 1.5-19 ~J.g/litre were found in patients undergoing regular haemodialysis (Hopfer et al., 1989; Nixon et al., 1989). Significantly higher 52 INORGANIC CONSTITUENTS serum nickel levels were observed in non-occupationally exposed subjects from a heavily nickel-polluted area compared with levels in subjects living in a control area (nickel concentrations in tap-water 109 ± 46 vs 0.6 ± 0.2 1-Lg/litre; serum nickel levels 0.6 ± 0.3 vs 0.2 ± 0.2 1-Lg/litre) (Hopfer et al., 1989). Tentative reference values for nickel in serum and urine have been proposed: 0.2 ~J.g/litre or lower in serum, and 1-3 1-Lg/litre in urine of healthy adults (Templeton et al., 1994b).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO SYSTEMS
5.1 Acute exposure
Effects on kidney function, including tubular and glomerular lesions, have been reported by several authors after parenteral administration of high nickel doses of between 1 and 6 mg/kg of body weight intraperitoneally in rabbits and rats (IPCS, 1991 ). Intramuscular injection of the insoluble compound nickel subsulfide caused acute kidney damage in mice (Rodriguez et al., 1996). Increased hepatic lipid peroxidation and increased serum levels of aspartate and alanine aminotransferase activity were observed when rats were parent~rally exposed to nickel chloride at nickel doses of7-44 mg/kg of body weight (Donskoy et al., 1986). Levels ofhepatic monooxygenases decreased when mice were administered 10 mg of nickel per kg of body weight subcutaneously (Iscan et al., 1995).
5.2 Short-term exposure
Body-weight gain, haemoglobin, and plasma alkaline phosphatase were significantly reduced in weanling rats exposed to nickel (as nickel acetate) at concentrations of 500 or 1000 mg/kg in the diet (equivalent to 25 or 50 mg/kg of body weight per day) for 6 weeks compared with controls (Whanger, 1973). No effects were observed in rats exposed to 100 mg/kg in the diet (equivalent to 5 mg/kg of body weight per day).
5.3 Long-term exposure
Rats (25 per sex per dose) were exposed to nickel (as nickel sulfate) in the diet at doses of 0, 100, 1000, or 2500 mg/kg (equivalent to 0, 5, 50, or 125 mg/kg of body weight per day) for 2 years (Ambrose et al., 1976). Growth was depressed in rats at 1000 and 2500 mg/kg of diet, but there were indications that decreased food consumption might explain the decreased body-weight gains, particularly at 2500 mg/kg of diet. However, no statistical analysis seems to have been performed. Survival was overall very poor, especially in the control groups and the 2500 mg/kg of diet groups. In females at 1000 and 2500 mg/kg of diet, the mean relative liver weights were decreased by about 20%, and the mean relative heart weights were increased by about 30% compared with the control group. No histological or gross pathological findings related to nickel exposure were observed. The highest nickel concentrations were found in the kidneys. The NOAEL in this study was 5 mg/kg of body weight per day. However, the study does not meet current standards for NICKEL 53 long-term studies, mainly because of the low survival rate. The observed changes in organ weights in female rats might in part be due to changes in food and water consumption. Also, both gross and histopathological examinations of the animals were negative, although there were 20-30% changes in relative organ weights. It can thus not be excluded that the observed changes in relative organ weights were related to changes in food and/or water consumption rather than to a toxic effect of nickel. Increased relative kidney weight was observed in rats exposed to nickel (as nickel sulfate) in drinking-water at daily doses of about 7 mg/kg of body weight for up to 6 months (Vyskocil et al., 1994). There was an increased excretion of albumin in urine in females. but there were no changes in total protein, beta-2-microglobulin, N-acetyl-beta-D-glucosaminidase, or lactate dehydrogenase in urine due to nickel exposure. In a 2-year study, dogs (three per sex per dose) were exposed to 0, 100, 1000, or 2500 mg of nickel per kg of diet (equivalent to 0, 2.5, 25, and 62.5 mg/kg of body weight per day). In the 2500 mg/kg of diet group, decreased weight gain and food consumption, higher kidney- and liver-to-body-weight ratios, and histological changes in the lung were observed. The NOAEL in this study was 25 mg/kg of body weight per day (Ambrose et al.. 1976).
5.4 Reproductive and developmental toxicity
Intraperitoneal administration of nickel nitrate ( 12 mg of nickel per kg of body weight) to male mice resulted in reduced fertilizing capacity of spermatozoa; no effects were seen at 8 mg of nickel per kg of body weight (Jacquet & Mayence, 1982). Reduced number of live pups and reduced body weight of fetuses were observed in rats exposed to single doses of nickel chloride ( 16 mg of nickel per kg of body weight) or nickel subsulfide (80 mg of nickel per kg of body weight) administered intramuscularly on day 8 and day 6, respectively (Sunderman et al., 1978). No congenital anomalies were found in the fetuses. In a three-generation study in rats at dietary levels of250, 500, or 1000 mg of nickel (administered as nickel sulfate) per kg of diet (equivalent to 12.5, 25, or 50 mg/kg of body weight per day), a higher incidence of stillborn in the first generation was observed compared with the control group (Ambrose et al., 1976). Body weights were decreased in weanlings at 1000 mglkg of diet in all generations. The number of pups born alive per litter and the number of pups weaned per litter were progressively fewer with increasing nickel dose, but no statistical analysis of the results is presented. Decreased weanling body weight is a clear-cut effect in the 1000 mg/kg of diet dose group. No teratogenic effects were observed in any generation at any dose level. No histological lesions were observed in the third generation at weaning. Decreased litter sizes were observed in a small-scale three-generation study in rats administered nickel in drinking-water at 5 mg/litre, corresponding to 0.2 mg/kg of body weight per day (Schroeder & Mitchener, 1971). Velazquez & Poirer (1994) and ATSDR (1996) described a two-generation study in rats. Nickel chloride was administered in drinking-water at concentrations of 0, 50, 250, or 500 mg/litre (equal to 0, 7, 31, and 52 mg of nickel per kg of body 54 INORGANIC CONSTITUENTS weight per day) from 90 days before breeding. Along with changes in maternal body weight and liver weight at the 500 mg/litre dose level in the P0 generation, there were also a dose-related decrease in live litter size and pup weight and increased neonatal mortality. In the F 1 generation, there was dose-related mortality between 3 and 7 weeks of age at the 250 and 500 mg/litre dose levels. For the F1 matings, there were also dose-related decreases in live litter size and increased mortality per litter, but this was significant only in the high-dose group. Decreased food intake and water intake were observed in the exposed animals. Also, the room temperature was up to 6°C higher than normal at certain times during gestation and the early postnatal days. Lower than normal levels of humidity were also recorded. Thus, the NOAEL in this study is considered to be 7 mg of nickel per kg of body weight per day; however, because of the problems referred to, it is difficult to make a direct association between the effects reported in this study and nickel exposure. Female Long-Evans rats were exposed for 11 weeks prior to mating and then during two successive gestation periods (G 1 and G2) and lactation periods (Ll and L2) to 0, 10, 50, or 250 ppm nickel as nickel chloride (equal to 0, 1.3, 6.8, and 31.6 mg of nickel per kg of body weight per day) in drinking-water (Smith et al., 1993). Dams drinking water containing nickel at 31.6 mg/kg of body weight per day consumed less liquid and more food per kg of body weight than did controls. Maternal weight gain was reduced during G 1 in the mid- and high-dose groups. There were no effects on pup birth weight, and weight gain was reduced only in male pups from darns in the mid-dose group. The proportion of dead pups per litter was significantly elevated at the high dose in L1 and at the low and high dose in L2 (the increase at the middle dose in L2 approached statistical significance), with a dose-related response in both experimental segments. The number of dead pups per litter was significantly increased at each dose in L2. It was noted that the number of litters with dead pups and the total number of dead pups per litter in the control group were less in L2 than in Ll. Plasma prolactin levels were reduced in darns at the highest dose level 1 week after weaning of the second litter. The authors concluded that 1.3 mg/kg of body weight per day represented the LOAEL in this study, although this is considered to be conservative, owing to variations in response between the successive litters. Alterations in milk composition were observed in lactating rats exposed to four daily subcutaneous injections of nickel at doses of 3-6 mg/kg of body weight (Dostal et al., 1989). Liver weights were decreased in pups whose darns received 6 mg of nickel per kg of body weight. These findings may explain the effects seen on litter size and body weights of the pups in studies described above.
5.5 Mutagenicity and related end-points
Nickel compounds are generally inactive in bacterial mutation assays but active in mammalian cell systems (IPCS, 1991 ). It was concluded that there was cell toxicity in all gene mutation studies indicating nickel-dependent lesions using mammalian cells. NICKEL 55
Chromosomal gaps, deletions and rearrangements, DNA-protein cross-links, and sister chromatid exchanges are reported in mammalian systems, including human cell systems. Chromosomal aberrations occur in all chromosomes but with preference to the heterochromatic centromeric regions (IPCS, 1991; Rossman, 1994). In several experimental systems, nickel ions have been shown to potentiate the effects of other mutagenic agents, which may be explained by the capacity of nickel to inhibit DNA repair (Lynn et al., 1994; Rossman, 1994).
5.6 Carcinogenicity
A nuniber of studies on the carcinogemc1ty of nickel compounds in experimental animals are available (IARC, 1990; Aitio, 1995). Generally, tumours are induced at the site of administration of the nickel compound. For instance, several nickel compounds induce injection-site sarcomas (Sunderman, 1984). Recently, a marked variation in the incidence of injection-site sarcomas between different strains of mice was reported (Rodriguez et al., 1996). There are only a limited number of studies on carcinogenic effects after oral exposure to nickel compounds. The incidence of tumours was not higher in rats exposed to drinking-water containing nickel at 5 mgllitre during their lifetime compared with control rats (Schroeder et al., 1974). As well, no difference in tumour incidence was observed in a lifetime study in rats exposed to 5, 50, or 125 mg of nickel per kg of body weight per day in the feed compared with controls (Ambrose et al., 1976). Owing to the high death rate and lack of information on cause of death, this study is of minor value in evaluating carcinogenicity after oral exposure to nickel.
5. 7 Other effects
Nickel salts affect the T -cell system and suppress the activity of natural killer cells in rats and mice (IPCS, 1991). Mitogen-dependent lymphocyte stimulation was inhibited in human lymphocytes (Sikora & Zeromski, 1995) and in spleens of mice exposed to nickel (IPCS, 1991). Dose-related decreased spleen proliferative response to lipopolysaccharide was observed in mice exposed to nickel sulfate in drinking water for 180 days. At the lowest dose (44 mg of nickel per kg of body weight per day), decreased thymus weight was observed, but there was no nickel-induced immunosuppression NK cell activity or response to T -cell mitogens. Parenteral administration of nickel to rabbits, chickens, and rats and oral administration of nickel to rabbits induce hyperglycaemia and reduce the levels of prolactin releasing factor in rats (IPCS, 1991 ). The myeloid system was affected (i.e. decrease in bone marrow cellularity and dose-related reductions in the bone marrow proliferative response) when mice were exposed to nickel sulfate in drinking-water at doses of 0, 44, 108, or 150 mg of nickel per kg of body weight per day for 180 days (Dieter et al., 1988). The LOAEL in this study was 44 mg of nickel per kg of body weight per day. 56 INORGANIC CONSTITUENTS
6. EFFECTS ON HUMANS
6.1 Acute exposure
A 2Yz-year-old girl died after ingesting about 15 g of nickel sulfate crystals. Cardiac arrest occurred after 4 hours; the autopsy revealed acute haemorrhagic gastritis (Daldrup et al., 1983). Thirty-two industrial workers accidentally drank water contaminated with nickel sulfate and nickel chloride (1.63 g of nickel per litre). The nickel doses in persons who developed symptoms were estimated to range from 7 to 35 mg/kg of body weight. Twenty workers developed symptoms, including nausea, vomiting, diarrhoea, giddiness, lassitude, headache, and shortness of breath. In most cases, these symptoms lasted for a few hours, but they persisted for 1-2 days in seven cases. Transiently elevated levels of urine albumin suggesting mild transient nephrotoxicity were found in two workers 2-5 days after exposure. Mild hyperbilirubinaemia developed on day 3 after exposure in two subjects, and elevated levels of blood reticulocytes were observed in seven workers on day 8 post-exposure. It is known from animal studies that nickel after intrarenal injection enhances the renal production of erythropoietin, which may explain the reticulo-cytosis, and that nickel induces microsomal haem oxygenase activity in liver and kidney, leading to a secondary hyperbilirubinaemia. Serum nickel concentrations ranged between 13 and 1340 j.tg/litre in persons with symptoms (Sunderman et al., 1988). Seven hours after ingesting nickel sulfate in drinking-water (50 j.lg of nickel per kg of body weight), a 55-year-old man developed left homonymous haemianopsia, which lasted 2 hours (Sunderman et al., 1989). Nickel intoxication in 23 patients receiving haemodialysis was reported (Webster et al., 1980). The dialysate was contaminated by leachate from a nickel plated stainless steel water heater tank. Symptoms such as nausea, vomiting, headache, and weakness occurred rapidly after exposure at plasma nickel concentrations of about 3 rug/litre and persisted for 3-13 hours after dialysis.
6.2 Skin irritation and hypersensitivity
Allergic contact dermatitis is the most prevalent effect of nickel in the general population. A recent epidemiological investigation showed that 20% of young (15-34 years) Danish women and 10% of older (35-69 years) women were nickel-sensitized, compared with only 2-4% of Danish men (15-69 years) (Nielsen & Menm~, 1992). The prevalence of nickel allergy was found to be 7-10% in previously published reports (Menne et al., 1989). Edetic acid (EDT A) reduced the number and severity of patch test reactions to nickel sulfate in nickel-sensitive subjects (Allenby & Goodwin, 1983). Systemically induced flares of dermatitis are reported after oral challenge of nickel-sensitive women with 0.5-5.6 mg of nickel as nickel sulfate administered in a lactose capsule (Veien, 1989). At the highest nickel dose (5.6 mg), there was a positive reaction in a majority of the subjects; at 0.5 mg, only a few persons responded with flares. Responses to oral doses of 0.4 or 2.5 mg of nickel did not NICKEL 57 exceed responses in subjects given placebos in double-blind studies (Jordan & King, 1979; Gawkrodger et al., 1986). After an oral dose of 1 mg of nickel. significantly higher levels of nickel were found in the urine of atopic patients (i.e. persons with a history of flexural dermatitis) compared with controls, indicating a higher gastrointestinal absorption of nickel in atopic persons (Hindsen et al., 1994). No such difference was found between nickel-allergic patients and controls. The small number of patients may explain these unexpected findings. There are several reports on the effects of diets low or high in nickel. but it is still a matter of discussion whether naturally occurring nickel in food may worsen or maintain the hand eczema of nickel-sensitive patients, mainly because results from dietary depletion studies have been inconclusive (Veien & Menne, 1990). In a single blind study, 12 nickel-sensitive women were challenged with a supplementary high nickel diet (Nielsen et al., 1990). The authors concluded that hand eczema was aggravated during the period (i.e. days 0-11) and that the symptoms thus were nickel induced. However, it should be noted that in some subjects the severity of the eczema (i.e. the number of vesicles in the palm of the hand) varied markedly between day 14 or 21 before the challenge period and the start of the challenge period. Oral hyposensitization to nickel was reported after six weekly doses of 5 mg of nickel in a capsule (Sjowall et al., 1978) and 0.1 ng of nickel sulfate daily for 3 years (Panzani et al., 1995). Cutaneous lesions were improved in eight patients with contact allergy to nickel after oral exposure to 5 mg of nickel weekly for 8 weeks (Bagot et al., 1995). Nickel in water (as nickel sulfate) was given to 25 nickel-sensitive women in daily doses of 0.01-0.04 mg/kg of body weight per day for 3 months after they had been challenged once with 2.24 mg of nickel (Santucci et al., 1988). In 18 women, flares occurred after the challenge dose, whereas only 3 out of 17 subjects had symptoms during the prolonged exposure period. Later, Santucci and eo-workers (1994) gave increasing oral doses of nickel in water (0.01-0.03 mg of nickel per kg of body weight per day) to eight nickel sensitive women for up to 178 days. A significant improvement in hand eczema was observed in all subjects after 1 month.
6.3 Carcinogenicity
The identification of nickel species hazardous to humans was investigated by the International Committee on Nickel Carcinogenesis in Man by analysing 10 previously studied cohorts of men occupationally exposed to nickel (ICNCM, 1990). It was concluded that occupational exposure to sulfidic and oxidic nickel at high concentrations causes lung and nasal cancers. There was no correlation between metallic nickel exposure and cancer in lung or nose. Soluble nickel exposure increased the cancer risk and may also enhance the risk associated with exposure to less-soluble nickel compounds. The Committee also concluded that there was no substantial evidence that nickel compounds may produce cancers other than in the lung or nose in occupationally exposed persons. Inhalation is an important route of exposure to nickel and its salts in relation to health risks. IARC (1990) concluded that nickel compounds are carcinogenic to 58 INORGANIC CONSTITUENTS humans (Group 1), whereas metallic nickel is possibly carcinogenic to humans (Group 2B). However, there is a lack of evidence of a carcinogenic risk from oral exposure to nickel.
7. PROVISIONAL GUIDELINE VALUE
In several limited studies, NOAELs were approximately 5 mg of nickel per kg of body weight per day. In a well conducted recent two-generation study on rats, dose-related increases in perinatal mortality were observed, giving a LOAEL of 1.3 mg of nickel per kg of body weight per day in the second litter, whereas the lowest-observed-effect level in the first litter was 31.6 mg of nickel per kg of body weight per day. These variations in response between successive litters make it difficult to draw firm conclusions. Also, a NOAEL of 7 mg of nickel per kg of body weight per day was derived from a more limited two-generation study in rats. The guideline value for nickel of 0.02 mg/litre is maintained because, on the basis of the available data, it is considered to be protective of public health. Owing to uncertainties about the effect level for perinatal mortality, the value is considered to be provisional.
8. REFERENCES
Aitio A (1995) Ntckel and mckel compounds. Stockholm, National Institute of Working Life, Nordic Council of Ministers, The Nordic Expert Group for Criteria Documentation of Health Risks from Chemicals, 61 pp. (Arbete och balsa 26). Alien HE, Halley-Henderson MA, Hass CN (1989) Chemical composition of bottled mineral water. Archives of environmental health, 44: I 02-116. Allenby CF, Goodwin BF (1983) Influence of detergent washing powders on minimal eliciting patch test concentrations of nickel and chromium. Contact dermatitis, 9:491--499. Ambrose AM et al. (1976) Long term toxicologic assessment of nickel in rats and dogs. Journal of food science and technology, 13:181-187. Andersen KE et al. (1983) Nickel in tap water. Contact dermatitis, 9:140-143. A TSDR (1996) Toxicological profile for mcke/. Draft for public comment. Atlanta, GA, US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry. Bagot M et al. (1995) Oral desensitization in nickel allergy induces a decrease in nickel-specific T cells. European Journal ofdermatology, 5:614--617. Becker W, Kumpulainen J (1991) Contents of essential and toxic mineral elements in Swedish market-basket diets in 1987. British journal ofnutritwn, 66:151-160. Bennett BG (1984) Environmental nickel pathways to man. In: Sunderman FW Jr, ed. Nickel in the human environment. Lyon, International Agency for Research on Cancer, pp. 487--495 (IARC Scientific Publications No. 53). Booth J (1990) Nickel in the diet and its role in allergic dermatitis. Journal of human nutrition and dietetics, 3:233-243. Casey CE, Robinson MF (1978) Copper, manganese, zinc, nickel, cadmium and lead in human foetal tissues. British journal of nutrition, 39:639-646. Dabeka RW (1989) Survey of lead, cadmium, cobalt and nickel in infant formulas and evaporated milks and estimation of dietary intakes of the elements by infants 0-12 months old. Science ofthe total environment, 89:279-289. Dabeka RW, McKenzie AD (1995) Survey of lead, cadmium, fluoride, nickel, and cobalt in food composites and estimation of dietary intakes of these elements by Canadians in 1986-1988. Journal of the Association ofOjjicwl Analytical Chemists International, 78:897-909. NICKEL 59
Daldrup T, Haarhoff K, Szathrhary SC (1983) [Fatal nickel sulfate poisoning.] Bellrdge zur Gerichtlichen Medizin, 41:141-144 (in German with English summary). Dieter MP et al. ( 1988) Evaluation of tissue disposition, myelopoietic, and immunologic responses in mice after long-term exposure to nickel sulfate in the drinking water. Journal of toxicology and environmental health, 24:356-372. Donskoy E et al. (1986) Hepatic toxicity of nickel chloride in rats. Annals of clinical laboratory science, 16:108-117. Dostal LA et al. ( 1989) Effects of nickel chloride on lactating rats and their suckling pups, and the transfer of nickel through rat milk. Toxicology and applied pharmacology, I 0 I :220-231. Flint GN, Packirisamy S (1995) Systemic nickel: the contribution made by stainless-steel cooking utensils. Contact dermatitis, 32:218-224. Flyvholm MA, Nielsen GO, Andersen A (1984) Nickel content of food and estimation of dietary intake. Zeitschriftfur Lebensmittel-Untersuchung und -Forschung, 179:427-431. Foulkes EC, McMullen OM (1986) On the mechanism of nickel absorption in the rat jejunum. Toxicology, 38:35-42. Gammelgaard B, Andersen JR (1985) Nickel in tap water. Contact dermatitis, 12:123. Gawkrodger DJ et al. ( 1986) Nickel dermatitis: the reaction to oral nickel challenge. British journal ofdermatology, 115:33-38. Grandjean P, Nielsen GO, Andersen 0 (1989) Human nickel exposure and chemobiokinetics. In: Maibach HI, Menne T, eds. Nickel and the skin- Immunology and toxicology. Boca Raton, FL, CRC Press, Inc., pp. 9-35. Haudrechy P et al. (1994) Nickel release from nickel-plated metals and stainless steels. Contact dermatitis, 31:249-255. Hindsen M, Christensen OB, Moiler H (1994) Nickel levels in serum and urine in five different groups of eczema patients following oral ingestion of nickel. Acta Dermato Venereologica, 74:176-178. Hopfer SM, Fay WP, Sunderman FW Jr (1989) Serum nickel concentrations in hemodialysis patients with environmental exposure. Annals ofclinical laboratory science, 19:161-167. IARC (1990) Nickel and nickel compounds. In: Chromium, mckel and weldmg. Lyon, International Agency for Research on Cancer, pp. 257-445 {IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 49). ICNCM (1990) Report of the International Committee on Nickel Carcinogenesis in Man. Scandinavian journal of work, environment & health, 16:1-82. IPCS (1991) Nickel. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria I 08). !scan M et al. (1995) Differential responses of hepatic monooxygenases and glutathione S transferases of mice to a combination of cadmium and nickel. Comparative biochemistry and phystology, IIIC:61-68. Ishimatsu S et al. (1995) Distribution of various nickel compounds in rat organs after oral administration. Biological trace element research, 49:43-52. ISO (1986) Water quality- Determination of cobalt, nickel, copper, zinc, cadmium and lead Flame atom1c absorption spectrometric methods. Geneva, International Organization for Standardization (ISO 8288-1986 (E)). ISO (1996) Water quality- Determination of 33 elements by inductively coupled plasma atomic emission spectroscopy. Geneva, International Organization for Standardization {ISO 11885:1996 (E)). Jacquet P, Mayence A (1982) Application of the in vitro embryo culture to the study of the mutagenic effects of nickel in male germ cells. Toxicology letters, 11:193-197. Jordan WP, King SE (1979) Nickel feeding in nickel-sensitive patients with hand eczema. Journal of the American Academy ofDermatology, I :506-508. Jorhem L, Sundstrom B (1993) Levels of lead, cadmium, zinc, copper, nickel, chromium, manganese and cobalt in foods on the Swedish market 1983-1990. Journal of food composition and analysis, 6:223-241. Levnedsmiddelstyrelsen (1995) Overvagmngssystem for levnedsm1dler 1988-1992. Soborg, Sundhetsministeriet (Publication No. 232). 60 INORGANIC CONSTITUENTS
Lynn S et al. (1994) Glutathione can rescue the inhibitory effects of nickel on DNA ligation and repa1r synthesis. Car.cmogenes1s, 15:2811-2816. MAFF (1985) Survey of alwnimum, antimony, chromium, cobalt, indium, nickel, thallium and tin in food. London, Ministry of Agriculture, Fisheries and Food (Food Surveillance Paper No. 15). Marzouk A, Sunderman FW Jr (1985) Biliary excretion of nickel in rats. Toxtcology letters, 27:65-71. McNeely MD, Nechay MW, Sunderman FW Jr (1972) Measurements of mckel in serum and urine as indices of environmental exposure to nickel. Cltmcal chemistry, 18:992-995. Menne T, Chnstophersen J, Green A (1989) Epidemiology of mckel dermatitis. In. Maibach HI, Men ne T, eds. Ntckel and the sktn tmmunology and toxicology. Boca Raton, FL, CRC Press, pp. 109-115. Meranger JC, Subramanian KS, Chalifoux C (1981) Survey for cadmium, cobalt, chromium, copper, nickel, lead, zinc, calcium, and magnesmm in Canadian drinking water supplies. Journal of the Assoctation of Official Analytical Chemists, 64:44-53. Morgan LG, Flint GN (1989) Nickel alloys and coatings: release of nickel In: Maibach HI, Menne T, eds. Nickel and the skm Immunology and toxtcology Boca Raton, FL, CRC Press, pp. 45-54. Myron DR et al (1978) Intake of mckel and vanadium by humans. A survey of selected diets. American journal ofclimcal nutritiOn, 31:527-531. Nielsen GD et al. (1990) Nickel-sensitive patients with vesicular hand eczema: oral challenge with a diet naturally h1gh in nickel. Brttish;ournal ofdermatology, 122·299-308. Nielsen GD, Andersen 0, Jensen M (1993) Toxicokinetics of nickel in mice studied w1th the gamma emitting isotope "Ni. Fundamental and applied toxtcology, 21:236-243. Nielsen NH, Menne T (1992) Allergic contact sensitization in an unselected Danish population. Acta Dermato Venereologtca, 72:456-460. Nixon DE et al (1989) Determination of serum nickel by graphite furnace atomic absorption spectrometry with Zeeman-effect background correction: values m a normal population and a population undergoing dialysis. Analyst, 114:1671-1674. Oakley D, Korte NE (1996) Nickel and chromium in groundwater samples as influenced by well construction and sampling methods. Groundwater momtonng review, Winter:93-99. Panzani RC et al. (1995) Oral hyposensitization to nickel allergy. preliminary clinical results. InternatiOnal archtves of allergy and applied immunology, I 07:251-254. Pedersen GA, Petersen J ( 1995) Undersogelse af mkkel-, chrom- og b(vafgtvelse fra el-kogekander samt kartlagmng af metalafgtvelse fra kaffemaskiner. [Investtgatwn of the release of nickel, chromtum and lead fi-om electrtc kettles and release of trace elements from coffeemachines.] Copenhagen, National Food Agency (Report ILF 1995.1) (in Danish with English summary). Punsar S et al. (1975) Coronary heart disease and drinking water. A search in two Finnish male cohorts for epidemiologic evidence of a water factor. Journal ofchronic dtseases, 28:259-287. Rasmussen G (1983) A/gtvelse af sporelementer (arsen, cadmtum, krom, kobber, nikkel, bly, antimon, tm og zmk) fra husholdmngsgestande. [Release of trace elements (arsenic, cadmtum, chromium, copper. mckel, lead, antimony, tm and zinc) from kitchen utenstls.] Copenhagen, National Food Agency (Publication No. 77) (in Danish with English summary). RIVM (1994) AttentiOn substanceo m Dutch envtronmental policy. Bilthoven, Rljkinstitut voor Volksgezondheid en Milieuhyg1ene (National Institute of Public Health and Environmental Protection) (Report No. 601014). RIWA (1994) Year(v report Parts A and B. Association of Rhine and Meuse Water Supply Companies.
, Rodriguez RE et al (1996) Relative susceptibility ofC57BL/6, (C57BL/6xC3H/He)F 1 and C3H/He mice to acute tOXICity and carcinogen1city of mckel subsulfide. Toxtcology, 107:131-140 Rossman TG (1994) Metal mutagenesis. In: Goyer RA, Cherian MG, eds. Toxicology of metals Berlm, Springer-Verlag, pp. 373-406. Santucci B et al. (1988) Nickel sensitivity: effects of prolonged oral intake of the element. Contact dermattiTs, 19:202-205. Santucci B et al. (1994) Serum and urine concentrations in nickel-sensitive patients after long prolonged oral administration. Contact dermatllTs, 30:97-10 I. NICKEL 61
Sarkar B (1984) Nickel metabolism In Sundennan FW Jr, ed. Nickel m the human environment Lyon, International Agency for Research on Cancer, pp. 367-384 (IARC Scientific Publications No. 53). Schroeder HA, Mitchener M (1971) Toxic effects of trace elements on the reproduction of mice and rats. Archives of environmental health, 23: I 02-106. Schroeder HA, Mitchener M, Nason AP (1974) Life-term effects of nickel in rats: survival, tumors, interactions with trace elements and tissue levels. Journal ofnutntion, 104:239-243. Schwenk W (I 992) Untersuchungen Iiber die Nickelabgabe nichtrostender Chrom-Nickel-Stiihle m Trinkwasser. [Nickel transfer from Cr-Ni stainless steel pipework into potable water.] GWF Wasser Abwasser, 133.281-286 (in German with English summary). Severa Jet al. (\995) Distribution of nickel in body fluids and organs of rats chronically exposed to mckel sulphate. Human and expenmentaltox1cology, 14:955-958. Sikora J, Zeromski J (I 995) The effects of mckel compounds on mitogen dependent human lymphocyte stimulation. Inrernalional ;ournal of Immunopathology and pharmacology, 8:79-85. Sjowall P, Christensen OB, Moller H (1978) Oral hyposensitization in nickel allergy. Journal of the Amencan Academy of Dermatology, I 7:774-778. Smart GA, Sherlock JC ( 1987) Nickel m foods and the diet. Food add1t1ves and contammants, 4:61-71. Smith MK et al. (I 993) Pen natal toxicity associated with nickel chloride exposure. Envtronmental research, 6 I :200-21 I. Solomons NW et al. (I 982) Bioavailability of nickel in man: effects of foods and chemically-defined dietary constituents on the absorption of inorganic nickel. Journal of nutritiOn, 112:39-50. Sunderman FW Jr ( 1984) Carcinogenicity of nickel compounds in ammals. In: Sunderman FW Jr, ed. Nickel m the human envwonmenr Lyon, International Agency for Research on Cancer, pp. 127-142 (I ARC Scientific Publications No. 53). Sunderman FW Jr et al. ( 1978) Embryotoxicity and fetal toxicity of nickel in rats. Tox1cology and applied pharmacology, 43:381-390. Sunderman FW Jr et al. (I 988) Acute nickel toxicity in electroplating workers who accidentally ingested a solution of nickel sulfate and nickel chloride. Amencan journal of mduslrial medtcme, 14:257-266. Sunderman FW Jr et al. (1989) Nickel absorption and kinetics m human volunteers. Proceedings of the Socwtyfor Expenmental Biology and Medic me, 191:5-11. Tallkvist J, Wmg AM, TJalve H (I 994) Enhanced mtestinal nickel absorption in iron-deficient rats. Pharmacology & toxicology, 75:244-249. Templeton DM, Xu SX, Stuhne-Sekalec L (1994a) Isotope-specific analysis of Ni by ICP-MS: applications of stable isotope tracers to biokinetic studies. Science of the total environment, 148:253-262. Templeton DM, Sunderman FW Jr, Herber RF (1994b) Tentative reference values for nickel concentrations in human serum, plasma, blood, and urine: evaluation according to the TRACY protocol. Science of the total environment, 148:243-251. Veien NK (1989) Nickel dermatitis. Its relationship to food and experimental oral challenge. In: Maibach HI, Menne T, eds. Nickel and the skin. immunology and toxicology. Boca Raton, FL, CRC Press, pp. 165-178. Veien NK, Andersen MR (1986) Nickel in Danish food. Acta Dermato Venereologica, 66:502-509. Veien NK, Menne T (1990) Nickel contact allergy and nickel-restricted diet. Seminars in dermatology, 9:197-205. Velazquez SF, Pmrer KA (1994) Problematic risk assessments for drinking water contaminants: selenium, aldicarb, and nickel. In: Wang RGM, ed. Water contamination and health IntegratiOn of exposure assessment, toxicology, and risk assessment. New York, NY, Dekker, pp. 467--495 (Environmental Science and Pollution Control Series, Vol. 9). Vyskocil A, Viau C, Cizkova M (1994) Chronic nephrotoxicity of soluble mckel in rats. Human & expenmenlaltoxicology, 13·689-693. Webster JD et al. (I 980) Acute nickel intoxication by dialysis. Annals ofinternal medicine, 92:631--633. Whanger PD (1973) Effects of dietary nickel on enzyme activities and mineral contents in rats. Tox1cology and applied pharmacology, 25:323-331.
NITRATE AND NITRITE
First draft prepared by G.J.A. Speijers National Institute ofPublic Health and Environmental Protection Bilthoven, The Netherlands
A health-based guideline value for nitrate of 50 mg/litre (expressed as nitrate ion) was recommended in the second edition of the WHO Guidelines for drinking water quality to prevent methaemoglobinaemia, a condition that occurs when nitrate is reduced to nitrite, in infants. Most clinical cases of methaemoglobinaemia occur at nitrate levels of 50 mg/litre and above, and almost exclusively in infants under 3 months of age. Accepting a relative potency for nitrite and nitrate with respect to methaemoglobin (metHb) formation of 10:1 (on a molar basis), a provisional guideline value for nitrite of 3 mg/litre (expressed as nitrite ion) was proposed. Because of the possibility of simultaneous occurrence of nitrite and nitrate in drinking-water, the sum of the ratios of the concentrations of each to its guideline value was not to exceed one. In 1995, JECF A established ADis for nitrate and nitrite and the US National Research Council published an evaluation of nitrate and nitrite in drinking-water. In light ofthis new information, the Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality determined that there was a need to reassess the existing guideline values for nitrate and nitrite.
1. GENERAL DESCRIPTION
1.1 Identity
Nitrate and nitrite are naturally occurring ions that are part of the nitrogen cycle. The nitrate ion (N03-) is the stable form of combined nitrogen for oxygenated systems. Although chemically unreactive, it can be reduced by microbial action. The nitrite ion (N02-) contains nitrogen in a relatively unstable oxidation state. Chemical and biological processes can further reduce nitrite to various compounds or oxidize it to nitrate (ICAIR Life Systems, Inc., 1987).
1.2 Physicochemical properties (ICAIR Life Systems, Inc., I 987/
Property Nitrate Nitrite Acid Conjugate base of strong Conjugate base of weak acid acid HN03; pKa =-1.3 HN02; pKa = 3.4
1 Conversion to nitrogen: I mg NO,-/Iitre = 0.226 mg No,--N/Iitre; I mg No,-llitre = 0.304 mg No,--N/Iitre. 64 INORGANIC CONSTITUENTS
Property Nitrate Nitnte Salts Very soluble in water Very soluble in water Reactivity Unreactive Reactive; oxidizes antioxidants, Fe2+ of haemoglobin (Hb) to Fe3+, and primary amines; nitrosates several amines and am ides
1.3 Major uses
Nitrate is used mainly in inorganic fertilizers. It is also used as an oxidizing agent and in the production of explosives, and purified potassium nitrate is used for glass making. Sodium nitrite is used as a food preservative, especially in cured meats. Nitrate is sometimes also added to food to serve as a reservoir for nitrite.
1.4 Environmental fate
In soil, fertilizers containing inorganic nitrogen and wastes contammg organic nitrogen are first decomposed to give ammonia, which is then oxidized to nitrite and nitrate. The nitrate is taken up by plants during their growth and used in the synthesis of organic nitrogenous compounds. Surplus nitrate readily moves with the ground water (US EP A, 1987; van Duijvenboden & Matthijsen, 1989). Under aerobic conditions, nitrate percolates in large quantities into the aquifer because of the small extent to which degradation or denitrification occurs. Under anaerobic conditions, nitrate may be denitrified or degraded almost completely to nitrogen. The presence of high or low water tables, the amount of rainwater, the presence of other organic material, and other physicochemical properties are also important in determining the fate of nitrate in soil (van Duijvenboden & Loch, 1983 ). In surface water, nitrification and denitrification may also occur, depending on the temperature and pH. The uptake of nitrate by plants, however, is responsible for most of the nitrate reduction in surface water. Nitrogen compounds are formed in the air by lightning or discharged into it from industrial processes, motor vehicles, and intensive agriculture. Nitrate is present in air primarily as nitric acid and inorganic aerosols, as well as nitrate radicals and organic gases or aerosols. These are removed by wet and dry deposition.
2. ANALYTICAL METHODS
Spectrometric techniques are used for the determination of nitrate in water. Detection limits range from 0.01 to 1 mg/litre (ISO, 1986, 1988). A molecular absorption spectrometric method is available for the determination of nitrite in potable water, raw water. and wastewater. The limit of detection lies within the range of0.005-0.0l mg/litre (ISO, 1984). A continuous-flow spectrometric method for the determination of nitrite, nitrate, or the sum of both in various types of water is suitable at concentrations ranging from 0.05 to 5 mg/litre for nitrite and from 1 to 100 mg/litre for nitrite/nitrate, both in the undiluted sample (ISO, 1996). NITRATE AND NITRITE 65
Nitrate and nitrite can also be determined in water by liquid chromatography, down to a level of 0.1 mg/litre for nitrate and 0.05 mg/litre for nitrite (ISO, 1992).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Atmospheric nitrate concentrations ranging from 0.1 to 0.4 j.tg/m3 have been reported, the lowest concentrations being found in the South Pacific (Prospero & Savoie, 1989). Higher concentrations ranging from 1 to 40 j.tg/m3 have also been 3 reported, with annual means of 1-8 j.tg/m • Mean monthly nitrate concentrations in air in the Netherlands range from 1 to 14 j.tg/m3 (Janssen et al., 1989). Indoor nitrate aerosol concentrations of 1.1-5.6 j.tg/m3 were found to be related to outdoor concentrations (Yocom, 1982).
3.2 Water
Concentrations of nitrate in rainwater of up to 5 mg/litre have been observed in industrial areas (van Duijvenboden & Matthijsen, 1989). In rural areas, concentrations are somewhat lower. The nitrate concentration in surface water is normally low (0-18 mg/litre) but can reach high levels as a result of agricultural runoff, refuse dump runoff, or contamination with human or animal wastes. The concentration often fluctuates with the season and may increase when the river is fed by nitrate-rich aquifers. Nitrate concentrations have gradually increased in many European countries in the last few decades and have sometimes doubled over the past 20 years. In the United Kingdom, for example, an average annual increase of 0.7 mg/litre has been observed in some rivers (Young & Morgan-Jones, 1980). The natural nitrate concentration in groundwater under aerobic conditions is a few milligrams per litre and depends strongly on soil type and on the geological situation. In the USA, naturally occurring levels do not exceed 4-9 mg/litre for nitrate and 0.3 mg/litre for nitrite (US EP A, 1987). As a result of agricultural activities, the nitrate concentration can easily reach several hundred milligrams per litre (WHO, 1985b ). For example, concentrations of up to 1500 mg/litre were found in groundwater in an agricultural area oflndia (Jacks & Sharrna, 1983). In the USA. nitrates are present in most surface water and groundwater supplies at levels below 4 mg/litre, with levels exceeding 20 mg/litre in about 3% of surface waters and 6% of groundwaters. In 1986, a nitrate concentration of 44 mg/litre ( 10 mg of nitrate-nitrogen per litre) was exceeded in 40 surface water and 568 groundwater supplies. Nitrite levels were not surveyed but are expected to be much lower than 3.3 mg/litre (US EPA, 1987). The increasing use of artificial fertilizers, the disposal of wastes (particularly from animal farming). and changes in land use are the main factors responsible for the progressive increase in nitrate levels in groundwater supplies over the last 20 years. In Denmark and the Netherlands, for example, nitrate concentrations are increasing by 0.2-1.3 mg/litre per year in some areas (WHO, 1985b). Because ofthe 66 INORGANIC CONSTITUENTS delay in the response of groundwater to changes in soil, some endangered aquifers have not yet shown the increase expected from the increased use of nitrogen fertilizer or manure. Once the nitrate reaches these aquifers, the aquifers will remain contaminated for decades, even if there is a substantial reduction in the nitrate loading of the surface. In most countries. nitrate levels in drinking-water derived from surface water do not exceed 10 mg/litre. In some areas, however, concentrations are higher as a result of runoff and the discharge of sewage effluent and certain industrial wastes. In 15 European countries, the percentage of the population exposed to nitrate levels in drinking-water above 50 mg/litre ranges from 0.5 to 10% (WHO, 1985b; ECETOC, 1988); this corresponds to nearly 10 million people. Individual wells in agricultural areas throughout the world especially contribute to nitrate-related toxicity problems, and nitrate levels in the well-water often exceed 50 mg/litre. Nitrite levels in drinking-water in the Netherlands are usually below 0.1 mg/litre. In 1993, a maximum value of0.21 mg/litre was detected (RIVM, 1993). Chloramination may give rise to the formation of nitrite within the distribution system, and the concentration of nitrite may increase as the water moves towards the extremities of the system. Nitrification in distribution systems can increase nitrite levels, usually by 0.2-1.5 mg of nitrite per litre, but potentially by more than 3 mg of nitrite per litre (A WW ARF, 1995).
3.3 Food
Vegetables and cured meat are in general the main source of nitrate and nitrite in the diet. but small amounts may be present in fish and dairy products. Meat products may contain <2.7-945 mg of nitrate per kg and <0.2-6.4 mg of nitrite per kg; dairy products may contain <3-27 mg of nitrate per kg and <0.2-1.7 mg of nitrite per kg (ECETOC, 1988). Several vegetables and fruits contain 200-2500 mg of nitrate per kg (van Duijvenboden & Matthijsen, 1989). The nitrate content of vegetables can be affected by processing of the food, the use of fertilizers, and growing conditions, especially the soil temperature and (day)light intensity (Gangolli et al.. 1994; WHO, 1995). Vegetables such as beetroot, lettuce, radish, and spinach often contain nitrate concentrations above 2500 mg/kg, especially when they are cultivated in greenhouses. Nitrite levels in food are very low (generally well below 10 mg/kg) and rarely exceed 100 mg/kg. Exceptions to this are vegetables that have been damaged, poorly stored, or stored for extended periods as well as pickled or fermented vegetables. In such circumstances, nitrite levels of up to 400 mg/kg have been found (WHO, 1995).
3.4 Estimated total exposure and relative contribution of drinking-water
Air pollution appears to be a minor source of nitrate exposure. In general, vegetables will be the main source of nitrate intake when nitrate levels in drinking water are below 10 mg/litre (Chilvers et al., 1984; US EP A. 1987; ECETOC, 1988). When nitrate levels in drinking-water exceed 50 mg/litre, drinking-water will be the major source of total nitrate intake, especially for bottle-fed infants. In the NITRATE AND NITRITE 67
Netherlands, the average population exposure is approximately 140 mg of nitrate per day (including the nitrate in drinking-water). The contribution of drinking-water to nitrate intake is usually less than 14%. For bottle-fed infants, daily intake from formula made with water containing 50 mg of nitrate per litre would average about 8.3-8.5 mg of nitrate per kg of body weight per day. The mean dietary intakes determined by the duplicate portion technique (WHO, 1985a) range from 43 to 131 mg of nitrate per day and from 1.2 to 3 mg of nitrite per day. Estimates of the total nitrate intake based on the proportion of nitrate excreted in the urine (Bartholomew et al., 1979) range from 39 to 268 mg/day, the higher values applying to vegetarian and nitrate-rich diets (ECETOC, 1988). The estimated total daily intake of nitrate ranged in the United Kingdom from 50 to 81 mg per person (Bonnell, 1995; Schuddeboom, 1995), in Denmark from 70 to 172 mg per person (Bonnell, 1995), and in Germany from 70 to 110 mg per person (Bonnell, 1995). According to the US EP A, the average nitrate intake from food is approximately 40-100 mg/day for males. The daily nitrite intake ranges from 0.3 to 2.6 mg/day, primarily from cured meat (NAS, 1981). Nitrite present in cured meat has been reported to account for up to 70% of total dietary intake of this substance, depending on the intake of such meat and the origin and type of cured meat consumed. Mean dietary nitrite intake from all food sources has been reported to range from <0.1 to 8.7 mg of nitrite per person per day for European diets (WHO, 1995).
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
4.1 Absorption, distribution, and elimination
Ingested nitrate is readily and completely absorbed from the upper small intestine. Nitrite may be absorbed directly from both the stomach and the upper small intestine. Part of the ingested nitrite reacts with gastric contents prior to absorption. Nitrate is rapidly distributed throughout the tissues. Approximately 25% of ingested nitrate is actively secreted into saliva, where it is partly (20%) reduced to nitrite by the oral microflora; nitrate and nitrite are then swallowed and re-enter the stomach. Bacterial reduction of nitrate may also take place in other parts of the human gastrointestinal tract, but not normally in the stomach; exceptions are reported in humans with low gastric acidity, such as artificially fed infants, certain patients in whom hydrochloric acid secretion is slower than normal, or patients using antacids (Colbers et al., 1995). In rats, active secretion and reduction of nitrate in saliva are virtually absent (Walker, 1995). Total nitrate reduction in rats is probably less than in humans. Absorbed nitrite is rapidly oxidized to nitrate in the blood. Nitrite in the 2 bloodstream is involved in the oxidation of Hb to metHb: the Fe + present in the 1 haem group is oxidized to its Fe + form, and the remaining nitrite binds firmly to this 1 oxidized haem. The Fe + form does not allow oxygen transport, owing to the strong binding of oxygen (Jaffe, 1981; US National Research Council, 1995). Therefore, methaemoglobinaemia can lead to cyanosis. 68 INORGANIC CONSTITUENTS
Nitrite has been shown to cross the placenta and cause the formation of fetal methaemoglobinaemia in rats. It may react in the stomach with nitrosatable compounds (e.g. secondary and tertiary amines or amides in food) to form N-nitroso compounds. Such endogenous nitrosation has been shown to occur in human as well as animal gastric juice both in vivo and in vitro, mostly at higher pH values, when both nitrite and nitrosatable compounds were present simultaneously (Shephard, 1995; WHO, 1996). The major part of the ingested nitrate is eventually excreted in urine as nitrate, ammonia, or urea, faecal excretion being negligible. Little nitrite is excreted (WHO, 1985b; ICAIR Life Systems, Inc., 1987; Speijers et al., 1989).
4.2 Endogenous synthesis of nitrate and nitrite
The excess nitrate excretion that has often been observed after low nitrate and nitrite intake originates from endogenous synthesis, which amounts, in normal healthy humans, to 1 mmol/day on average, corresponding to 62 mg of nitrate per day or 14 mg of nitrate-nitrogen per day. Gastrointestinal infections greatly increase nitrate excretion, as a result, at least in part, of increased endogenous (non-bacterial) nitrate synthesis, probably induced by activation of the mammalian reticuloendothelial system (WHO, 1985b, 1996; Speijers et al., 1989; Wishnok et al., 1995). This endogenous synthesis of nitrate complicates the risk assessment of nitrate. Increased endogenous synthesis of nitrate, as reported in animals with induced infections and inflammatory reactions, was also observed in humans. Infections and non-specific diarrhoea played a role in the increased endogenous synthesis of nitrate (Tannenbaum et al., 1978; Green et al., 1981; Hegesh & Shiloah, 1982; Bartholomew & Hill, 1984; Lee et al., 1986; Gangolli et al., 1994 ). These observations are all consistent with the induction of one or more nitric oxide synthases by inflammatory agents, analogous to the experiments described in animals and macrophages. This induction in humans has been difficult to demonstrate directly, but administration of [' 5N]arginine to two volunteers resulted in the incorporation of 15N into urinary nitrate in both individuals, confirming the arginine~nitric oxide pathway in humans (Leaf et al., 1989). Nitrate excretion in excess of nitrate intake by humans was reported in 1916, but this result remained obscure until the end of the 1970s, when it was re-examined because of the potential involvement of nitrate in endogenous nitrosation. A relatively constant daily production of about 1 mmol of nitrate was confirmed. A major pathway for endogenous nitrate production is conversion of arginine by macro phages to nitric oxide 'lnd citrulline, followed by oxidation of the nitric oxide to nitrous anhydride and then reaction of nitrous anhydride with water to yield nitrite. Nitrite is rapidly oxidized to nitrate through reaction with Hb. In addition to macrophages. many cell types can form nitric oxide, generally from arginine. Under some conditions, bacteria can form nitric oxide by reduction of nitrite. These processes can lead to nitrosation of amines at neutral pH, presumably by reaction with nitrous anhydride. The question of whether the arginine~nitrate pathway can be associated with increased cancer risk via exposure to N-nitroso compounds remains open. Nitric oxide is mutagenic towards bacteria and human cells in culture; it NITRATE AND NITRITE 69 causes DNA strand breaks, deamination (probably via nitrous anhydride), and oxidative damage; and it can activate cellular defence mechanisms. In virtually all of these cases. the biological response is paralleled by the final nitrate levels. Thus, while endogenously formed nitrate may itself be of relatively minor toxicological significance, the levels of this substance may potentially serve as indicators for those potentially important nitric oxide-related processes that gave rise to it (Wishnok et al., 1995). As mentioned above, both in vitro and in vivo studies showed that nitrate can be reduced to nitrite by bacterial and mammalian metabolic pathways. via the widespread nitrate reductase (Gangolli et al.. 1994). In humans, saliva is the major site for the formation of nitrite. About 5% of dietary nitrate is converted to nitrite (Spiegelhalder et al., 1976; Eisenbrand et al., 1980; Waiters & Smith, 1981; Gangolli et al.. 1994 ). A direct correlation between gastric pH, bacterial colonization, and gastric nitrite concentration has been observed in healthy people with a range of pH values from I to 7 (Mueller et al., 1983, 1986). In individuals with gastrointestinal disorders and achlorhydria, high levels of nitrite can be reached (6 mg/litre) (Rudell et al., 1976. 1978; Dolby et al.. 1984). The situation in neonates is not clear. It is commonly accepted that infants younger than 3 months may be highly susceptible to gastric bacterial nitrate reduction, as the pH is generally higher than in adults (Speijers et al., 1989). However. the presence of acid-producing lactobacilli in the stomach may be important, as these organisms do not reduce nitrate and may maintain a pH low enough to inhibit colonization by nitrate-reducing bacteria (Bartholomew et al.. 1980). As mentioned above, nitrite may also be produced via the arginine-nitric oxide pathway but would be undetectable because of the rapid oxidation to nitrate. One possible example of nitrite production by this route, however, is the methaemoglobinaemia observed in infants suffering from diarrhoea (Gangolli et al., 1994 ).
5. EFFECTS ON LABORATORY ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
The acute oral toxicity of nitrate to laboratory animals is low to moderate.
LD50 values of 1600-9000 mg of sodium nitrate per kg of body weight have been reported in mice, rats, and rabbits. Ruminants are more sensitive to the effects of nitrate as a result of high nitrate reduction in the rumen; the LD 50 for cows was 450 mg of sodium nitrate per kg of body weight. Nitrite is more toxic than nitrate:
LD50 values of 85-220 mg of sodium nitrite per kg of body weight have been reported for mice and rats (Speijers et al., 1989: WHO, 1996).
5.2 Short-term exposure
In a 13-week study in which nitrite was given to rats in drinking-water, a dose-related hypertrophy of the adrenal zona glomerulosa was observed at all dose levels (100, 300, 1000, or 3000 mg of potassium nitrite per litre). Increased metHb 70 INORGANIC CONSTITUENTS levels were seen only in the highest dose group (Til et al., 1988). WHO (1995) concluded that the NOEL in this study was 100 mg of potassium nitrite per litre (equivalent to 5.4 mg/kg of body weight per day expressed as nitrite ion), because the hypertrophy seen at this dose was not significantly different from the controls. An additional 13-week study in which nitrite was also given in drinking water, including lower doses of potassium nitrite and two doses of sodium nitrite (equimolar to the low and high doses of potassium nitrite), confirmed the finding of the adrenal hypertrophy of the zona glomerulosa for potassium nitrite and also revealed hypertrophy in the animals given sodium nitrite. The NOEL for the adrenal hypertrophy of the zona glomerulosa was 50 mg of potassium nitrite per litre (equivalent to 5 mg of potassium nitrite per kg of body weight per day) (Kuper & Til, 1995). Since then, studies designed to clarify the etiology of this hypertrophy and to establish its significance for human health have been partly performed and are currently in progress. The studies already performed confirmed the adrenal hypertrophy in another rat strain. However, the effects were seen only at higher dose levels. It was also seen that the hypertrophy was still present after a 30-day recovery period but had disappeared after a 60-day recovery period. At present, the mechanism of hypertrophy induced by nitrite is not clear (Boink et al., 1995). A variety of experimental and field studies in different mammals identified inorganic nitrate as a goitrogenic agent. It could be shown in rats by oral and parenteral application of potassium nitrate (Wyngaarden et al., 1953; Bloomfield et al., 1961; Alexander & Wolff, 1966; Wolff, 1994), of nitrate in hay (Lee et al., 1970), and of sodium nitrate (Horing et al., 1985; Seffner & Horing, 1987a,b). Antithyroid effects of nitrate were also found in sheep (Bloomfield et al., 1961) and in pigs by application of potassium nitrate (Jahreis et al., 1986, 1987). Furthermore, nitrate was goitrogenic to livestock: pigs (Korber et al., 1983), cattle (Korber et al., 1983, 1985), sheep (Korber et al., 1983), and goats (Prassad, 1983).
5.3 Long-term exposure
The only observed effect of nitrate in rats after 2 years of oral administration was growth inhibition; this was seen at dietary concentrations of 5% sodium nitrate and higher. The NOEL in this study was 1%, which corresponds to 3 70 mg of nitrate per kg of body weight per day (Speijers et al., 1989; WHO, 1996). A more recent long-term study was solely a carcinogenicity study, in which the highest dose levels of 1820 mg of nitrate per kg of body weight per day did not show carcinogenic effects. However, this level could not be considered as a NOEL, because complete histopathological examinations were not performed (WHO, 1996). One of the long-term effects of nitrite reported in a variety of animal species is vitamin A deficiency; this is probably caused by the direct reaction of nitrite with the vitamin. The most important effects reported in long-term animal studies were an increase in metHb level and histopathological changes in the lungs and heart in rats receiving nitrite in drinking-water for 2 years. The LOAEL, which gave a metHb level of 5%, was 1000 mg of sodium nitrite per litre; the NOEL was 100 mg of sodium nitrite per litre, equivalent to 10 mg of sodium nitrite per kg of body weight per day (or 6.7 mg/kg of body weight per day expressed as nitrite ion) (Speijers et al., 1989). NITRATE AND NITRITE 71
5.4 Reproductive and developmental toxicity
The reproductive behaviour of guinea-pigs was impaired only at very high nitrate concentrations (30 000 mg of potassium nitrate per litre); the NOEL was 10 000 mg/litre (Speijers et al., 1989; WHO, 1996). In rabbits, dose levels of 250 or 500 mg of nitrate per litre administered during 22 weeks revealed no detrimental effects on reproductive performance after successive gestations. In sheep and cattle, no abortions were observed at dose levels causing severe methaemoglobinaemia (Speijers et al., 1989; WHO, 1996). Nitrite appeared to cause fetotoxicity in rats at drinking-water concentrations equivalent to 200 and 300 mg of sodium nitrite per kg of body weight per day, causing increased maternal metHb levels. However, after similar doses in feed in other studies, no embryotoxic effects were observed in rats. In a reproductive toxicity study in guinea-pigs at dose levels of 0, 50, or 60 mg of sodium nitrite per kg of body weight per day given by subcutaneous injection, fetal death followed by abortion occurred at the highest dose level. Teratogenic effects were not observed in reported studies in mice and rats (Speijers et al., 1989; WHO, 1996).
5.5 Mutagenicity and related end-points
Nitrate is not mutagenic in bacteria and m=malian cells in vitro. Chromosomal aberrations were observed in the bone marrow of rats after oral nitrite uptake, but this could have been due to exogenous N-nitroso compound formation. Nitrite is mutagenic. It causes morphological transformations -in in vitro systems; mutagenic activity was also found in a combined in vivo-in vitro experiment with Syrian hamsters. The results of in vivo experiments were controversial (Speijers et al., 1989; WHO, 1996).
5.6 Carcinogenicity
Nitrate is not carcinogenic in laboratory animals. Some studies in which nitrite was given to mice or rats in the diet showed slightly increased tumour incidence; however, the possibility of exogenous N-nitroso compound formation in these studies could not be excluded. In studies in which high levels of nitrite and simultaneously high levels of nitrosatable precursors were administered, increased tumour incidence was seen (Speijers et al., 1989; WHO, 1996). These types of tumours could be characteristic of the presumed corresponding N-nitroso compound endogenously formed. However, this increase in tumour incidence was seen only at extremely high nitrite levels, in the order of 1000 mg/litre of drinking-water. At lower nitrite levels, tumour incidence resembled those of control groups treated with the nitrosatable compound only. On the basis of adequately performed and reported studies, it may be concluded that nitrite itself is not carcinogenic to animals (Speijers et al., 1989; WHO, 1996). 72 INORGANIC CONSTITUENTS
6. EFFECTS ON HUMANS
6.1 Methaemoglobinaemia
The toxicity of nitrate to humans is mainly attributable to its reduction to nitrite. The major biological effect of nitrite in humans is its involvement in the oxidation of normal Hb to metHb, which is unable to transport oxygen to the tissues. The reduced oxygen transport becomes clinically manifest when metHb concentrations reach 10% of normal Hb concentrations and above; the condition, called methaemoglobinaemia, causes cyanosis and, at higher concentrations, asphyxia. The normal metHb level in humans is less than 2%; in infants under 3 months of age, it is less than 3%. The Hb of young infants is more susceptible to metHb formation than that of older children and adults. This higher susceptibility is believed to be the result of the large proportion offetal Hb still present in the blood of these infants. This fetal Hb is more easily oxidized to metHb. In addition, there is a deficiency in the metHb reductase responsible for the reduction of metHb back to Hb. The net result is that a dose of nitrite causes a higher metHb formation in these infants than in adults. With respect to exposure to nitrate, these young infants are also more at risk because of a relatively high intake of nitrate and, under certain conditions, a higher reduction of nitrate to nitrite by gastric bacteria due to the low production of gastric acid (Speijers et al., 1989; WHO, 1996). The higher reduction of nitrate to nitrite in the young infants is not quantified very well, and it appears that gastrointestinal infections increase the risk of higher yield of nitrite and thus a higher metHb formation (ECETOC, 1988; Speijers et al., 1989; Moller, 1995; Schuddeboom, 1995; WHO, 1996). Other groups especially susceptible to metHb formation include pregnant women and people deficient in glucose-6-phosphate dehydrogenase or metHb reductase (Speijers et al., 1989).
6.1.1 Adults and children above the age of 3 months
Cases of methaemoglobinaemia have been reported in adults consuming high doses of nitrate by accident or as a medical treatment. Fatalities were reported after single intakes of 4-50 g of nitrate (equivalent to 67-833 mg of nitrate per kg of body weight) (Speijers et al., 1989; WHO, 1996), many of which occurred among special risk groups in whose members gastric acidity was reduced. Toxic doses - with metHb formation as a criterion for toxicity- ranged from 2 to 9 g (equivalent to 33-150 mg of nitrate per kg of body weight) (WHO, 1996). In a controlled study, an oral dose of 7-10.5 g of ammonium nitrate and an intravenous dose of 9.5 g of sodium nitrate did not cause increased metHb levels in adults, although vomiting and diarrhoea occurred (Speijers et al., 1989; WHO, 1996). Accidental human into xi cations have been reported as a result of the presence of nitrite in food. The oral lethal dose for humans was estimated to range from 33 to 250 mg of nitrite per kg of body weight, the lower doses applying to children and elderly people. Toxic doses giving rise to methaemoglobinaemia ranged from 0.4 to 200 mg/kg of body weight (WHO, 1996). NITRATE AND NITRITE 73
Another source of information with respect to nitrite toxicity in humans is the use of sodium nitrite as medication for vasodilation or as an antidote in cyanide poisoning. Doses of 30-300 mg per person (equivalent to 0.5-5 mg/kg of body weight) were reported not to cause toxic effects (WHO, 1996). Few cases of methaemoglobinaemia have been reported in older children. A correlation study among children aged 1-8 years in the USA showed that there was no difference in metHb levels between 64 children consuming high-nitrate well water (22-111 mg of nitrate-nitrogen per litre) and 38 children consuming low nitrate water (<1 0 mg of nitrate-nitrogen per litre). These concentrations correspond to 100-500 and <44 mg of nitrate per litre, respectively. All the metHb levels were within the normal range, suggesting that older children are relatively insensitive to the effects of nitrate (Craun et al., 1981 ).
6.1.2 Infants under 3 months of age
Cases of methaemoglobinaemia related to low nitrate appear to be restricted to infants. In infants under the age of 3 months, the conversion of nitrate to nitrite and metHb formation are high, as discussed above. Gastrointestinal disturbances play a crucial role, the reduction of nitrate to nitrite in the stomach being enhanced by bacterial growth at the high pH in the stomach of these infants. Toxic effects can therefore be induced at a much lower dose of nitrate than in adults. According to Com~ & Breimer (1979), assuming an 80% reduction of nitrate to nitrite in these young infants, the toxic dose ranged from 1.5 to 2.7 mg of nitrate per kg of body weight, using 10% formation of metHb as a toxicity criterion. However, in reported cases of methaemoglobinaemia, the amounts of nitrate ingested were higher: 37.1-108.6 mg/kg of body weight, with an average of 56.7 mg of nitrate per kg of body weight (WHO, 1996). In studies in which a possible association between clinical cases of infantile methaemoglobinaemia or subclinically increased metHb levels and nitrate concentrations in drinking-water was investigated, a significant relationship was usually found, most clinical cases (97.7%) occurring at nitrate levels of 44.3-88.6 mg/litre or higher (Walton, 1951; WHO, 1996), and almost exclusively in infants under 3 months of age (Walton, 1951). Some cases of infant methaemoglobinaemia have indeed been described in which increased endogenous nitrate (nitrite) synthesis as a result of gastrointestinal infection appeared to be the only causative factor (WHO, 1996). As most cases of infantile methaemoglobinaemia reported in the literature have been associated with the consumption of private and often bacterially contaminated well-water, the involvement of infections is highly probable. Most of these studies may be therefore less suitable from the point of view of the quantitative assessment of the risk of nitrate intake for healthy infants. On the other hand, bottle-fed infants under 3 months of age have a high probability of developing gastrointestinal infections because of their low gastric acidity, which is another important reason to consider these infants as a risk group. 74 INORGANIC CONSTITUENTS
6.2 Carcinogenicity
Nitrite was shown to react with nitrosatable compounds in the human stomach to form N-nitroso compounds. Many of these N-nitroso compounds have been found to be carcinogenic in all the animal species tested, although some of the most readily formed compounds, such as N-nitrosoproline, are not carcinogenic in humans. The N-nitroso compounds carcinogenic in animal species are probably also carcinogenic in humans. However, the data from a number of epidemiological studies are at most only suggestive. The endogenous formation of N-nitroso compounds is also observed in several animal species, if relatively high doses of both nitrite and nitrosatable compounds are administered simultaneously. Thus, a link between cancer risk and endogenous nitrosation as a result of high intake of nitrate and/or nitrite and nitrosatable compounds is possible (Speijers et al., 1989; WHO, 1996). Several reviews of epidemiological studies have been published; most of these studies are geographical correlation studies relating estimated nitrate intake to gastric cancer risk. The US National Research Council found some suggestion of an association between high nitrate intake and gastric and/or oesophageal cancer (NAS, 1981). However, individual exposure data were lacking, and several other plausible causes of gastric cancer were present. In a later WHO review (WHO, 1985b), some of the earlier associations appeared to be weakened following the introduction of individual exposure data or after adjustment for socioeconomic factors. No convincing evidence was found of an association between gastric cancer and the consumption of drinking-water in which nitrate concentrations of up to 45 mg/litre were present. No firm evidence was found at higher levels either, but an association could not be excluded because of the inadequacy of the data available. More recent geographical correlation and occupational exposure studies also failed to demonstrate a clear relationship between nitrate intake and gastric cancer risk, although these studies were well designed. A case-control study in Canada, in which dietary exposure to nitrate and nitrite was estimated in detail, showed that exogenous nitrite intake, largely from preserved meat, was significantly associated with the risk of developing gastric cancer (ECETOC, 1988). On the other hand, case-control studies based on food frequency questionnaires tend to show a protective effect of the estimated nitrate intake on gastric cancer risk. Most likely this is due to the known strong protective effect of vegetables and fruits on the risk of gastric cancer (Moller, 1995; WHO, 1996). Studies that have assessed the effect of nitrate from sources other than vegetables, such as the concentration in drinking-water or occupational exposure to nitrate dusts, have not shown a protective effect against gastric cancer risk. For other types of cancer, there are no adequate data with which to establish any association with nitrite or nitrate intake (Gangolli et al., 1994; Moller, 1995; WHO, 1996). It has been established that the intake of certain dietary components present in vegetables, such as vitamins C and E, decreases the risk of gastric cancer. This is generally assumed to be at least partly due to the resulting decrease in the conversion of nitrate to nitrite and in the formation of N-nitroso compounds. It is possible that any effect of a high nitrate intake per se is masked in correlation studies by the NITRATE AND NITRITE 75 antagomzmg effects of simultaneously consumed dietary protective components. However, the absence of any link with cancer in occupational exposure studies is not in agreement with this theory. The known increased risk of gastric cancer under conditions of low gastric acidity could be associated with the endogenous formation of N-nitroso compounds. High mean levels of N-nitroso compounds, as well as high nitrate levels, were found in the gastric juice of achlorhydric patients, who must therefore be considered as a special risk group for gastric cancer from the point of view of nitrate and nitrite (NAS, 1981; WHO, 1985b, 1996; ECETOC, 1988; Speijers et al., 1989).
6.3 Other effects
Congenital malformations have been related to high nitrate levels in drinking-water in Australia; however, these observations were not confirmed. Other studies also failed to demonstrate a relationship between congenital malformations and nitrate intake (WHO, 1985b; ECETOC, 1988). Studies relating cardiovascular effects to nitrate levels in drinking-water gave inconsistent results (WHO, 1985b ). Possible relationships between nitrate intake and effects on the thyroid have also been studied, as it is known that nitrate competitively inhibits iodine uptake. In addition to effects of nitrate on the thyroid observed in animal studies and in livestock, epidemiological studies revealed indications for an antithyroid effect of nitrate in humans. If dietary iodine is available at an adequate range (corresponding to a daily iodine excretion of 150-300 ~-tg/day), the effect of nitrate is weak, with a tendency to zero. The nitrate effect on thyroid function is strong if a nutritional iodine deficiency exists simultaneously (Horing et al., 1991; Ho ring, 1992). Hettche (1956a,b) described an assoc1atlon between high nitrate concentrations in drinking-water and goitre incidence in 1955. As well, Horing & Schiller (1987), Sauerbrey & Andree (1988), Horing et al. (1991), Horing (1992), and van Maanen et al. (1994) found that inorganic nitrate in clrinking-water is a manifested factor of endemic goitre. A dose-response relationship could be demonstrated by Horing et al. ( 1991) (nitrate in drinking-water vs incidence of goitre) as well as by van Maanen et al. (1994) (nitrate in drinking-water vs thyroid volume). Both the experimental and epidemiological studies give the impression that nitrate in drinking-water has a stronger effect on thyroid function than nitrate in food. The differences in nitrate kinetics after ingestion through drinking-water and through food could be the cause of the difference in thyroid effects. However, no adequate studies regarding this question exist at present. Furthermore, some of the above-mentioned studies demonstrate that dietary iodine deficiency is much more effective than nitrate exposure in causing goitre. In addition to the effect of nitrite on the adrenal zona glomerulosa in rats, a study in humans indicated that sodium nitrite (0.5 mg of sodium nitrite per kg of body weight per day, during 9 days) caused a decreased production of adrenal steroids, as reflected by the decreased concentration of 17-hydroxysteroid and 17-ketosteroids in urine (Til et al., 1988; Kuper & Til, 1995). Similar results were also found in rabbits (Violante et al., 1973). Although the mechanism is not clear, 76 INORGANIC CONSTITUENTS the effects of nitrite seen in rats seem relevant for the hazard assessment for humans, unless mechanistic studies prove otherwise.
7. GUIDELINE VALUES
With respect to chronic effects, JECF A recently re-evaluated the health effects of nitrate/nitrite, confirming the previous ADI of0-3.7 mg/kg of body weight per day for nitrate ion and establishing an ADI of 0-0.06 mg/kg of body weight per day for nitrite ion (WHO, 1995). However, it was noted that these ADis do not apply to infants below the age of 3 months. Bottle-fed infants below 3 months of age are most susceptible to methaemoglobinaemia following exposure to nitrate/nitrite in drinking-water. For methaemoglobinaemia in infants (an acute effect), it was confirmed that the existing guideline value for nitrate ion of 50 mg/litre is protective. For nitrite, human data reviewed by JECF A support the current provisional guideline value of 3 mg/litre, based on induction of methaemoglobinaemia in infants. Toxic doses of nitrite responsible for methaemoglobinaemia range from 0.4 to more than 200 mg!kg of body weight. Following a conservative approach by applying the lowest level of the range (0.4 mg/kg of body weight), a body weight of 5 kg for an infant, and a drinking water consumption of 0.75 litre, a guideline value for nitrite ion of 3 mg/litre (rounded figure) can be derived. The guideline value is no longer provisional. Because of the possibility of the simultaneous occurrence of nitrite and nitrate in drinking-water, the sum of the ratios of the concentrations (C) of each to its guideline value (GV) should not exceed one, i.e.:
+ :s; 1
It seems prudent to propose a guideline value for nitrite associated with chronic exposure based on JECF A's analysis of animal data showing nitrite-induced morphological changes in the adrenals, heart, and lungs. Using JECFA's ADI of 0.06 mg/kg of body weight per day, assuming a 60-kg adult ingesting 2 litres of drinking-water per day, and allocating 10% of the ADI to drinking-water, a guideline value of 0.2 mg of nitrite ion per litre (rounded figure) can be calculated. However, owing to the uncertainty surrounding the relevance of the observed adverse health effects for humans and the susceptibility of humans compared with animals, this guideline value should be considered provisional. Because of known interspecies variation in the conversion of nitrate to nitrite, the animal model was not considered appropriate for use in human risk assessment for nitrate. Chloramination may give rise to the formation of nitrite within the distribution system, and the concentration of nitrite may increase as the water moves towards the extremities of the system. All water systems that practise chloramination should closely and regularly monitor their systems to verify disinfectant levels, microbiological quality, and nitrite levels. If nitrification is detected (e.g. reduced NITRATE AND NITRITE 77 disinfectant residuals and increased nitrite levels), steps should be taken to modify the treatment train or water chemistry in order to maintain a safe water quality. Efficient disinfection must never be compromised.
8. REFERENCES
Alexander WD, Wolff J ( 1966) ThyrOidal iodide transport. 8. Relation between transport, goitrogenic and antigo1trogenic properties of certain anions. Endocrmology, 78:581-590. A WWARF (1995) Nllrtjicatwn occurrence and control m chlorammated water systems. Denver, CO, American Water Works Association Research Foundation. Bartholomew BA, Hill MJ (1984) The pharmacology of dietary nitrate and the origm of unnary nitrate. Food chenustry and toxzcology, 22:789-795 Bartholomew B et al. ( 1979) Possible use of urmary nitrate as a measure of total nitrate intake. Proceedzngs of the Nutntwn Soczety, 38: 124A. Bartholomew BA et al. (1980) Gastnc bacteria, nitrate, nitrite and nitrosamines in patients with pernicious anaemia and in patients treated with cimetidine !ARC sczenttjic publicatwns, 31 :595-608. Bloomfield RA et al. (1961) Effect of dietary nitrate on thyroid function. Sczence, 134:1961 Boink BTJ, Dormans JAMA, Speijers GJA (1995) The role of mtrite and/or nitrate in the etiology of the hypertrophy of the adrenal zona glomerulosa of rats. In: Health aspects of nitrate and its metabolztes (partzcularly nitrite). Proceedings of an internatwnal workshop, B!lthoven (Netherlands), 8-10 November 1994. Strasbourg, Council of Europe Press, pp. 213-228. Bonnell A (1995) Nitrate concentrations in vegetables. In: Heaith aspects of nztrate and lis metabolites (partzcularly mtnte) Proceedings of an mternatwnal workshop, Bzlthoven (Netherlands), 8-10 November 1994 Strasbourg, Council of Europe Press, pp 11-20. Chilvers C, Insk1p H. Caygill C (1984) A survey of dietary nitrate in well-water users. International journal of epzdemwlogy, 13:324-331. Colbers EPH et al. (1995) A pzlot study to mvestzgate nllrate and mtnte kinetics 111 healthy volunteers wllh both normal and artificiallv increased gastrzc pH after sodium mtrale ingestion. Bilthoven, Rijksintituut voor de Volksgezondheid en Milieuhygiene (National Institute of Public Health and Environmental Protection) {RIVM Report No. 235802001 ). Corn~ WJ, Breimer T (1979) Nztrate and mtrzte zn vegetables Wagenmgen, Centre for Agricultural Publishmg DocumentatiOn (Literature Survey No. 39). Craun GF, Greathouse DG, Gunderson DH (1981) Methaemoglobin levels in young children consuming high nitrate well water in the United States. lnternatwnal JOUrnal of epzdemwlogy. 10·309-317. Dolby JM et al. ( 1984) Bacterial colonizatiOn and n1tnte concentration in the achlorhydric stomachs of patients With pnmary hypogammaglobulinaemia or classical pernicious anaemia. Scandmavtan Journal ofgastroenterology, 19: I 05-110. ECETOC ( 1988) Nil rate and drznkmg water. Brussels, European Chemical Industry Ecology and Toxicology Centre (Technical Report No. 27). Eisenbrand G et al. (1980) Carcinogenic1ty of N-nJtroso-3-hydroxypyrrolidine and dose-response study With N-nitrosopiperidme in rats. !ARC scientific pub/icatzons, 31:657-666 Gangolli SD et al. (1994) Assessment: mtrate, nitnte and N-nitroso compounds European JOUrnal of pharmacology, environmental toxicology and pharmacology sectwn, 292:1-38. Green LC et al. (1981) Nitrate biosynthesis in man. Proceedings of the Nattonal Academy ofSczences of the Umted States of Amenca, 78:7764-7768. Hegesh E, Shiloah J (1982) Blood mtrates and infantile methaemoglobmaemia. C/znica Chzmzca Acta, 125.107-115. Hettche HO (1956a) Epidemiolog1e und Atwlogie der Struma in I 00 Jahren Forschung. [Epidemiology and etiology of gOitre in 100 years of research ] Archives uber Hygiene, Baktenologie, 140.79-105. Hettche HO ( 1956b) Zur Atwlogie und Pathogenese der Struma endemica. [On the etiology and pathogenesis of endemic goitre.] Ze1tblatt Allgemezne Pathologie, 95:187-193. 78 INORGANIC CONSTITUENTS
Horing H (1992) Der Eintluss von Umweltchemicalien auf die SchilddrUse. [The influence of environmental chemicals on the thyroid.) Bundesgesundheltsblatt, 35: 194-197. Horing H, Schiller F (1987) Nitrat und Schilddrtise- Ergebnisse epidemiologischer Untersuchungen. [Nitrate and thyroid; results of epidemiological studies.) Schriften Reihe for Gesundheit und Umwelt, Suppl. I :38-46. Horing H, Nagel M, Haerting J (1991) Das nitratbedingte Strumarisiko in einem Endemiegebiet. [The nitrate-dependent endemic thyroid areas.] In: Oberla K, Rienhoff 0, Victor N, eds. Quanfltaflve Methoden in der Epidemwlogie. Berlin, l. Guugenmoos-Holzmann, pp. 147- 153 (Medizinische Jnformatik und Statistik, 72). Horing H et al (1985) Zum Einfluss subchronischer Nitratapplikation mit dem Trinkwasser auf die Schilddrtise der Ratte (Radiojodtest). [The influence of subchronic nitrate administration in drinking-water on the thyroid.) Schriften Reihefur Gesundheit und Umwelt, 1:1-15. ICAIR Life Systems, Inc. (1987) Drmkmg water cnteria document on mtrate/nitrite. Washington, DC, US Environmental Protection Agency, Office of Drinking Water. ISO (1984) Water quality- Determmation of mtnte- Molecular absorption spectrometric method. Geneva, International Organization for Standardization (ISO 6777/1-1984 (E)). ISO (1986) Water quality- determinatwn ofmtrate- Part 1: 2,6-Dimethylphenol spectrometric method, Part 2 · 4-Fiuoropheno/ spectrometric method after disttllation. Geneva, International Organization for Standardization (ISO 7890-1,2: 1986 (E)). ISO (1988) Water quality - determmation of nitrate - Part 3· Spectrometric method using sulfosalicylic acid Geneva, International Organization for Standardization (ISO 7890- 3:1988 (E)). ISO (1992) Water quality- Determmation of dissolved fluoride, chlonde, nitrite, orthophosphate, brom1de, mtrate and sulfate using liqwd chromatography of ions Geneva, International Orgamzation for Standardization (ISO I 0304-1:1992 (E)). ISO (1996) Water quality - Determination of mtrite nitrogen and mtrate nitrogen and the sum of both by flow anaZvsis (contmuous flow analysis and flow in;ection analysis) Geneva, International Organization for Standardization (ISO 13395:1996 (E)). Jacks G, Sharma VP (1983) Nitrogen circulation and nitrate in ground water in an agricultural catchment in southern India. Environmental geology, 5(2):61-64. Jaffe ER ( 1981) Methaemoglobinaemia. Climcal haematology, I 0:99-122. Jahreis Get al. (1986) Effect of chronic dietary nitrate and different iodine supply on porcine thyroid function, somatomedin-C-level and growth. Experimental and clinical endocnnology, Le1p=1g, 88:242-248. Jahreis G et al. (1987) Growth impairment caused by dietary nitrate intake regulated via hypothyroidism and decreased somatomedin. Endocrmologw Expenmentalis Bratislava, 21:171-180 Janssen LHJM, Visser H, Roemer FG ( 1989) Analysis of large scale sulphate, nitrate, chloride and ammonium concentrations in the Netherlands using an aerosol measuring network. Atmosphenc environment, 23(12):2783-2796. Korber R, Groppel F, Leirer R (1983) Untersuchungen zum Jod- und Schilddrtisenstoffwechsel bei Ktihen und Schafen unter experimenteller Nitratbelastung. [Research on iodine and thyroid metabolism in cows and sheep under experimental nitrate exposure.)ln: Anka M et al., eds. 4. Spurenelementensymposmm der Karl-Marx Universitat Le1pz1g, Leipzig, pp. 178-186. Korber R, Rossow N, Otta J (1985) Beitrag zum Jodmangelsyndrom der Landwirtschaftlichen Nutztiere Rind, Schaf und Schwein. [The addition to iodine-deficiency syndrome of cow, sheep and pig.) Monatshejiefur Veterinaermedizin, 40:220-224. Kuper F, Til HP (1995) Subchronic toxicity experiments with potassium nitrite in rats. In: Health aspects of nitrate and its metabolttes (particularly nitrite). Proceedings of an internatwnal workshop, Bdthoven (Netherlands), 8-10 November 1994. Strasbourg, Council of Europe Press, pp. 195-212 Leaf CD, Wishnok JS, Tannenbaum SR ( 1989) L-arginine is a precursor for nitrate biosynthesis in humans. Bwchem1cal and biophysical research commumcatwns, 163:1032-1037. Lee C, Weiss R, Horvath DJ (1970) Effects of nitrogen fertilization on the thyroid function of rats fed 40 percent orchard grass diets. Joumal ofnutntion, 100:1121-1126. NITRATE AND NITRITE 79
Lee K et al. (1986) Nitrate, nitrite balance and de novo synthesis of nitrate in humans consuming cured meat. Amencan ;ournal ofclinical nutrition, 44:188-194. Mi:iller H (1995) Adverse health effects of nitrate and its metabolites: epidemiological studies in humans. In: Health aspects of mtrate and its metabolites (particularly nitrite). Proceedings of an mternational workshop, Bilthoven (Netherlands), 8-10 November I994. Strasbourg, Council of Europe Press, pp. 255-268. Mueller RL et al. ( 1983) [Endogenous synthesis of carcinogenic N-nitroso compounds: bacterial flora and nitrite formation in the healthy human stomach.] Zentralblatt fur Bakteriologie, Mzkrobiologie, und Hygzene B, 178:297-315 (in German). Mueller RL et al. (1986) Nitrate and nitrite in normal gastric juice. Precursors of the endogenous N nitroso compound synthesis. Oncology, 43:50-53. NAS (1981) The health effects of mtrate, nitnte, and N-nitroso compounds. Part I of a two-part study by the Committee on Nitrite and Alternative Curing Agents in Food. Report by the US National Research Council, National Academy of Sciences. Washington, DC, National Academy Press. Prassad J (1983) Effect of high nitrate diet on thyroid glands in goats. Indian journal of animal sciences (New Delhi), 53:791-794. Prospero JM, Savoie DL (1989) Effect of continental sources of nitrate concentrations over the Pacific Ocean. Nature, 339(6227):687--689. RIVM (1993) Handhaving Milieuwetten I99511997. De kwalitelt van het drinkwater in Nederland in 1993. [Maintenance of the envzronmentallaw 199511997 The quality of the drinking water in the Netherlands in I993.] Bilthoven, Rijksintituut voor de Volksgezondheid en Milieuhygiene (National Institute of Public Health and Environmental Protection) (RIVM Report No. 731011007). Rudell WS et al. ( 1976) Gastric JUice nitrite. a risk factor for cancer in the hypochlorhydric stomach? Lancet, 2:1037-1039. Rudell WS et al. (1978) Pathogenesis of gastric cancer in pernicious anaemia. Lancet, I :521-523. Sauerbrey G, Andree B ( 1988) Untersuchungen uber die endemische Struma und ihre Beziehung zu verschzedenen Trinkwasser m vzer Gemeinden des Bezirkes Suhl [Research on the endemic goitre and the relation to different drinking-water of four communities of Suhl.] Berlin, University of Berlin (Dissertation). Schuddeboom LJ ( 1995) A survey of the exposure to nitrate and nitrite in foods (including drinking water). In: Health aspects of mtrate and its metabolites (par/lcularly mtnte). Proceedings of an mternational workshop, Bdthoven (Netherlands), 8-10 November 1994. Strasbourg, Council of Europe Press, pp. 41-74. Seffner W, Hi:iring H ( 1987a) Zum Einfluss von subchronischer Nitratapplikation im Trmkwasser auf die Schtlddriise der Ratte-Morphologische Untersuchungen. [On the influence of subchronic nitrate application in drinking-water on the thyroids of rats.] Schrzfte fur Gesundheit und Umwelt, 3:15-32. Seffner W, Horing H ( 1987b) Zum Einfluss einer chronischen Fluorapplikation auf die durch Nitrat ausgeli:isten Schilddriisenveranderungen. [On the influence of chronic fluorine application on the thyroid changes induced by nitrate.] Schriftefur Gesundheit und Umwelt, 3:15-32. Shephard SE (1995) Endogenous formation of N-nitroso compounds in relatton to the intake of nitrate or nitrite. In: Health aspects of mtrate and its metabolites (partzcularly nitrite). Proceedmgs of an mternational workshop, Bilthoven (Nether/and;), 8-IO November I994. Strasbourg, Council of Europe Press, pp. 137-150. Speijers GJA et al. (1989) Integrated criteria document nitrate, effects Appendix to RIVM Report No. 758473012 Bilthoven, Rijksintituut voor de Volksgezondheid en Milieuhygiene (National Institute of Public Health and Environmental Protection) (RIVM Report No. A 7584730 12). Spiegelhalder 8, Eisenbrand G, Preussmann R (1976) Influence of dietary nitrate on nitrite content of human saliva: possible relevance to in vivo formation of N-nitroso compounds. Food and cosmetics toxicology, 14:545-548. Tannenbaum SR et al (1978) Nitnte and nitrate are formed by endogenous synthesis in the human intestine. Sczence, 200:1487-1489. 80 INORGANIC CONSTITUENTS
Til HP et al. ( 1988) Evaluation of the oral toxicity of potassium nitrite in a 13-week drinking-water study in rats. Food chemtsfry and toxtcology, 26(1 0)·851-859. US EPA ( 1987) Esttmated na/tonal occurrence and exposure to m/rate and nifn/e m public drinking wafer supp!tes Washington, DC, US Environmental Protection Agency, Office of Drinking Water. US National Research Council ( 1995) Ntfrate and nitrite in dnnking water. Subcommittee on Nitrate and Nitrite in Drinkmg Water, Committee on Toxicology, Board on Environmental Studies and Toxicology, CommissiOn on Life Science. Washington, DC, National Academy Press. van DUIJvenboden W, Loch JPG (1983) Nitrate in the Netherlands: a serious threat to groundwater. Aqua, 2:59--60. van Duijvenboden W, Matthijsen AJCM (1989) Integrated cnteria document nitrate. Bilthoven, Rijksintituut voor de Volksgezondheid en Milieuhygiene (Nationallnstttute of Public Health and Environmental Protection) (RlVM Report No. 758473012). van Maanen JM et al. (1994) Consumption of drinking water with high nitrate levels causes hypertrophy of the thyroid. Toxtcology lefters, 72·365-374 Violante A, Cianetti A, Ordine A (1973) Studio della funzionella cortico surrenalica in corso di intossicazione con sodia nitrio. [Adrenal cortex function during subacute poisoning with sodium nitrite.] Quaderm Sclavo dt Dtagnosttca Clinica e di Laboratorto, 9:907-920. Walker R (1995) The conversion of nitrate into nitrite in several animal species and man. In: Health aspects of mlrale and tts metabo/ites (parttcularly mtrtte) Proceedmgs of an tn/ernational workshop, Btlthoven (Netherlandv). 8-10 November 1994 Strasbourg, Council of Europe Press, pp. 1!5-123. Waiters CL, Smith PLR ( !981) The effect of water-borne nitrate on salivary nitnte. Food chemistry and toxtcology, lli:297-302 Walton G (1951) Survey of literature relating to infant methaemoglobinaemia due to nitrate contaminated water Amencan Journal ofpub!tc health, 41:986-996. WHO ( 1985a) Gutdelmes for the study of dtetary mtake of chemical conlammants. Geneva, World Health Organization (WHO Offset PublicatiOn No. 87). WHO (1985b) Health ha:::ards from mtrate m drtnkmg-water. Report on a WHO meetmg, Copenhagen. 5-9 March 1984 Copenhagen, WHO Regional Office for Europe (Environmental Health Senes No 1). WHO (1995) Evaluatton of certam food addtttves and conlammants. Geneva, World Health Organization, Jomt FAO/WHO Expert Committee on Food Additives, pp. 29-35 (WHO Technical Report Series No. 859). WHO (1996) Toxtcologtcal evalualton of certam food addtfives and contammants. Prepared by the Forty-Fourth Meeting of the Joint FAO/WHO Expert Committee on Food Additives (JECFA) Geneva, World Health Organization, International Programme on Chemical Safety (WHO Food Additives Series 35) Wishnok JS et al (1995) Endogenous fonnat10n of nitrate. In: Health aspects of nitrate and its metabolttes (parltcular~v nitrtte) Proceedmgs of an mternattonal workshop, Btlthoven (Netherlands), 8-10 November 1994. Strasbourg, Council of Europe Press, pp. 151-179. Wolff J (1994) Transport of Iodide and other anions in the thyroid. Physiology revtews, I :45-90. Wyngaarden JB, Stanbury JB, Rabb B (1953) The effects of iodide, perchlorate, thiocyanate, and nitrate administration upon IOdide concentrating mechanism of the rat thyroid. Endocrinology, 52:568-574 Yocom JE (1982) Indoor/outdoor atr quahty relationships: a cntical review. Journal of the Atr Pollutton Control Assocwtion, 32:500-606. Young CP, Morgan-Jones M (1980) A hydrogeochemical survey of the chalk groundwater of the Banstead area, Surrey, with particular reference to nitrate. Journal of the lnstt/ute of Water Engmeers and Sctenttsts, 34:213-236. URANIUM'
First draft prepared by M. Giddings Bureau of Chemical Hazards, Environmental Health Directorate Health Protection Branch, Health Canada, Ottawa, Ontario, Canada
At the time of publication of the 1993 Guidelines for drinking-water quality, adequate short- and long-term studies on the chemical toxicity of uranium were unavailable, and therefore a guideline value for uranium was not derived. Instead, it was recommended that the limits for radiological characteristics of uranium be adopted. The equivalent for natural uranium, based on these limits, is approximately 140 Jlg/litre. As new data on the chemical toxicity of uranium are now available for use in the derivation of a guideline value, the 1995 Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality recommended that uranium be included in the 1998 Addendum.
1. GENERAL DESCRIPTION
1.1 Identity
Uranium occurs naturally in the +2, +3, +4, +5, and +6 valence states, but it is most commonly found in the hexavalent form. In nature, hexavalent uranium is commonly associated with oxygen as the uranyl ion, UO/+. Naturally occurring uranium ("'U) is a mixture of three radionuclides e34U, 235U, and mu), all of which decay by both alpha and gamma emissions (Cothem & Lappenbusch, 1983; Lide, 1992-93). Natural uranium consists almost entirely of the mu isotope, with the 235U and 234U isotopes respectively comprising about 0.72% and 0.0054% of natural uranium (Greenwood & Eamshaw, 1984). Uranium is widespread in nature, occurring in granites and various other mineral deposits (Roessler et al., 1979; Lide, 1992-93).
Compound CAS no. Molecular formula Uranium 7440-61-1 u
Uranyl ethanoate 541-09-3 C4H60 6U Uranyl chloride 7791-26-6 c1p2u
Uranyl nitrate 36478-76-9 N20 8U
Uranium dioxide 1344-57-6 uo2
This revtew addresses only the chemtcal aspects of uranium toxicity. Information pertinent to the derivation of a guideline based on radiological effects is presented in the second edition of the Gu1delmes for dnnkmg-water quality. 82 INORGANIC CONSTITUENTS
1.2 Physicochemical properties (Lide, 1992-93)
Compound Melting Boiling point Density at 20°C Water solubility point (0 C} (OC) (g/cm3) (gllitre) u 1132 3818 19.0 insoluble
C4H60 6U 110 275 2.9 76.94 (decomposes)
Cl20 2U 578 (decomposes) 3200 NzOsU 60.2 118 2.8 soluble
uo2 2878 10.96 insoluble
1.3 Major uses
Uranium is used mainly as fuel in nuclear power stations, although some uranium compounds are also used as cataolysts and staining pigments (Berlin & Rudell, 1986).
1.4 Environmentalfate
Uranium is present in the environment as a result of leaching from natural deposits, release in mill tailings, emissions from the nuclear industry, the combustion of coal and other fuels, and the use of phosphate fertilizers that contain uranium.
2. ANALYTICAL METHODS
Uranium in water is most commonly measured by solid fluorimetry with either laser excitation or ultraviolet light following fusion of the sample with a pellet of carbonate and sodium fluoride (detection limit 0.1 J.tg/litre) (Kreiger & Whittaker, 1980). Sample preparation for this method is tedious, however, and there is interference from other metals. Uranium can also be determined by inductively coupled plasma mass spectrometry, which has the same detection limit (0.1 J.tg/litre) and a between-run precision of less than 6% (Boomer & Powell, 1987). Alpha spectrometry has been used for the determination of uranium in bottled waters (Gans, 1985) and environmental media (Singh & Wrenn, 1988), although the recovery is often highly variable owing to the low specific activity of natural uranium (Singh & Wrenn, 1988).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Mean levels of uranium in ambient air have been reported to be 0.02 ng/m3 in Tokyo (based on a 1979-1981 survey) (Hirose & Sugimura, 1981) and 0.076 ng/m3 URANIUM 83 in New York (based on two samples, each a composite of two weekly air filter collections, from 1985 and 1986) (Fisenne et al., 1987). On the assumption of a daily 3 3 respiratory volume of 20 m and a mean urban airborne concentration of 0.05 ng/m , the daily intake of uranium from air would be about 1 ng. Tobacco smoke (from two packages of cigarettes per day) contributes less than 50 ng of inhaled uranium per day (Lucas & Markun, 1970).
3.2 Water
In a survey of 130 sites (approximately 3700 samples) in Ontario, Canada, conducted between 1990 and 1995, the mean of the average uranium concentrations (range 0.05-4.21 )lg/litre; detection limit 0.05 )lg/litre) in treated drinking-water was 0.40 )lg/litre (OMEE, 1996). Uranium concentrations of up to 700 1-!g/litre have been found in private supplies in Canada (Moss et al., 1983; Moss, 1985). The mean concentration of uranium in drinking-water in New York City, USA, ranged from 0.03 to 0.08 1-!g/litre (Fisenne & Welford, 1986). A mean uranium concentration of 2.55 )lg/litre was reported in drinking-water from 978 sites in the USA in the 1980s (US EP A, 1990, 1991 ). In five Japanese cities, the mean level in potable water supplies was 0.9 ng/litre (Nozaki et al., 1970). The daily uranium intake from water in Finland· has been estimated to be 2.1 )lg (Kahlos & Asikainen, 1980). The daily intake from drinking-water in Salt Lake City, USA, is estimated to be 1.5 1-!g (Singh et al., 1990). On the basis of the results of the survey from Ontario (OMEE, 1996), the daily intake of uranium from drinking-water in Canada is estimated to be 0.8 /-!g.
3.3 Food
Uranium has been detected in a variety of foodstuffs. The highest concentrations are found in shellfish, and lower levels have been measured in fresh vegetables, cereals, and fish. The average per capita intake of uranium in food has been reported to be 1.3 1-!g/day (Fisenne et al., 1987) and 2-3 1-!g/day (Singh et al., 1990) in the USA and 1.5 1-!g/day in Japan (Nozaki et al., 1970). In a review of naturally occurring sources of radioactive contamination in food, dietary intakes of 238U were found to range from 12 to 45 mBq/day in several European countries, from 11 to 60 mBq/day in Japan (the higher values were found in uranium mining areas), and from 15 to 17 mBq/day in the USA. The average daily dietary intake was in the order of 20 mBq, or about 4 )lg. It was often difficult to determine whether these dietary intakes included intake from drinking-water, and it was emphasized that intake from drinking-water has sometimes been found to be equal to intake from the diet (Harley, 1988). In a study by Cheng et al. (1993), the mean uranium concentration in nine different beverages was 0. 98 1-!g/litre (range 0.26-1.65 1-!g/litre ), and the mean concentration of uranium in mineral water was 9.20 )lg/litre. Landa & Councell (1992) performed leaching studies to determine the quantity of uranium leaching from 33 glass items and two ceramic items in which uranium was used as a colouring agent. Uranium-bearing glasses leached a 84 INORGANIC CONSTITUENTS maximum of 30 f-lg of uranium per litre, whereas the ceramic-glazed items released approximately 300 000 f-lg of uranium per litre.
3.4 Estimated total exposure llfd relative contribution of drinking-water
The daily intake of uranium from each source for adults is estimated to be: air, 0.001 f-tg; drinking-water, 0.8 f-tg; food, 1.4 ).!g. Thus, the total daily intake is approximately 2.2 f-lg, or 0.037 f-tglkg of body weight for a 60-kg adult, the majority of which originates from food.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
Although ubiquitous in the environment, uranium has no known metabolic function in animals and is currently regarded as non-essential (Berlin & Rudell, 1986). Absorption of uranium from the gastrointestinal tract depends upon the solubility of the uranium compound (Berlin & Rudell, 1986), previous food consumption (Sullivan et al., 1986; La Touche et al., 1987), and the concomitant administration of oxidizing agents, such as the iron(III) ion and quinhydrone (Sullivan et al., 1986). The average human gastrointestinal absorption of uranium is 1-2% (Wrenn et al., 1985). The absorption of a uranium dose of approximately 800 mg/kg of body weight in starved female Sprague-Dawley rats increased from 0.17 to 3.3% when iron(III) (190 mg/kg of body weight) was administered simultaneously (Sullivan et al., 1986). Absorption of uranium in starved rats administered doses of uranium by gavage was reported to increase with dose; the degree of absorption ranged from 0.06 to 2.8% for doses between 0.03 and 45 mg of uranium per kg of body weight (La Touche et al., 1987). Only 0.06% of ingested uranium was absorbed in Sprague Dawley rats and New Zealand white rabbits fed ad libitum and having free access to drinking-water containing up to 600 mg of uranyl nitrate hexahydrate per litre for up to 91 days (Tracy et al., 1992). Following ingestion, uranium rapidly appears in the bloodstream (La Touche et al., 1987), where it is associated primarily with the red cells (Fisenne & Perry, 1985); a non-diffusible uranyl-albumin complex also forms in equilibrium with a 3 diffusible ionic uranyl hydrogen carbonate complex (U02HC0 +) in the plasma (Moss, 1985). Because of their high affinity for phosphate, carboxyl, and hydroxyl groups, uranyl compounds readily combine with proteins and nucleotides to form stable complexes (Moss, 1985). Clearance from the bloodstream is also rapid, and the uranium subsequently accumulates in the kidneys and the skeleton, whereas little is found in the liver (La Touche et al., 1987). The skeleton is the major site of uranium accumulation (Wrenn et al., 1985); the uranyl ion replaces calcium in the hydroxyapatite complex of bone crystals (Moss, 1985). Based on the results of studies in experimental animals, it appears that the amount of soluble uranium ·accumulated internally is proportional to the intake from ingestion or inhalation. It has been estimated that the total body burden of uranium in humans is 40 f-lg, with approximately 40% of this being present in the muscles, URANIUM 85
20% in the skeleton, and 10%, 4%, I%, and 0.3% in the blood, lungs, liver, and kidneys, respectively (Igarashi et al., 1987). Once equilibrium is attained in the skeleton, uranium is excreted in the urine and faeces. Urinary excretion in humans has been found to account for approximately 1% of total excretion, averaging 4.4 j.lg/day (Singh et al., 1990), the rate depending in part on the pH of tubular urine (Berlin & Rudell, 1986). Under alkaline conditions, most of the uranyl hydrogen carbonate complex is stable and is excreted in the urine. If the pH is low, the complex dissociates to a variable degree, and the uranyl ion may then bind to cellular proteins in the tubular wall, which may then impair tubular function. The half-life of uranium in the rat kidney has been estimated to be approximately 15 days. Clearance from the skeleton is considerably slower; half lives of 300 and 5000 days have been estimated, based on a two-compartment model (Wrenn et al., 1985). In another study using a 10-compartment model, overall half lives for the clearance of uranium from the rat kidney and skeleton were determined to be 5-11 and 93-165 days, respectively (Sontag, 1986). The overall elimination half-life of uranium under conditions of normal daily intake has been estimated to be between 180 and 360 days (Berlin & Rudell, 1986).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
Reported oral LD50s of uranyl ethanoate dihydrate for rats and mice are 204 and 242 mg/kg of body weight, respectively (Domingo et al., 1987). Among the most common signs of acute toxicity are piloerection, significant weight loss, and haemorrhages in the eyes, legs, and nose. The most common renal injury caused by uranium in experimental animals is damage to the proximal convoluted tubules, predominantly in the distal two-thirds (Berlin & Rudell, 1986; Anthony et al., 1994; Domingo, 1995); the rate of effects varies with dosage level (Leggett, 1989). It has recently been shown that uranyl inhibits both Na+ transport-dependent and Na+ transport-independent ATP utilization as well as mitochondrial oxidative phosphorylation in the renal proximal tubule (Leggett, 1989; Domingo, 1995). At doses not high enough to destroy a critical mass of kidney cells, the effect appears to be reversible, as some of the cells are replaced; however, the new epithelial lining differs morphologically, and possibly functionally, from normal epithelium (Wrenn et al., 1985; Berlin & Rudell, 1986). Histopathologically, the regenerated cells are simple flattened cells with no microvilli on luminal surfaces and with reduced numbers of mitochondria (Leggett, 1989). There is some evidence that tolerance may develop following repeated exposure to uranium (Yuile, 1973; Durbin & Wrenn, 1976; Campbell, 1985). This tolerance does not, however. prevent chronic damage to the kidney, as the regenerated cells are quite different; although histopathologically it may appear that the repair process is well advanced, the urinary biochemical changes return to normal only slowly (Leggett, 1989). Alterations causing thickening of the 86 INORGANIC CONSTITUENTS glomerular basement membrane of the kidney, which results from the storage of uranium in the kidney, can be prolonged and severe enough to cause permanent damage (McDonald-Taylor et al., 1992). Persistent ultrastructural changes in the proximal tubules of rabbits have also been reported to be associated with the kidney's ability to store uranium (McDonald-Taylor et al., 1997). Cell damage in the proximal tubules was significantly more severe in animals allowed up to a 91-day recovery period than in animals in the no-recovery group.
5.2 Short-term exposure
Forty male Sprague-Dawley rats given 0, 2, 4, 8, or 16 mg of uranyl ethanoate dihydrate per kg of body weight per day (equivalent to doses of 0, 1.1, 2.2, 4.5, or 9.0 mg of uranium per kg of body weight per day) in drinking-water for 2 weeks exhibited a variety of biochemical effects, including increases in blood glucose levels at ~4 mg ofuranyl ethanoate dihydrate per kg of body weight per day, decreases in aspartate aminotransferase and alanine aminotransferase values at ~8 mg of uranyl ethanoate dihydrate per kg of body weight per day, increases in several other haematological parameters at 16 mg of uranyl ethanoate dihydrate per kg of body weight per day, and increases in total protein levels in all treated groups (Ortega et al., 1989). The authors considered the NOAEL to be 2 mg of uranyl ethanoate dihydrate per kg of body weight per day (1.1 mg of uranium per kg of body weight per day). Groups of 15 male and 15 female weanling Sprague-Dawley rats consumed water containing <0.001 (control), 0.96, 4.8, 24, 120, or 600 mg of uranyl nitrate hexahydrate per litre (equivalent to doses of <0.0001, 0.06, 0.31, 1.52, 7.54, and 36.73 mg of uranium per kg of body weight per day in males and <0.0001, 0.09, 0.42, 2.01, 9.98, and 53.56 mg of uranium per kg of body weight per day in females) for 91 days (Oilman et al., 1997a). Histopathological changes were observed mainly in the liver, thyroid, and kidney. In the liver, treatment-related lesions were seen in both sexes at all doses and were generally non-specific nuclear and cytoplasmic changes. The thyroid lesions were not considered specific to the uranium treatment. The kidney was the most affected tissue. In males, statistically significant treatment related kidney lesions (reported at all doses) included nuclear vesiculation, cytoplasmic vacuolation, and tubular dilation. Other statistically significant lesions in males (~4.8 mg of uranyl nitrate hexahydrate per litre) included glomerular adhesions, apical displacement of the proximal tubular epithelial nuclei, and cytoplasmic degranulation. In females, statistically significant changes in the kidney included nuclear vesiculation of the tubular epithelial nuclei (all doses) and anisokaryosis (all doses except 4.8 mg of uranyl nitrate hexahydrate per litre). However, the most important changes in the female were the capsular sclerosis of glomeruli and reticulin sclerosis of the interstitial membranes; these changes occurred in all dose groups and are considered to be "nonreparable lesions." Significant treatment-related liver changes were also reported in hepatic nuclei and cytoplasm in both sexes at the lowest exposure level. The LOAEL for adverse effects on the kidney and liver of male and female rats, based on the frequency of degree of degenerative lesions in the renal proximal convoluted tubule, was URANIUM 87 considered to be 0.96 mg of uranyl nitrate hexahydrate per litre (equivalent to 0.09 mg of uranium per kg of body weight per day in females and 0.06 mg of uranium per kg of body weight per day in males). The reason for the difference in sensitivity between males and females is not clear, but it did not appear to be due to differences in pharmacokinetics, as accumulation of uranium in renal tissue did not differ significantly between the two sexes at all doses. In a similar study, groups of 10 male New Zealand white rabbits were given uranyl nitrate hexahydrate in drinking-water at concentrations of <0.001 (controls), 0.96, 4.8, 24, 120, or 600 mg/litre (determined to be equivalent to doses of 0, 0.05, 0.2, 0.88, 4.82, and 28.7 mg of uranium per kg of body weight per day) for 91 days (Gilman et al., 1997b ). Histopathological changes were observed in the kidney tubule, liver, thyroid, and aorta. Histopathological findings were observed in the kidney tubules at doses above 0.96 mg ofuranyl nitrate hexahydrate per litre. When compared with controls, significant treatment-related changes included cytoplasmic vacuolation, anisokaryosis, nuclear pyknosis, and nuclear vesiculation; the incidence of nuclear vesiculation and anisokaryosis appeared to be dose-related, with nuclear vesiculation having the higher frequency and severity. Other treatment-related changes included tubular dilation, hyperchromicity, tubular atrophy, changes in the interstitium collagen, and reticulin sclerosis. In total, 11 different morphological indicators of tubular injury were observed in the highest exposure group. The LOAEL, based on the nuclear changes in the kidney, was considered to be 0.96 mg of uranyl nitrate hexahydrate per litre (equivalent to 0.05 mg of uranium per kg of body weight per day). It should be noted, however, that these rabbits were not Pasteurella-free, and four of them contracted a Pasteurella infection during the course of the study. In the same study, 10 Pasteurella-free female rabbits were exposed to drinking-water containing <0.001 (controls), 4.8, 24, or 600 mg ofuranyl nitrate hexahydrate per litre (equivalent to doses of 0, 0.49, 1.32, and 43.02 mg of uranium per kg of body weight per day) for 91 days. Dose-related and treatment related nuclear changes in the kidney tubule included anisokaryosis and vesiculation, which were significantly different from effects observed in controls at all doses. Other treatment-related changes in the kidney included cytoplasmic vacuolation, tubular atrophy, and nuclear pyknosis. In general, histopathological changes in the kidney in females were generally less marked than in males. The LOAEL was considered to be 4.8 mg ofuranyl nitrate hexahydrate per litre (equivalent to 0.49 mg of uranium per kg of body weight per day). In both sexes, histopathological changes in the liver, thyroid, and aorta were similar. In the liver, changes may have been treatment-related, although very mildly affected animals were seen in all groups, and changes in the thyroid were mild. Changes in the aorta were not dose-dependent. It should be noted that no similar aortic changes were observed in the 91-day uranyl nitrate hexahydrate studies in rats (Gilman et al., 1997a). It is interesting to note, however, that even though the female rabbits consumed on average 65% more water than the males and their average uranium intake was approximately 50% greater on a mg/kg of body weight per day basis, their average tissue levels were not similarly raised. The differences between the males and females, both qualitative and quantitative, suggest pharmacokinetic parameter differences, which contrasts with the findings in the rat study by the same authors (Gilman et al., 1997a). 88 INORGANIC CONSTITUENTS
In an additional study to observe the reversibility of renal injury in Pasteurella-free male New Zealand white rabbits, groups of 5-8 animals were given <0.00 1 (control), 24, or 600 mg of uranyl nitrate hexahydrate per litre (equivalent to 0, 1.36, and 40.98 mg of uranium per kg of body weight per day) in drinking-water for 91 days, with a recovery period of up to 91 days (Oilman et al., 1997c). Minor histopathological lesions were seen in the liver, thyroid, and aorta. In the kidney, tubular injury with degenerative nuclear changes, cytoplasmic vacuolation, and tubular dilation was observed in the high-dose group, which did not exhibit consistent resolution even after a 91-day recovery period. In general, the male rabbits did not respond as dramatically as those in the earlier study (Oilman et al., 1997b ), although the histopathological changes observed in this study were similar to those noted in the female rabbits of the previous study. Animals in this study consumed approximately 33% more uranium per day than the males in the previous study (Oilman et al., 1997b ), yet uranium residues in kidney tissue were 30% less, which would appear to indicate that Pasteurella-free rabbits are less sensitive than the non-Pasteurella-free strain to the effects of the uranyl ion in drinking-water. Based on the histopathological data in the kidney, a LOAEL for the male New Zealand rabbits in this study is estimated to lie between 24 and 600 mg of uranyl nitrate hexahydrate per litre.
5.3 Long-term exposure
In an early series of experiments, very high doses (up to 20% in the diet) of a variety of uranium compounds were fed to rats, dogs, and rabbits for periods ranging from 30 days to 2 years (Maynard and Hodge, 1949). On the basis of very limited histopathological investigations, renal damage was reported in each species.
5.4 Reproductive and developmental toxicity
Adverse reproductive effects, in terms of total number of litters and average number of young per litter, were reported in rats given 2% uranyl nitrate hexahydrate for 7 months (Maynard & Hodge, 1949). More recent studies have examined the teratogenic/embryotoxic effects and reproductive outcomes of uranyl acetate dihydrate in Swiss albino mice. Domingo et al. (1989a) evaluated the developmental toxicity of uranium by treating groups of 20 pregnant Swiss mice by gavage to doses of 0, 5, 10, 25, or 50 mg of uranyl acetate dihydrate per kg of body weight per day (equivalent to 0, 2.8, 5.6, 14, and 28 mg of uranium per kg of body weight per day) on days 6-15 of gestation; the animals were sacrificed on day 18. Although all dams survived, there was a dose-related reduction in maternal weight gain, a significant decrease in daily feed intake, and a significant increase in liver weights. Exposure related fetotoxicity, including reduced fetal body weights and length, increased incidence of stunted fetuses per litter, increased incidence of both external and internal malformations, and increased incidence of developmental variations, was observed in the fetuses of mice at ~2.8 mg of uranium per kg of body weight per day. At doses ~14 mg of uranium per kg of body weight per day, specific malformations included .cleft palate and bipartite sternebrae, and developmental URANIUM 89 variations included reduced 'ossification and unossified skeletal variations. There was no evidence of embryolethality at any dose. Based on both the maternal and fetotoxic effects, a LOAEL of 2.8 mg of uranium per kg of body weight per day could be considered. A second study by Domingo et al. (1989b) evaluated the effect of uranium on late fetal development, parturition, lactation, and postnatal viability. Groups of 20 female mice were treated by gavage from day 13 of pregnancy until day 21 of lactation to doses of 0, 0.05, 0.5, 5, or 50 mg of uranyl acetate dihydrate per kg of body weight per day (equivalent to 0, 0.028, 0.28, 2.8, and 28 mg of uranium per kg of body weight per day). Maternal deaths (2/20 at 2.8 mg of uranium per kg of body weight per day, and 3/20 at 28 mg of uranium per kg of body weight per day) were attributed to the treatment; however, maternal toxicity was not evident from changes in body weight or food consumption, although relative liver weight was significantly reduced in all treatment groups. Decreases in pup viability, as indicated by significant decreases in litter size on day 21 of lactation, and significant decreases in the viability and lactation indexes were observed in the high-dose group. Based on developmental effects in pups, a NOEL of2.8 mg of uranium per kg of body weight per day was established. Paternain et al. (1989) studied the effects of uranium on reproduction, gestation, and postnatal survival in mice. Groups of25 mature male Swiss mice were exposed to oral doses ofO, 5, 10, or 25 mg ofuranyl acetate dihydrate per kg ofbody weight per day (equivalent to 0, 2.8, 5.6, and 14 mg of uranium per kg of body weight per day) for 60 days prior to mating with mature females (25 per group). Females were exposed for 14 days prior to mating, and exposure continued through mating, gestation, parturition, and nursing of litters; half the treated dams were sacrificed on day 13 of gestation. No treatment -related effects on mating or fertility were observed. Embryolethality (number of late resorptions and dead fetuses) was significantly increased and the number of live fetuses was decreased in the high-dose group. Lethality in pups (at birth and at day 4 of lactation) was significantly increased at 2:5.6 mg of uranium per kg of body weight per day, and pup growth (decreases in weight and length) and development of offspring, from birth and during the entire lactation period, were significantly affected in the high-dose group. Unspecified degenerative changes in the testes of rats have also been reported following chronic administration of uranyl nitrate hexahydrate and uranyl fluoride in the diet (Maynard and Hodge, 1949; Maynard et al., 1953; Malenchenko et al., 1978). In a more recent study, male Swiss mice were exposed for 64 days to uranyl acetate dihydrate in drinking-water at doses of 0, l 0, 20, 40, or 80 mg/kg of body weight per day (equivalent to 0, 5.6, 11.2, 22.4, and 44.8 mg of uranium per kg of body weight per day) prior to mating with untreated females for 4 days (Llobet et al., 1991). With the exception of interstitial alterations and vacuolization of Leydig cells at the highest dose, no effects were observed in testicular function/spermatogenesis. There was, however, a significant, non-dose-related decrease in the pregnancy rate of these animals. 90 INORGANIC CONSTITUENTS
5.5 Mutagenicity and· related end-points
Uranyl nitrate was cytotoxic and genotoxic in Chinese hamster ovary cells at concentrations ranging from 0.01 to 0.3 mmol/litre. There was a dose-related decrease in the viability of the cells, a decrease in cell cycle kinetics, and increased frequencies of micronuclei, sister chromatid exchanges, and chromosomal aberrations (Lin et al., 1993). The authors suggest that the data provide a possible mechanism for the teratogenic effects observed in the studies by Domingo et al. (1989a). The genotoxic effects in this study were thought to occur through the binding of the uranyl nitrate to the phosphate groups of DNA. Chromosomal aberrations have also been induced in male mouse germ cells exposed to enriched uranyl fluoride; however, these aberrations may have been produced by the radioactivity of the test compound (Hu & Zhu, 1990).
5.6 Carcinogenicity
Although bone cancer has been induced in experimental animals by injection or inhalation of soluble compounds of high-specific-activity uranium isotopes or mixtures of uranium isotopes, no carcinogenic effects have been reported in animals ingesting soluble or insoluble uranium compounds (Wienn et al., 1985).
6. EFFECTS ON HUMANS
Nephritis is the primary chemically induced effect of uranium in humans (Hursh & Spoor, 1973). Little information is available on the chronic health effects of exposure to environmental uranium in humans. In Nova Scotia, Canada, clinical studies were performed on 324 persons exposed to variable amounts of naturally occurring uranium in drinking-water (up to 0.7 mg/litre) supplied from private wells. No relationship was found between overt renal disease or any other symptomatic complaint and exposure to uranium. However, a trend towards increasing excretion of urinary ~ 2 -microglobulin and increasing concentration of uranium in well-water was observed; this raises the possibility that an early tubular defect was present and suggests that this parameter might be useful as an index of subclinical toxicity. The group with the highest uranium concentrations in well-water failed to follow this trend, but this was attributed to the fact that most of the individuals in this group had significantly reduced their consumption of well-water by the time the measurements were made, leading to the conclusion that the suspected tubular defect might well be rapidly reversible (Moss et al., 1983; Moss, 1985). In a pilot study conducted in 1993 in three communities in Saskatchewan, Canada, there was a statistically significant association (p = 0.03) between increasing but normal levels of urine albumin (measured as mg/mmol creatinine) and the uranium cumulative index. The cumulative index was calculated for each study participant as the product of the uranium concentration in drinking-water, the number of cups of water consumed per day. and the number of years lived at the current residence (Mao et al., 1995). The study was conducted with 100 participants URANIUM 91 in three different areas with mean uranium levels ranging from 0.71 (control) to 19.6 f.lg/litre. Urine albumin levels ranged from 0.165 to 16.1 mg/mmol creatinine, with eight participants having "elevated" urine albumin concentrations (>3.0 mg/mmol creatinine). Three participants had serum creatinine concentrations of>120 f.imol/litre (range 50-170 f.imol!litre ), which is reportedly indicative of prevalent renal damage. It should be noted, however, that diabetics were not excluded from the study, although diabetic status and age, known risk factors for renal dysfunction, were factored into the statistical analysis of the results. According to the authors, microalbuminuria has been shown to be a sensitive indicator of early renal disease.
7. PROVISIONAL GUIDELINE VALUE
There are insufficient data regarding the carcinogemc1ty of uranium in humans and experimental animals. The guideline value for the chemical toxicity of uranium was therefore derived using a TDI approach. As no adequate chronic study was identified, the TDI was derived using the results of the most extensive subchronic study conducted to date in which uranium was administered in drinking water to the most sensitive sex and species (Gilman et al., 1997a). In the 91-day study in rats, the LOAEL for degenerative lesions in the proximal convoluted tubule of the kidney in males was considered to be 0.96 mg of uranyl nitrate hexahydrate per litre, which is equivalent to 0.06 mg of uranium per kg of body weight per day. A TDI of 0.6 f.lg/kg of body weight per day was derived using the LOAEL of 60 f.ig/kg of body weight per day and an uncertainty factor of 100 (for intra- and interspecies variation). There is no need to apply an additional uncertainty factor to account for the use of a LOAEL instead of a NOAEL because of the minimal degree of severity of the lesions being reported. Also, an additional uncertainty factor for the length of the study (91-day) is not required because the estimated half-life of uranium in the kidney is 15 days, and there is no indication that the severity of the renal lesions will be exacerbated following continued exposure. This TDI yields a guideline value of 2 f.lg/litre (rounded figure), assuming a 60-kg adult consuming 2 litres of drinking-water per day and a 10% allocation of the TDI to drinking-water. This value would be protective, based on associations for subclinical renal effects reported in preliminary epidemiological studies. Several methods are available for the removal of uranium from drinking water, although some of these methods have been tested at laboratory or pilot scale only. Coagulation using ferric sulfate or aluminium sulfate at optimal pH and coagulant dosages can achieve 80-95% removal of uranium, whereas at least 99% removal can be achieved using lime softening, anion exchange resin, or reverse osmosis processes. In areas with high natural uranium levels, a value of 2 f.lg/litre may be difficult to achieve with the treatment technology available (WRc, 1997). The guideline value for uranium is provisional because it may be difficult to achieve with the treatment technology available, because of limitations in the key study, namely the lack of a dose-response relationship (no NOEL) despite the wide range of administered doses, and because of insufficient information on the degree and severity of the pathological examinations. It should be noted that there are several human studies under way that may provide helpful additional data. 92 INORGANIC CONSTITUENTS
8. REFERENCES
Anthony ML et al (1994) Studies of the b1ochemical toxicology ofuranyl nitrate in the rat. Archtves of toxicology, 68:43-53. Berlin M, Rudell B ( 1986) Uranium. In: Friberg L, Nordberg GF, Vouk VB, eds. Handbook on the toxicology of metals, 2nd ed. Amsterdam, Elsev1er Science Publishers, pp. 623--637. Boomer DW, Powell MJ (1987) Determination of uranium in environmental samples using inductively coupled plasma mass spectrometry Anall'tical chenustry, 59:2810-2813. Campbell DCC (1985) The development of an ammal model with whtch to study the nephrotoxtc effects ofuramum-contammated dnnking water. Hahfax, Nova Scotia, Dalhousie University (M.Sc. thes1s). Cheng YL, Lin JY, Hao XH ( 1993) Trace uranium determmation in beverages and mineral water using fissiOn track techniques. Nuclear tracks and radtation measurements, 22(1-4):853-855. Cothern CR, Lappenbusch WL ( 1983) Occurrence of uranium in drinking water in the US. Health phystcs, 45·89-99 Domingo JL ( 1995) Chem1cal toxicity of uranmm. Toxicology and ecotoxtcology news, 2(3):74-78. Domingo JL et al. (1987) Acute toxicity of uranium m rats and mice. Bulletm of environmental contammation and toxtcology, 39:168-174. Domingo JL et al. ( 1989a) The developmental toxicity ofuramum in mice. Toxtcology, 55(1-2).143-152. Domingo JL et al. (1989b) Evaluation of the perinatal and postnatal effects of uramum in mice upon oral administration. Archtves ofenvironmental health, 44(6)·395-398. Durbin PW, Wrenn ME (1976) Metabolism and effects of uranium in animals. In: Conference on occupatwnal health e>.penence with uramum. Washington, DC, US Energy Research and Development Administration, pp. 68-129 (available from US National Technical InformatiOn Serv1ce) Fisenne IM, Perry PM (1985) lsotop1c U concentration in human blood from New York City donors. Health phystcs, 49:1272-1275. Fisenne I M, Welford GA (1986) Natural U concentration in soft tissues and bone of New York City residents. Health phystcs, 50(6)·739-746. 235 2 231 228 0 212 226 228 Fisenne IM et al. (1987) The daily mtake of • "· U, ·" · Th and • Ra by New York City residents Health phystcs, 53:357-363. Gans I (1985) Natural radionuclides in mineral waters Sctence of the total envtronment, 45:93-99. Gilman AP et al. (1997a) Uranyl nitrate: 28-day and 91-day toxicity studies in the Sprague-Dawley rat. Fundamental and applted toxtcology (in press). Gilman AP et al. (1997b) Uranyl nitrate: 91-day toxicity studies in the New Zealand white rabbit Fundamental and app!ted toxtcology (in press). Gilman AP et al. (1997c) Uranyl mtrate. 91-day exposure and recovery stud1es in the New Zealand white rabbit Fundamental and app!ted toxtcology (in press). Greenwood NN, Earnshaw A (1984) Chenustry ofthe elements. Oxford, Pergamon Press Harley JH (1988) Naturally occurring sources of radioactive contaminatiOn. In: Harley JH, Schmidt GO, Sliini G, eds. Radionuclides m the food cham. Berlm, Springer-Verlag. Hirose K, Sugimura Y (1981) Concentration of uranium and the activity ratio of 214Uf"U in surface air: effect of atmosphenc burn-up of Cosmos-954. Meteorology and geophysics, 32:317 [cited in F1senne & Welford, 1986] Hu Q, Zhu S (1990) Induction of chromosomal aberrations in male mouse germ cells by uranyl fluonde containing enriched uranium. Mutatwn research, 244:209-214. Hursh JB, Spoor NL (1973) Data on man. In· Hodge HC et al., eds. Handbook of expertmental pharmacology Vol 36 Uramum. plutomum, transplutomc elements Berlin, Spnnger Verlag, pp. 197-240. Igarashi Y, Yamakawa A, I ked a N ( 1987) Plutonmm and uran1um in Japanese human tissues. Radtotsotopes, 36.433-439. Kahlos H, Asikainen M (1980) Internal radiation doses from radioactivity of drinking water in Finland. Health phystcs, 39: I 08-111. URANIUM 93
Kreiger HL, Whtttaker EL (1980) Prescnbed procedures for measurement of radioactivity in drink1ng water. Washington, DC, US Environmental Protection Agency (EPA-600/4-80-032) [cited in Blanchard RL et al. (1985) Radiological sampling and analytical methods for national primary drinking water regulations. Health physics, 48(5):587-600]. Landa ER, Councell TB (1992) Leaching of uranium from glass and ceramic foodware and decorative items. Health physrcs, 63.343-348. La Touche YD, Willis DL, Dawydmk 01 (1987) Absorption and biokmetics of U in rats following an oral admmistration ofuranyl nitrate solution. Health physics, 53(2): 147-162. Leggett R W ( 1989) The behav1our and chemical toxicity of U in the kidney: a reassessment. Health physiCS, 57(3):365-383. Lide OR, ed. (1992-93) Handbook of chem1stry and phys1cs Boca Raton, FL, CRC Press. Lin RH et al. (1993) Cytogenetic toxicity of uranyl nitrate in Chinese hamster ovary cells. Mutation research. 319·197-203 Llobet JM et al. ( 1991) Influence of chronic exposure to uranium on male reproduction in mice. Fundamental and applied toxicology, 16:821-829. Lucas HF, Markun F (1970) Thorium and uranium in blood, urme and c1garettes. In: Argonne Natronal Laboratory Radratron Physics DIVISIOn Annual Report, Part 2 Argonne, IL, Argonne National Laboratory, pp. 47-52 (ANL-7760). Malenchenko AF, Barkun NA, Guseva GF (1978) Effect of uranium on the induction and course of experimental autoimmune ochitis and thyrOiditis. Journal of hyg1ene, epidemiology, mrcrobrology and unmunology, 22(3):268-277 Mao Y et al. (1995) Inorganic components of drinking water and microalbuminuria. Environmental research, 71·13 5-140. Maynard EA, Hodge HC (1949) Stud1es of the toxicity of various uranium compounds when fed to experimental animals. In: Voeglm C, ed. Pharmacology and toxicology of uranium compounds. New York, NY, McGraw-Hill, pp. 309-376. Maynard EA, Downs WL, Hodge HC (1953) Oral tox1city of uranium compounds. In: Voegtlin C, Hodge HC, eds. Pharmacology and toxicology of uranium compounds Chronic inhalatron and other studies New York, NY, McGraw-Hill, pp. 1121-1369. McDonald-Taylor CK, Singh A, Gilman A (1997) Uranyl nitrate-induced proximal tubule alterations m rabbits: a quantitative analysis. Journal oftoxrcologic pathology, 25(4):381-389. McDonald-Taylor CK et al ( 1992) Uranyl nitrate-induced glomerular basement membrane alteratiOns m rabb1ts: a quantitative analysis. Bulletin of environmental contammatron and toxrcology, 48:367-373. Moss MA ( 1985) Chrome low level uranium exposure via drmking water- clinical mvestigations in Nova Scotw. Halifax, Nova Scotia, Dalhousie University (M.Sc. thesis). Moss MA et al. (1983) Uranium in drinking water- report on clinical studies in Nova Scotia. In: Brown SS, Savory J, eds. Chemical toxrcology and clinical chemistry of metals. London, Academic Press, pp 149-152. Nozaki T et al. (1970) Neutron activation analys1s of uranium in human bone, drinking water and daily diet. Journal of radroanalytrcal chemistry, 6:33--40 OMEE (1996). Momtorrng data for uranium - 1990-1995. Toronto, Ontario, Ontario Ministry of Environment and Energy, Ontario Drinking Water Surveillance Program. Ortega A et al. ( 1989) Evaluation of the oral toxicity of uranium in a 4-week drinking-water study in rats Bullet m of environmental contammatron and tox1cology, 42:935-941. Paternain JL et al. (1989) The effects of uranium on reproduction, gestation, and postnatal survival in mice. Ecotox1cology and environmental safety, 17:291-296. Roessler CE et al ( 1979) Uranium and radium-226 in Flonda phosphate materials. Health phys1cs, 37:267-269. Singh NP. Wrenn ME (1988) Determinat1ons of actinides in biological and environmental samples. Science of the total env1romnent, 70: 187-203. Singh NP et al. ( 1990) Daily U intake in Utah res1dents from food and drinking water. Health physics, 59(3):333-337. 94 INORGANIC CONSTITUENTS
Sontag W (1986) Multicompartment kinetic models for the metabolism of americium, plutonium and uranium in rats Human toxzcology, 5:163-173. Sullivan MF et al. ( 1986) Influence of oxidizing or reducing agents on gastrointestina1 absorption of U, Pu, Am, Cm and Pm by rats. Health physzcs, 50(2):223-232. Tracy BL et al. (I 992) Absorption and retention of uranium from drinking water by rats and rabbits. Health physzcs, 62(1)·65-73. US EPA (I 990) Occurrence and exposure assessment for uranium m public drinking water supplies. Report prepared by Wade Miller Associates, Inc. for the Office of Drinking Water, US Environmental Protection Agency, 26 April I 990 (EPA Contract No. 68-03-35 !4). US EPA (1991) Revzew of RSC analysis. Report prepared by Wade Miller Associates, Inc. for the US Environmental Protection Agency, 9 May 1991 [follow-up to US EPA, 1990]. WRc (1997) Treatment technology for alummium, boron and uramum. Document prepared for WHO by the Water Research Centre, Medmenham, and reviewed by S. Clark, US EPA; A. van Dijk-Looijaard, KIW A, Netherlands; and D. Green, Health Canada. Wrenn ME et al. (1985) Metabolism of ingested U and Ra. Health physics, 48:601--633. Yuile CL (1973) Animal expenments. In: Hodge HC et al., eds. Handbook of experimental pharmacology Vat 36. Uramum, plutomum, transplutonic elements. Berlm, Springer Verlag, pp. 165-195. CYANOBACTERIAL TOXINS: MICROCYSTIN-LR
First draft prepared by S. Gupta Bureau of Chemical Hazards, Environmental Health Directorate Health Protection Branch, Health Canada, Ottawa, Ontario, Canada
No guideline values were proposed for cyanobacterial toxins in the second edition of the WHO Guidelines for drinking-water quality. As microcystins (produced by cyanobacteria, or blue-green algae) are extremely toxic and are often associated with poisonings in humans and animals, the Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality decided that guideline values for cyanobacterial toxins were needed.
1. GENERAL DESCRIPTION
1.1 Identity
The cyanobacteria, also known as blue-green algae, owe their name to the presence of photosynthetic pigments. Cyanobacteria are a major group of bacteria that occur throughout the world. Freshwater cyanobacteria may accumulate in surface water supplies as "blooms" and may concentrate on the surface as blue-green "scums." Some species of cyanobacteria produce toxins, which are classified according to their mode of action into hepatotoxins (e.g. microcystins), neurotoxins (e.g. anatoxins), skin irritants, and other toxins. Both hepatotoxins and neurotoxins are produced by cyanobacteria commonly found in surface water and therefore are of relevance to water supplies (Carmichael, 1992; Fawell et al., 1993). The hepatotoxins are produced by various species within the genera Microcystis, Anabaena, Oscillatoria, Nodularia, Nostoc, Cylindrospermopsis, and Umezakia, although not all strains do so (Fawell et al., 1993; A WW A, 1995). Most hepatotoxins (all cyclic heptapeptides) are microcystins. At least 50 congeners of microcystins are known (Carmichael, 1994), and several of these may be produced during a bloom. The chemical structure of microcystins includes two variable amino acids and an unusual aromatic amino acid, ADDA (3-amino-9-methoxy-2,6,8- trimethyl-1 0-phenyldeca-4,6-dienoic acid), containing a substituted phenyldecadienoic acid (Botes et al., 1985). Different microcystins have different lipophilicities and polarities, which could affect their toxicity. Microcystin-LR was the first microcystin chemically identified; to date, most work has been conducted using this microcystin. It has been associated with most of the incidents of toxicity involving microcystins in most countries (Fawell et al., 1993). Microcystin-LR is a cyclic heptapeptide with a molecular weight of about 1000 daltons. Neurotoxins are not considered as widespread in water supplies, and they do not appear to pose the same degree of risk from chronic exposure as microcystins (Fawell et al., 1993; AWWA, 1995). The neurotoxins, such as anatoxin-a and -a(s), 96 ORGANIC CONSTITUENTS are highly toxic nerve poisons but have short biological half-lives. On acute exposure, the neurotoxins cause death within minutes to a few hours, depending on the species, the amount of toxin ingested, and the amount of food in the stomach (Carmichael, 1992). Dog poisonings in Scotland were reported to be due to the consumption of Oscillatoria containing anatoxin-a (Codd, 1992). Toxic cyanobacteria also produce cytotoxic alkaloids, the most recently described being from the species Cylindrospermopsis raciborskii, which occurs in freshwater lakes, rivers, and drinking-water storage reservoirs in tropical areas. The molecular structure is a tricyclic guanidine linked to a hydroxymethyl uracil with a molecular weight of 415 daltons (Ohtani et al., 1992). These alkaloids have been implicated in a variety of health effects, ranging from gastroenteritis to kidney disease (Falconer, 1994).
1.2 Occurrence and growth of cyanobacteria
The occurrence of a particular genus and sp~cies of cyanobacteria around the world is apparently influenced by regional differences in water chemistry and climatic conditions. For example, Cylindrospermopsis is produced in tropical waters but has not been found in temperate climates. Similarly, Microcystis and Anabaena blooms occur widely in the temperate regions of the world (A WW A, 1995). In general, 50-75% of bloom isolates can produce toxins, often with more than one toxin being present. Toxic and non-toxic blooms of the same species can be found together (Skulberg et al., 1993; A WWA, 1995; Codd & Bell, 1996). The overall toxicity of a bloom can be uncertain, because variations can occur in toxin concentration over a short time and spatially within a water body experiencing a bloom (Hrudey et al., 1994). There is no simple method to distinguish the toxic from the non-toxic forms. The unpredictability of toxin production within any given bloom "renders them potentially dangerous and suspect at all times" (Ressom et al., 1994 ), and prevention of cyanobacterial blooms is therefore the key to the control of toxic blooms. The growth of cyanobacteria and the formation of blooms are influenced by physical, chemical, and biological factors, which were recently reviewed by Pearson et al. (1990), Ressom et al. (1994), and A WWA (1995) and are discussed below. As a result of the interplay of these factors, there may be large yearly fluctuations in the levels of cyanobacteria and their toxins. There is also a seasonal variation in predominating species. Cyanobacterial blooms persist in water supplies that contain adequate levels of essential inorganic nutrients such as nitrogen and phosphorus, water temperatures generally between 15 and 30°C, and pH levels between 6 and 9. Blooms usually occur in late summer or early fall and are most common in eutrophic or hypereutrophic bodies of water. The amount of daylight needed to optimize growth depends on the species. In addition, some cyanobacteria, such as Microcystis aeruginosa, can regulate their buoyancy in response to available light. This characteristic allows cyanobacteria to migrate through thermal gradients and use nutrients confined to cooler deeper water below. Buoyancy is controlled mainly through the production of carbohydrates from CYANOBACTERIAL TOXINS: MICROCYSTIN-LR 97 photosynthesis. This control mechanism breaks down if there is too little carbon dioxide available. Although buoyancy cannot be adjusted during the night, the organisms will float to the surface because of their reduced carbohydrate content as a result of respiration at night. Turbulence and high water flows are unfavourable to the growth of cyanobacteria, as they interfere with the organisms' ability to maintain a position in the water column. Heavy rain storms can increase runoff and nutrient levels in the water, which encourages the formation of blooms. The formation of surface scum is enhanced by calm weather conditions. Initially, there may be high barometric pressure and light to moderate winds, accompanied by constant circulation in a water body in which large numbers of cyanobacteria are maintaining their position in the water column to take advantage of those conditions. If the wind stops and circulation also stops, the cyanobacteria may suddenly become "overbuoyant." If they cannot adjust their buoyancy fast enough or at all (at night), then the blooms will float to the surface and form surface scum. Thus, scums are often formed overnight. The scum may drift downwind and may settle at lee shores and quiet bays, where the cyanobacteria may release their toxins and eventually die (Ressom et al., 1994).
1.3 Toxin production and persistence
The two main factors that have been shown to affect toxin production are light and temperature. The optimum temperature for toxin production in cyanobacteria is between 20 and 25°C (Gorham, 1964; van der Westhuizen & Eloff, 1985; Watanabe & Oishi, 1985), which suggests that cyanobacteria are most toxic during periods with warm weather and in areas with warm climates. However, the optimum temperature may change from country to country. Light intensity more than light quality is an important factor in toxin production in M aeruginosa. Both toxicity and the ratio of the toxin to protein production are enhanced by both red and green light compared with white light (Utkilen & Gjolme, 1992). The toxicity of cyanobacteria increases with an increase in light intensity below 40 microeinsteins/m2 per second (van der Westhuizen & Eloff. 1985; Watanabe & Oishi, 1985; Utkilen & Gjolme, 1992) and therefore decreases with water depth (Utkilen & Gjolme, 1992). However, when mixing of water from different depths occurs, especially during periods of high winds, this may not be true. Some laboratory studies have shown that pH, nitrogen, phosphorus, and carbon dioxide could also influence the growth of microcystins. The presence of six different microcystins in floating scums of M aeruginosa in Hartbeespoort Dam, South Africa, was monitored for 2.5 years. The toxins were either not detectable or in low concentrations during the winter and reached maximum concentrations during the summer. The total concentrations of four of the toxins (5--415 f.lg/g dry scum) were directly correlated to solar radiation, surface water temperature, pH, and per cent oxygen saturation. No significant correlations were found between total toxin concentrations in scum samples and organic and inorganic nutrient concentrations in surface water (Wicks & Theil, 1990). 98 ORGANIC CONSTITUENTS
Kotak et al. (1995) studied the patterns of occurrence of microcystin-LR (measured as f.!g/g biomass of M aeruginosa by high-performance liquid chromatography, or HPLC) in three hypereutrophic hard-water lakes in central Alberta, Canada, over three seasons. Microcystis aeruginosa was highly variable temporally (differences up to 3 orders of magnitude) within each lake over 1 year, between years in an individual lake, and between lakes in a year. Seasonal changes in microcystin-LR concentration were positively correlated to the abundance and biomass of the M aeruginosa, total and total dissolved phosphorus concentration, pH, and chlorophyll. Surprisingly, there was a negative correlation between microcystin-LR concentration and nitrate concentration and no correlation with water temperature. Over a 24-hour period, the concentration of microcystin-LR in M aeruginosa decreased more than 6-fold at night compared with concentrations during the day. Codd & Bell (1996) determined the effect of temperature and nutrient supply on microcystin levels in cultures of M aeruginosa in water bodies in the United Kingdom. The amount of microcystin per unit cyanobacterial dry weight was higher when nitrate levels were in excess and highest between 20 and 25°C, but was reduced above 25°C and below 20°C. As toxin production varies greatly among different strains of the same species, genetic differences and metabolic processes may also be important in the production of these toxins. Studies have shown that the ability to produce toxins can vary temporally and spatially at a particular site or within the bloom itself (Hrudey et al., 1994; Ressom et al., 1994). Cyanobacterial toxins either are membrane-bound or occur free within the cells. In laboratory studies, most of the toxin release occurs as cells age and die and passively leak their cellular contents, although active release of toxins can also occur from young growing cells (Pearson et al., 1990). Watanabe & Oishi (1983) investigated the toxicity of a cultured strain of M aeruginosa through Jag, exponential, and stationary growth phases. Maximum toxicity was observed between the exponential and stationary growth phases. The maximum cellular content of the toxin in two cultured Microcystis strains occurred at the late stage of exponential growth (Watanabe et al., 1989). Microcystins and alkaloid toxins are degraded in natural waters, but there may be a Jag phase before significant degradation takes place. Studies conducted using microcystin-LR at 10 f..lg/litre in reservoir water in the United Kingdom suggest a half-life of less than a week (Cousins et al., 1996). Codd & Bell (1996) also found that microcystin was readily biodegraded in ambient waters, with a half life of about 1 week. Generally, if a Jag phase exists, it is about 9 or 10 days long. In one study, microcystin was present up to 21 days following treatment of a bloom with an algicide (Jones & Orr, 1994). This could have been due to the shock dose with copper sulfate. Microcystin-LR is very stable in water and resistant to pH extremes and temperatures up to 300°C (Wannemacher, 1989). Biodegradation and photolysis are means by which released microcystin-LR can naturally degrade in water (Kenefick et al., 1993; Tsuji ~tal., 1994 ). CYANOBACTERIAL TOXINS: MICROCYSTIN-LR 99
2. ANALYTICAL METHODS
Ressom et al. (1994), Lambert et al. (1994b ), and A WWA (1995) reviewed the methods available for the analysis of microcystins in drinking-water. In comparing the various analytical methods being used for microcystin-LR and other microcystins, including their detection limits, it may be useful to distinguish screening methods, such as the mouse bioassay, enzyme-linked immunosorbent assay (ELISA), and phosphatase bioassay, which are conducted before clean-up and indicate the presence of toxins in samples, from methods that are conducted for the identification and quantification of the various individual microcystins (Harada, 1994). Often, more than one toxin may be present in a sample. The consensus among those using analytical methods is that a single method will not suffice for the identification and accurate quantification of many microcystins. The best approach for monitoring is to use a combination of screening and more sophisticated quantification methods. It is important to measure total microcystins, which includes microcystins occurring free in water and microcystins bound to or inside cyanobacterial cells and which includes all microcystins, not just microcystin-LR. Thus, sample preparation may need to include sanification (to break up cells) and a variety of extraction procedures in order to isolate the different (i.e. more lipophilic or polar) microcystins. Most of the existing studies on the levels of microcystins in water supplies have not clearly indicated whether total or free microcystins were measured. The mouse bioassay plays an important role as a screening tool, as it gives the total toxic potential of the sample within a few hours and it can distinguish hepatotoxins from neurotoxins. The assay determines the minimum amount of toxin required to kill a mouse and compares this value with lethal doses of a known amount of toxin. The disadvantage is that it does not detect toxins at low levels, especially in finished drinking-water, and it does not identify the specific toxic agent (Lambert et al., 1994a). As some cyanobacteria can produce both microcystin and anatoxin, the presence of microcystin can be overlooked with the mouse bioassay. The problem arises because anatoxin can kill the mouse within minutes, whereas microcystin can kill the animal within an hour. Bhattacharya and colleagues (1996) used a modified mouse liver slice culture technique for rapidly screening large numbers of cultures and bloom samples of cyanobacterial species for cytotoxicity and hepatotoxicity. Following microcystin-LR treatment, the hepatocytes were swollen with granulated cytoplasm. Congestion and haemorrhage were also evidenced by eosinophilic debris. The method is sensitive enough to detect toxins at the microgram level (Bhattacharya et al., 1996). The protein phosphatase bioassay is another screening method for the quantification of microcystin-LR in water samples (Lambert et al., 1994a). This method is sensitive to subnanogram levels of microcystins in finished and raw water samples. This is a quick method, and many samples can be quantified in a few hours. However, it is not specific to microcystins and will indicate the presence of other substances inhibiting protein phosphatases. This should not be a problem when monitoring a particular area where the potentially occurring species and their possible toxins are known. 100 ORGANIC CONSTITUENTS
An ELISA using polyclonal antibodies and with a detection limit of 0.2 ng/ml has been published (Chu et al., 1990). It is likely that methods using monoclonal or polyclonal antibodies raised against a single toxin (e.g. microcystin LR) will have problems of cross-reactivity with other microcystins. There are several HPLC methods for the identification and quantification of microcystins and other cyanobacterial toxins. Many HPLC methods are variations on methods developed by Siegelman et al. (1984) and Harada et al. (1988). HPLC can distinguish between microcystin analogues, provided standards are available for reference (Lambert et al., 1994b). The HPLC-ultraviolet (UV) detection method is more sensitive than the mouse bioassay, but it does not detect microcystin at levels lower than 1 J..lg/litre in waters containing high levels of natural organic matter (Lambert et al., 1994b). In the United Kingdom, an official HPLC-UV method with a detection limit of 0.5 J..lg/litre has been developed by the Water Research Centre. This method is designed to measure only free microcystin-LR; it does not measure "total" microcystin (free plus cell-bound) or microcystins other than microcystin-LR. It could be modified, however, to measure total microcystins (Fawell et al., 1993). Lawton et al. (1994) developed a reverse-phase HPLC method to determine numerous variants of microcystin and nodularin (a hepatotoxin) in both raw and treated water by a single procedure within 24 hours. This method involves filtration to separate cyanobacterial cells from water, allowing intracellular and extracellular toxin levels to be assessed. The filtered water is subjected to trace enrichment using a 18C solid-phase extraction cartridge, followed by identification and determination by photodiode array HPLC. Recoveries of microcystin-LR, -RR, -L Y, -LF, and nodularin were good when raw and treated water samples were spiked with a mixture ofmicrocystins and nodularin at concentrations as low as 0.25 J..lg/litre. Liquid chromatography-mass spectrometry can analyse microcystin-LR, -YR. and -RR separately at 37, 42, and 23 ng/litre, respectively. It is a simple method that does not require a clean-up procedure because of its high selectivity (Tomoyasu & Keiji, 1996).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
The major route of human exposure to cyanobacterial toxins is the consumption of drinking-water. A minor exposure route is the recreational use of lakes and rivers; for microcystin-LR, however, absorption through skin contact is unlikely, as the toxin does not readily cross cell membranes (Eriksson et al., 1990). Some people are also exposed to cyanobacterial toxins through the consumption of certain algal food tablets. An additional, minor route of exposure is through inhalation while taking showers; microcystin-LR, however, is very water soluble and non-volatile, so inhalation and absorption through the lungs are unlikely, unless the toxin is inhaled as an aqueous aerosol in air (Lambert et al., 1994b ). The extent to which cyanobacterial toxins move up the food-chain (e.g. freshwater mussels and fish) has been investigated recently (Falconer et al., 1992b; Negri & Jones, 1995). The duration of toxin exposure would generally be shorter in colder countries than in those with milder climates. CYANOBACTERIAL TOXINS: MICROCYSTIN-LR 101
The levels of microcystin-LR in the lakes and dugout ponds of Alberta, Canada, ranged from 4 to 605 ~J.g/g dry weight ofbiomass (Kotak et al., 1993) or up to 1500 ~J.glg (Hrudey et al., 1994). More than 70% of over 380 bloom biomass samples from 19 lakes in Alberta between 1990 and 1992 showed detectable levels (>I 1-1-g of microcystin-LR per g dry weight of biomass) of toxin (Hrudey et al., 1994). Similarly, levels of microcystin-LR from natural blooms of Microcystis in Japan, between 1989 and 1991, ranged from 27 to 622 ~J.glg dry weight of biomass (Park et al., 1993 ). In the same blooms, the levels of microcystin-RR and microcystin-YR ranged from 11 to 979 ~J.glg dry weight of biomass and from 9 to 356 ~J.glg dry weight of biomass, respectively, with a total maximum level of microcystins of 1732 ~J.glg dry weight ofbiomass (Park et al., 1993). For two Alberta drinking-water supplies, the raw water intake levels of microcystin ranged from 0.15 to 4.3 ~J.g/litre, with a large coefficient of variation of 59% for hourly fluctuations over an 11.5-hour period. In treated water, levels ranged from 0.09 to 0.64 ~J.g/litre, with a small coefficient of variation of 10%. Over a 5-week period, similar coefficients of variation were obtained in the two types of samples (Hrudey et al., 1994). In the summer of 1993, microcystin-LR was detected (>0.5 ~J.g/litre) in water samples collected from Shoal Lake, Manitoba, Canada, and from within the drinking-water distribution system following the presence of M aeruginosa blooms in Shoal Lake (Jones, 1996). Following this, in 1995, 160 surface water supplies, located mainly in southwestern Manitoba, were chosen for algal study. Treated water samples were analysed only for those sites in which raw water supplies were found to have detectable levels of toxins (detection limit 0.1 ~-tg/litre). Toxin was present in 68% of the treated water samples collected from both the municipal water supply and dugouts used for domestic and livestock consumption. Thus, it appears that conventional treatment methods may be only partially successful in removing the toxins. Toxin concentrations ranged from <0.1 to 1.0 ~J.g/litre in raw water samples and from <0.1 to 0.6 ~J.g/litre in treated water samples. Fastner et al. (1995) detected seven different microcystins in 9 of 12 eutrophic water bodies in Germany in 1993. The microcystin concentration was up to 800 ~J.g/g dry weight for a single microcystin. Cytototoxins were also observed in six field samples. With appropriate water treatment, maximum exposure to total microcystins is probably less than 1 ~J.g/litre, based on the above data. Average exposure generally would probably be well below this level. Not all water supplies, however, are treated by filtration or adsorption; many are untreated or simply chlorinated. Cylindrospermopsis cultured from a drinking-water supply reservoir (Hawkins et al., 1985) has been shown to contain the toxic alkaloid cylindrospermopsin at 0.5% dry weight of algae (Ohtani et al., 1992).
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
The most likely route of exposure to cyanobacterial toxins is via oral ingestion. However, there have been no pharmacokinetic studies with orally 102 ORGANIC CONSTITUENTS administered microcystins. After intravenous or intraperitoneal injection of sublethal doses of variously radiolabelled toxins in mice and rats, microcystin appears to be transported by bile acids transporter in both the intestine (Falconer et al., 1992a) and the liver (Runnegar et al., 1991). About 70% of the toxin is rapidly localized in the liver (Brooks & Codd, 1987; Meriluoto et al., 1990; Lin & Chu, 1994). The kidney and intestine also accumulate significant amounts of microcystin-LR (Meriluoto et al., 1990; Robinson et al., 1991a). Plasma half-lives of microcystin-LR, after intravenous administration, were 0.8 and 6.9 minutes for the alpha and beta phases of elimination, but the concentration of radioactive eH-microcystin-LR) label in the liver did not change throughout a 6-day study period (Robinson et al., 1991a). Microcystin-LR was excreted rapidly, with 75% of the total excretion occurring within 12 hours. The remaining 24% of the administered dose was excreted after 6 days, about 9% via the urinary route and 15% slowly (1% per day) via the faecal route. Microcystin-LR does not readily cross cell membranes and does not enter most tissues. It crosses the ileum through the multispecific organic ion transport system (Runnegar et al., 1991) and mainly enters hepatocytes, where it is covalently bound to a 40 000-dalton protein (protein phosphatase 2A and possibly protein phosphatase 1) in the cytosol (Robinson et al., 1991b). The liver plays a large role in the detoxification of microcystins (Brooks & Codd, 1987; Robinson et al., 199lb). Detoxification "products were seen in urine, faeces, and liver cytosolic fractions (Robinson et al., 1991a), but these products have not been structurally identified. The detoxification products of microcystin-LR are more water soluble than the parent toxin.
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO SYSTEMS
5.1 Acute exposure
Fatalities in animals have been reported following the consumption of water containing large numbers (>106/ml) of cyanobacterial cells (Beasley et al., 1989; Carmichael, 1992).
Microcystin-LR is an extremely acute toxin. The LD 50 by the intraperitoneal route is approximately 25-150 J.tg/kg of body weight in mice; the oral (by gavage)
LD50 is 5000 J.tglkg of body weight in mice, and higher in rats (Fawell et al., 1994).
The intraperitoneal LD50s of several of the commonly occurring microcystins (microcystin-LA, -YR, and -YM) are similar to that of microcystin-LR, but the intraperitoneal LD50 for microcystin-RR is about 10-fold higher (Kotak et al., 1993; Rinehart et al., 1994). The microcystins are primarily hepatotoxins. After acute exposure to microcystin by intravenous or intraperitoneal injection, severe liver damage occurs, characterized by a disruption of liver cell structure, a loss of sinusoidal structure, increases in liver weight due to intrahepatic haemorrhage, haemodynamic shock, heart failure, and death. Other organs affected include the kidney and lungs (Hooser et al., 1990). Intestinal damage is a consequence of the transport of microcystins through the lining cells, which are damaged in a similar manner to hepatocytes (Falconer et al., 1992a). CYANOBACTERIAL TOXINS: MICROCYSTIN-LR 103
5.2 Short-term exposure
Microcystin-LR was administered orally by gavage to groups of 15 male and 15 female mice at 0, 40, 200, or 1000 J.!g/kg of body weight per day for 13 weeks. No definite treatment-related changes were noted at the lowest dose. At 200 J.!g/kg of body weight per day, there was slight liver pathology in some male and female mice. At the highest dose level, all male and most female mice showed liver changes, which included chronic inflammation, focal degeneration of hepatocytes, and haemosiderin deposits. In male mice at the two highest dose levels, serum transaminases were significantly elevated, serum gamma glutamyl transferase was significantly reduced, and there were small but significant reductions in total serum protein and serum albumin. In female mice, changes in transaminases were observed, but only at the highest dose level. Also at the highest dose level, food consumption in male and female mice was increased by 14 and 20%, respectively, but body weight was 7% lower in both sexes compared with control mice. The NOAEL for microcystin-LR was considered to be 40 J.!g/kg of body weight per day (Fawell et al., 1994). In another study, extract from M aeruginosa was given to groups of five pigs in their drinking-water for 44 days at dose levels calculated to be equivalent to microcystin doses of 280, 800, or 1310 J.!g/kg of body weight per day (Falconer et al., 1994). The extract contained at least seven microcystin variants, with microcystin-YR tentatively identified as the major constituent. A no-adverse-effect level (NAEL) for microcystins of280 J.!glkg of body weight per day was reported by the authors, with liver injury (evident from histopathology and changes in serum enzymes) observed at the two highest dose levels. However, one pig was also affected at the lowest dose level, and it is appropriate to consider the 280 J.!glkg of body weight per day dose level as a LOAEL.
5.3 Long-term exposure
An oral repeated-dose study was conducted with M aeruginosa extract supplied to mice at five concentrations (equivalent to 750-12 000 J.!g of microcystin-YM per kg of body weight per day) in their drinking-water for up to 1 year. At the higher concentrations, increased mortality, increased incidences of bronchopneumonia, and chronic liver injury were noted. No liver cancer was seen, but the authors indicated that there may have been some evidence of tumour promotion. No clear NOAEL was established in this study (Falconer et al., 1988).
5.4 Reproductive and developmental toxicity
To investigate the effects of microcystin-LR on the embryonic and fetal development of the mouse, four groups of 26 time-mated female mice of the Crl:CD-1 (ICR) BR strain were dosed once daily by oral gavage with aqueous solutions of microcystin-LR from days 6 to 15 of pregnancy, inclusive. The dose levels were 0, 200, 600, or 2000 J.!g/kg of body weight per day. Maternal clinical signs, body weights, and food consumption were recorded. On day 18 of pregnancy, 104 ORGANIC CONSTITUENTS the females were killed, a necropsy was performed, and the fetuses were examined for abnormalities. Only treatment at 2000 Jlg/kg of body weight per day was associated with maternal toxicity and mortality. Nine of the 26 females died or were sacrificed prematurely during the dosing period. At necropsy, a number of females had abnormal livers, and retardation of fetal weight and skeletal ossification were observed at the maximum dose. There was no evidence of embryolethality, teratogenicity, or embryonic growth retardation at all dose levels. There was no apparent effect of treatment at any dose level on litter size, post-implantation loss, or the sex distribution of the live fetuses. The NOAEL for any aspect of developmental toxicity was 600 Jlg/kg of body weight per day (Fa well et al., 1994 ). These data are in agreement with those of Falconer et al. (1988), which similarly provided no evidence of teratogenicity, embryonic mortality, or reduction in fertility in mice exposed to microcystin at 750 Jlg/kg of body weight per day in drinking-water from weaning (for 17 weeks prior to mating) through mating.
5.5 Mutagenicity and related end-points
No mutagenic response was observed for purified toxins derived from Microcystis in the Ames Salmonella assay (strains TA98, TA100, and TA102) with or without S9 activation. The Bacillus subtilis multigene sporulation test was also negative with regard to mutagenicity using both the 168 and hcr-9 strains (Repavich et al., 1990). In contrast, results of a study in which the purified toxins were tested against human lymphocytes suggested that the toxins may be clastogenic, as indicated by dose-related increases in chromosomal breakage (Repavich et al., 1990).
5.6 Carcinogenicity
There has been some evidence of tumour promotion in animal studies. In a modified two-stage carcinogenesis mouse skin bioassay, dimethylbenzanthracene (DMBA) (500 J.!g) in acetone was applied to the skin of four out of six groups of 20 3-month-old Swiss female mice. After 1 week, the DMBA-treated mice received drinking-water, Microcystis extract in drinking-water (actual microcystin-YM dose not provided), croton oil (as a positive control) applied to the skin (0.5% in 0.1 m! acetone twice a week) plus drinking-water, or croton oil plus Microcystin extract; the control mice received drinking-water or Microcystis extract in drinking-water. After 52 days from initiation, substantial skin tumours and ulcers were visible on the DMBA-treated mice consuming Microcystis extract. There was a significant increase in the mean weight of skin tumours per mouse in treated mice given the Microcystzs extract compared with water. The actual number of tumours per mouse and the weights of the tumours in relation to the weights of the animals were not provided. It was concluded by the authors that oral consumption of Microcystis extract in drinking-water may act as a promoter. However, the mechanism of action is not clear, as microcystins have difficulty penetrating epidermal cells. The tumour weight per mouse in DMBA-treated mice given both croton oil and the algal extract was slightly lower than in those given croton oil and drinking-water. These latter findings could not be explained by the author (Falconer, 1991 ). CYANOBACTERIAL TOXINS: MICROCYSTIN-LR 105
In a two-stage carcinogenicity bioassay, groups of I 0-19 7-week-old male Fischer 344 rats were initiated by intraperitoneal injection with 200 mg of diethylnitrosamine (DEN) per kg of body weight, followed by partial hepatectomy at the end of the third week. Tumour promotion was assessed by intraperitoneal injection of microcystin-LR at I or I 0 !J.glkg of body weight from the third week of the experiment. 3 or 5 times per week. Tumour promotion, as indicated by an increase in glutathione S-transferase placental form (GST-P) positive liver foci, was seen after 8 weeks in animals dosed with 10 11g of microcystin-LR per kg of body weight (Nishiwaki-Matsushima et al., 1992). In a second similar experiment. the same authors used dose levels of I 0 !J.g/kg of body weight per day before, and I 0, 25, or 50 11g/kg of body weight per day after, partial hepatectomy. It was found that the increase in GST-positive foci was dose related. However. these results should be interpreted with caution, because the doses used were very high and could have caused significant damage to hepatocytes. Microcystin-LR had no effect when given to non-initiated rats or to rats that had received partial hepatectomy but no promotion dose of microcystin-LR. Microcystin-LR was found to be a potent inhibitor of eukaryotic protein serine/threonine phosphatases I and 2A both in vlfro (Honkanen et al., 1990; Mackintosh et al., 1990) and in vivo (Runnegar et al., 1993). Such substances are considered to be non-phorbol ester (TP A) type tumour promoters. The inhibition of protein phosphatase 2A by microcystin-LR can be effectively reversed in the presence of polyclonal antibodies against microcystin-LR (Lin & Chu, 1994). The inhibition of protein phosphatase type I and type 2 activities by microcystin-LR is similar to that of the known protein phosphatase inhibitor and tumour promoter okadaic acid (Nishiwaki-Matsushima et al., 1992). The implications for low-level chronic exposure to microcystins are not known.
6. EFFECTS ON HUMANS
Blue-green algae have been known to cause animal and human poisoning in lakes, ponds, and dugouts in various parts of the world for over I 00 years. Through the recreational use of contaminated water, cyanobacterial blooms of Microcystis, Anabaena, and others have been linked to incidence of human illness in many countries. but no fatalities have been reported (Lambert et al., 1994b). In Canada. human illnesses have been reported in Saskatchewan, with symptoms including stomach cramps, vomiting, diarrhoea, fever, headache, pains in muscles and joints, and weakness (Dillenberg & Dehnel, 1960). Similar symptoms as well as skin, eye, and throat irritation and allergic responses to cyanobacterial toxins in water have also been reported in other countnes (Ressom et al., 1994). The reported instances of illnesses are few, but. because they are difficult to diagnose, such illnesses may in fact be more common than has been reported (Carmichael & Falconer, 1993). In Saskatchewan, Canada, I 0 children became sick with diarrhoea after swimming in a lake covered with cyanobacteria. Anabaena cells, which produce microcystin-LR, were found in the stools of one child (Dillenberg & Dehnel, 1960). In the United Kingdom. I 0 of 18 army recruits on a military exercise in a reservoir with a 106 ORGANIC CONSTITUENTS bloom of M aeruginosa suffered abdominal pain, nausea, vomiting, diarrhoea, sore throat, dry cough, blistering at the mouth, and headache. Two were hospitalized and developed an atypical pneumonia. Serum enzymes indicating liver damage were elevated. Microcystin-LR was identified in the bloom material (Pearson et al., 1990). Nevertheless, substances other than microcystin-LR may have been present, and some of the observed effects were probably due to other materials in the water. In the USA and Australia, several different cyanobacterial toxins have been implicated in human illness from certain municipal water supplies, often after algal blooms had been treated with copper sulfate (Bourke et al., 1983; Falconer, 1989; Ressom et al., 1994 ). In most cases, the cyanobacteria and sometimes the toxins involved have been identified, but the levels of toxin associated with illness have not been established in any of the outbreaks. The Palm Island mystery disease, affecting about 140 people, largely children, in Australia, occurred after a dense cyanobacterial bloom on a water supply was treated with copper sulfate. Within a week, severe illness was seen, characterized by vomiting, hepatomegaly, and kidney dysfunction, with loss of electrolytes, glucose, and plasma protein; recovery took 1-3 weeks (Byth, 1980). The cause of illness was not identified until a subsequent cyanobacterial bloom in the same water supply reservoir was shown to be highly toxic (Hawkins et al., 1985). The causative organism was C. raciborskii, containing the cytotoxic alkaloid cylindrospermopsin, a powerful inhibitor of protein synthesis (Terao et al., 1994). Possible liver damage, as evidenced by significant increases in gamma glutamyl transferase, was seen in persons drinking water from supplies containing blooms of Microcystis after treatment with copper sulfate (Malpus Dam, Armidale, Australia) compared with persons drinking uncontaminated water (Falconer, 1989). In Salisbury, Rhodesia, seasonal acute childhood gastroenteritis during the years 1960-1965 was linked to annual blooms of Microcystis in the lake serving as the water supply. An adjacent water supply was not similarly affected and was not associated with this disease (Zilberg, 1966). El Saadi & Cameron (1993) reported on 26 cases (aged 1-64 years) with skin diseases and multiple systemic symptoms associated with exposure (some via drinking-water) to river water or rainwater in Australia during 1991-1992. The water was stored in open tanks and contained Anabaena blooms. Further case-control studies in the same area are ongoing. Recently, an epidemiological survey in Haimen city (Jian-Su province) and Fusui county (Guangxi province) in China found a close relationship between the incidence of primary liver cancer and the use of drinking-water from ponds and ditches (Ueno et al., 1996). In 1993 and 1994, microcystin concentrations ranged from 0.058 to 0.460 J.Lg/litre; the highest concentrations occurred from June to September. Microcystin was not detected in deep well-water. A similar survey on 26 drinking-water samples in the Guangxi province showed a high frequency of microcystins in the water of ponds/ditches and rivers, but no microcystins were found in shallow and deep wells. According to Ueno et al. (1996), the combined effect of microcystin toxin from the drinking-water of ponds/ditches and rivers or both and other carcinogens such as aflatoxin B 1 found in food may be the cause of the high incidence of primary liver cancer in Haimen city and other areas in China. CYANOBACTERIAL TOXINS: MICROCYSTIN-LR 107
Cyanobacterial blooms tend to occur repeatedly in the same water supply. Therefore, some human populations are at risk of repeated ingestion of cyanobacterial toxins. However, the available data are not sufficient to allow a quantitative assessment of human exposure.
7. PROVISIONAL GUIDELINE VALUE
There are insufficient data to derive a guideline value for cyanobacterial toxins other than microcystin-LR. Only a guideline value for this compound is derived. A 13-week study in mice with microcystin-LR (Fawell et al., 1994) is considered the most suitable for the derivation of a guideline value. In this study, a NOAEL of 40 f.!g/kg of body weight per day was determined for liver pathology. A TDI of 0.04 J.lg/kg of body weight per day can be calculated by applying an uncertainty factor of I 000 (I 00 for intra- and interspecies variation, I 0 for limitations in the database, in particular lack of data on chronic toxicity and carcinogenicity) to the NOAEL. An allocation factor of 0.80 is used for the proportion of daily exposure arising from drinking-water, because there is little exposure from any other source and route. The resulting guideline value for total microcystin-LR (free plus cell-bound) is I J..tg/litre (rounded figure) in drinking-water. The guideline value thus calculated is supported by a 44-day study in which pigs were exposed, in their drinking-water, to an extract from M aeruginosa containing microcystin-LR. The guideline value of I J..tg/litre for microcystin-LR is provisional, as the database is limited, new data for the toxicity of cyanobacterial toxins are being generated, and the guideline value covers only microcystin-LR.
8. REFERENCES
A WW A (1995) Cyanobacterial (blue-green algal) toxms a resource gurde Denver, CO, A WW A Research Foundation and American Water Works Association Beasley VR et al. ( 1989) Algae intoxication m livestock and waterfowl. Vetennary climes of North Amenca, food am mal practtce, 5:345-361. Bhattacharya R et al. ( 1996) Liver shce culture for assessmg hepatotoxicity of freshwater cyanobacteria. Human and expenmentaltoxtcology, 15 ·I 05-110. Boles DP et al. (1985) Structural studies on cyanoginosins -LR,-YR,-Y A, and-YM, peptide toxins from 1\ftcrocystts aerugmosa. Journal of the Chenucal Soctety, Perktn transactwns, 1: 2747-2748. Bourke ATC et al (1983) An outbreak of hepato-enteritis (the Palm Island mystery disease) possibly caused by algal mtoxication Tox1con (Suppl 3)'45-48. Brooks WP, Codd GA (1987) Distnbution of Microcystts aerugmosa peptide toxin and interactiOns with hepatic m1crosomes in mice. Pharmacology and toxicology, 60:187-191. Byth S ( 1980) Palm Island mystery disease. Medtcal JOUrnal of Australta, 2"42-49. Carmichael WW (1992) A review: cyanobacteria secondary metabolites- the cyanotoxins. Journal of app!ted bactenology. 72:445-459 Carmichael WW (1994) The toxins of cyanobacteria. Scientific Amencan, 270(1 ):78-86. Carmichael WW, Falconer IR (1993) Diseases related to freshwater blue-green algal toxms, and control measures. In: Falconer IR, ed Algaltoxms m seafood and dnnkmg water. London, Academic Press, pp 187-209 Chu FS, Huang X, We1 RD (1990) Enzyme-linked Immunosorbent assay for microcystins in blue-green algal blooms. Journal ofthe Assoctatwn of Official Analytical Chemtsts, 73:451-456. 108 ORGANIC CONSTITUENTS
Codd GA (1992) Eutrophication, blooms and toxms of cyanobactena (blue-green algae), and health. In: Keller AZ, Wilson HC, eds The changmg face of Europe D1sasters. pollutwn and the env1ronment Vol 1./ Aquatic problems Proceedings of the Fourth Disaster Prevention and LimitatiOn Conference. Bradford, Umversity of Bradford, pp. 33-62. Codd GA, Bell SG (1996) The occurrence and fiile of blue-green algal toxms m fi'eshwaters London, Her Majesty's Stationery Office (National Rivers Authority Rand 0 Report 29). Cousins IT et al. (1996) Biodegradation of m1crocystin-LR by indigenous mixed bacterial populations Water re5earch, 30:481-485. Dillenberg HO, Dehnel MK ( 1960) Toxic water bloom in Saskatchewan. Canadwn Med1cal Assocwtwn JOUrnal, 83: 1151-1154. El Saadi 0, Cameron AS (1993) Illness associated with blue-green algae. Med1cal Journal of Australia, 158:792-793. Eriksson JE et al. (1990) Hepatocellular uptake of 3H-dihydromicrocystin-LR, a cyclic peptide toxin Bwchmuca Bwphys1ca Acta, 1025.60-66. Falconer IR ( 1989) Effects on human health of some toxic cyanobacteria (blue-green algae) in reservoirs, lakes and rivers. Tox1c1ty assessment, 4·17 5-184. Falconer IR ( 1991) Tu m or promotion and liver injury caused by oral consumption of cyanobacteria. Environmental toXIcology and water quality, 6.177-184. Falconer IR ( 1994) Health implicatiOns of cyanobacterial (blue-green algal) toxins. In. Tox1c cyanohacterw - a global penpecllve, pp. 2-6 Adelaide, Australian Centre for Water Quality Research Falconer IR et al. (1988) Oral toxicity of a bloom of the cyanobacterium M1crocyst1s aerugmosa administered to mice over periods up to 1 year. Journal of toxicology and env1romnental health, 24.291-305. Falconer IR et al (l992a) Effect of the cyanobacterial (blue-green algal) toxins from M1crocyst1s aerugmosa on isolated enterocytes from the chicken small intestine. Tox1con, 30·790-793. Falconer IR, Choice A, Hosia W (l992b) Toxicity of edible mussels growing naturally m an estuary during a water bloom of the blue green algae Nodularw sp111n1gena Environmental toxicology and water quality, 7:119-123. Falconer IR et al ( 1994) Toxicity of the blue-green alga (cyanobacterium) Aflcrocyst1s aerugmosa m drinking water to growing pigs, as an animal model for human mjury and nsk assessment. Em•1ronmental toxicology and water quality, 9·131-139 Fastner J, Heinze R, Chorus J (1995) Microcystin content, hepatotoxicity and cytotoxiCity of cyanobacteria m some German water bodies. Water scwnce and technology, 32: 165-170. Fawell JK et al. ( 1993) Blue-green algae and their toxins - analysis, toxicity, treatment and environmental control. Water supp~v. 11·1 09-121. Fawell JK, James CP, James HA (1994) Toxms from blue-green algae toxicological assessment of nucrocystin-LR and a method for 1ts determmatwn in water Medmenham, Marlow, Bucks, Water Research Centre, pp. 1-46. Gorham P (1964) Toxic algae. In Jackson OF, ed. Algae and man New York, NY, Plenum Publishmg Corp., pp. 307-336. Harada K (1994) Strategy for trace analysis of m1crocystins in complicated matrix. In: Tox1c cyanobactena - a global perspecfll'e, pp 49-51 Adelaide, Australian Centre for Water Quality Research. Harada K et al. ( 1988) Improved method for punfication of toxic peptides produced by cyanobacteria. Toxicon. 26.433-439 Hawkins PR et al. (1985) Severe hepatotoxicity caused by the tropical cyanobacterium (blue green algae) Cylmdrospermops1s raCihorskit (Woloszynska) Applted environmental nucrobwlogy, 50:1292-1295. Honkanen RE et al. ( 1990) Characterization of m1crocystin-LR, a potent inhibitor of type 1 and type 2a protem phosphatases Journal of bwlogtcal chenustry, 265: 19401-19404 Hooser SB et al. (1990) M1crocystm-LR-induced ultrastructural changes m rats. r·etennarJ' pathology, 27 9-15 CYANOBACTERIAL TOXINS: MICROCYSTIN-LR 109
Hrudey SE, Lambert TW, Kenefick SL (1994) Health risk assessment of microcystins in drinking water supplies. In: Toxic cyanobacteria - a global perspective, pp. 7-12. Adelaide, Australian Centre for Water Quality Research. Jones G (1996) Tox1c algae study summary Mamtoba Environment Submitted to Health Canada, February. Jones GJ, Orr PT (1994) Release and degradation of mtcrocystin following algicide treatment of a Microcystis aerugmosa bloom in a recreational lake, as determined by HPLC and protein phosphatase inhibition assay. Water research, 28:871-876. Kenefick SL et al. (1993) Toxin release from Microcystis aeruginosa after chemical treatment. Water sc1ence and technology, 27:433-440. Kotak BG et al. (1993) Occurrence and toxicological evaluation of cyanobacterial toxins in Alberta lakes and farm dugouts. Water research, 27:495-506. Kotak BG, Lam AK-Y, Prepas EE (1995) Variability of the hepatotoxin mtcrocystm-LR in hypereutrophtc dnnking water lakes. Journal ofphvcology, 31 :248-263. Lambert TW et al. (1994a) Quantitation of the microcystin hepatotoxins in water at environmentally relevant concentrations with the protein phosphatase bioassay. Environmental sc1ence and technology, 28·753-755 Lambert TW, Holmes CFB, Hrudey SE (1994b) Microcystin class of toxins: health effects and safety of drmking water supplies. Environmental rev1ew, 2:167-186 Lawton LA, Edwards C, Codd A (1994) Extraction and high-performanc6liquid chromatographic method for the determination ofmicrocystins in raw and treated waters. Ana~vst, 119:1525-1530. Lin J-R, Chu FS (1994) In v1tro neutralizatiOn of the inhibitory effect of microcystin-LR to protein phosphatase 2A by antibody against the toxin Tox1con, 32.605--613 Mackintosh C et al (1990) Cyanobacterial microcystin-LR ts a potent and specific inhtbttor of protein phosphatases 1 and 2A from both mammals and higher plants. Federatw11 of European Biochemical Soc1et1es fetters, 264:187-192 Meriluoto JA et al. (1990) Synthesis, organotropism and hepatocellular uptake of two tritium-labeled epimers of dihydromicrocystm-LR, a cyanobactenal peptide toxm analog. Tox1con. 28.1439-1446. Nagata S et al (1995) Determmation of mtcrocystin in environmental water by highly sensittve immunoassay. Japa11ese JOUrnal of toxicology and environmental health, 41·1 0 Negri AP, Jones GJ (1995) Btoaccumulation of paralytic shellfish poisoning toxms (PSP) from the cyanobacterium Anabaena c1rcinglis by the fresh water mussel Alathyria. Toxicon, 33:667--678 N1shtwaki-Matsush1ma R et al ( 1992) Liver tumor promotion by the cyanobacterial cyclic peptide toxin m icrocystin LR Journal of cancer research and chmcaf oncology, I 18:420-424. Ohtam I, Moore RE, Runnegar MTC (1992) Cylindrospermopsm: a potent hepatotoxin from the blue-green algae C!'imdrospermops1s rac1borsk11. Journal of the Amencan Chenucal Soc1ety, 114:7942-7944 Park HO et al (1993) Hepatotoxin (microcystin) and neurotoxin (anatoxin-a) contained in natural blooms and strams of cyanobacteria from Japanese fresh waters. Natural toxms, 1.353-360. Pearson MJ et al. (1990) Tox1c blue-green algae. A report by the UK National Rivers Authority, pp. 1-128 (Water Quality Series No 2). Repavich WM et al. (1990) Cyanobacteria (blue-green algae) in Wisconsm waters acute and chronic toxicity. Water research, 24:225-231. Ressom R et al (1994) Health effects of tox1c cya11ohactena (blue-green algae) Canberra, Australian National Health and Medical Research Counctl, pp 1-108 Rinehart KL, Namikoshi M, Choie BW ( 1994) Structure and biosynthesis of toxins from blue-green algae (cyanobacteria). Journal of app!ted phycofogy, 6·159-176. Robmson NA et al (1991 a) Ttssue distributiOn, excretion and hepatic biotransformation of mtcrocystin-LR in mice. Journal of pharmacaiO?J' and expenme11tal therapy, 256 176-182. Robmson NA, Matson CF, Pace JG ( 1991 b) Assoctatton of microcystm-LR and its biotransformatiOn product wtth a hepattc-cytosohc protem. Joumaf of bwchenucaftoxlcofo?J', 6:171-180. Runnegar MT, Gerdes RG, Falconer IR ( 1991) The uptake of the cyanobacterial hepatotoxin microcystm by i>olated rat hepatocytes. Tox1co11, 29 43-51 110 ORGANIC CONSTITUENTS
Runnegar MT, Kong SM, Berndt N (1993) Protein phosphatase inhibition and in vivo hepatotoxicity ofmicrocystins. AmencanJournal ofphysiology, 265:G224-G230. Siegelman HW et al. (1984) Toxins of M1crocyst1s aeruginosa and their hematological and histopathological effects. In: Ragelis EP, ed. Seafood toxins. American Chemical Society, pp. 407-413 (American Chemical Society Symposium Series No. 262). Skulberg OM et al. (1993) Taxonomy of toxic Cyanophyceae (cyanobacteria). In: Falconer IR, ed. Algal toxins in seafood and drinking water London, Academic Press, pp. 146-164. Terao Ketal. (1994) Electron microscopic studies on experimental poisoning in mice induced by cylindrospermopsin isolated from blue-green algae Umzakia naatans. Toxicon, 32:833-843. Tomoyasu U, Keiji M (1996) Development of the quantity determination method for Microcystin using liquid chromatography-mass spectrometry. Journal of the Japan Water Works Association, 65(7):25-35. Tsuji K et al. ( 1994) Stability of microcystins from cyanobacteria: effect of light on decomposition and isomerization. EnVIronmental science and technology, 28:173-177. Ueno Y et al. (1996) Detection of microcystins, a blue-green algal hepatotoxin, in drinking water sampled in Haimen and Fusui, endemic areas of primary liver cancer in China, by highly sensitive immunoassay. Carcinogenesis, 17(6): 1317-1321. Utkilen H, Gjolme N (1992) Toxin production by M1crocystis aerugmosa as a function of light in continuous cultures and its ecological significance. Applied environmental microbiology, 58:1321-1325. van der Westhuizen AJ, Eloff JN (1985) Effect of temperature and light on the toxicity and growth of the blue-green alga Microcyst is aeruginosa. Zeitsc11rift fur Pjlanzenphysiolog1e, 110:157-163. Wannemacher RW (1989) Chemical stablltty and laboratory safety of naturally occurrmg toxins. Fort Detnck, Frederick, MD, US Army Medical Research Institute of Infectious Disease Watanabe MF, Oishi S (1983) A highly toxic strain of blue green alga Microcyst1s aeruginosa Isolated from Lake Suwa. Bulletin of the Japanese Soctety ofScientific F1shenes, 49:1759. Watanabe MF, Oishi S (1985) Effects of environmental factors on toxicity of a cyanobacterium (Microcystis aerugmosa) under culture conditions. Applied environmental microbwlogy, 49:1342-1344. Watanabe MF et al. (1989) Heptapeptide toxin production during the batch culture of two M1crocystis species ( cyanobacteria). Journal ofapplied phycology, I: 161-165. Wicks JR, Theil PG (1990) Environmental factors affecting the production of peptide toxins in floating scums of the cyanobacterium M1crocysfiS aeruginosa in a hypertrophic African reservoir. Environmental science and technology, 24: 1413-1418. Zilberg 8 (1966) Gastroenteritis in Salisbury European children- a five-year study. Central African Journal ofmedlcme, 12:164-168 (1966) [cited in Falconer, 1994]. EDETIC ACID (EDTA)
First draft prepared by J. Fawell Water Research Centre, Medmenham, United Kingdom
In view of the possibility of zinc complexation, a provisional guideline value for edetic acid (EDT A) of 200 !J.gllitre was derived in the 1993 Guidelines for drinking-water quality, assuming consumption of 1 litre of drinking-water by a 10-kg child and allocating 10% of the TDI (190 Jlglkg of body weight) to drinking water. This TDI was derived from the 1973 JECF A ADI of 2.5 mg/kg of body weight for calcium disodium edetate as a food additive (1.9 mg/kg of body weight as the free acid). An extra uncertainty factor of 10 was introduced to reflect the fact that the JECFA ADI had not been considered since 1973, as well as concern over zinc complexation. The guideline value was considered to be provisional because it was based on a rather old (1973) evaluation by JECF A. As JECFA evaluated the toxicological studies available on sodium iron EDT A and the calcium/sodium salts of EDT A in 1993, the Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality recommended that EDTA be re-evaluated in the 1998 Addendum.
1. GENERAL DESCRIPTION
1.1 Identity
CAS no.: 60-00-4 Molecular formula: CJOHI6N20s
Edetic acid (ethylenediaminetetraacetic acid) and its salts are commonly referred to as EDT A. Other names include N,N -1 ,2-ethanediylbis[N-( carboxy methyl)glycine ], Versene acid, and (ethylenedinitrilo )tetraacetic acid.
1.2 Physicochemical properties
Property Value Physical appearance Colourless crystals Solubility in water 0.5 g/litre at 25°C
1.3 Organoleptic properties
EDTA has a slightly salty taste. 112 ORGANIC CONSTITUENTS
1.4 Major uses
EDT A has been used extensively in medicine as a chelating agent for Jhe removal of toxic heavy metals. The disodium salt of EDT A is a common component in many eye drops and contact lens wetting and cleansing solutions. EDT A is also used in a number of personal care and hygiene products, such as shampoos, liquid soaps. creams, and lotions. Household disinfectants often contain EDT A, especially if fatty acid soaps are used in the disinfectant formulation. These soaps are sensitive to calcium and magnesium, and the chelating agent prevents the formation of hard-water soap curds (Hart, 1984). EDTA is also used as a food additive in a range of products, including canned shrimp and prawns, canned mushrooms, and frozen french fries. It is added to salad dressings to prevent rancidity. EDTA is used in many industrial processes, m agriculture, m photochemicals, pharmaceuticals, and textiles, and in galvanizing and paper manufacturing. The usage of EDT A in West Germany in 1986 by industry was: metal processing and galvanizing technology, 30%; detergents. 20%; photographic industry, 20%; textiles, 10%; paper, 5%; and miscellaneous (antioxidants in soaps and cosmetics, pharmaceuticals, and foodstuffs), 15%. The total use over the year was about 15 000 t (Brauch & Schullerer, 1987).
1.5 Enviroumental fate
Once EDT A is present in the aquatic environment, its speciation will depend on the water quality and the presence of trace metals with which it can combine. The fate and behaviour of the different complexes may vary considerably. The iron(III) EDT A complex is the most labile because it is very photo-active. Svenson et al. (1989) calculated a half-life of 11 minutes for the photolysis of iron( Ill) EDT A dissolved in water and irradiated with sunshine equivalent to the annual maximum intensity. The photo-oxidation of free EDTA in water at pH 9.0 has a measured half life of 36 years for an initial hydroxyl radical concentration of 5 x 1o- 9 mol/litre and a 12-hour daylight duration. The iron(III) EDTA will exchange slowly with other trace metals after discharge to the environment, a process that will be dependent upon the pH of the water, as each trace metal has an optimum pH for chelation. Other metal complexes of EDTA are much more persistent and are not readily biodegraded in the aquatic environment. In soil-water systems, degradation has been observed, but the extent varies with soil type and length of exposure. The removal of EDT A from communal wastewater by biodegradation in sewage purification plants is very limited. A limited removal takes place by adsorption on sludge. Similarly, elimination of EDTA by different treatment methods of drinking-water is negligible, including filtration on activated carbon. The most effective elimination is by ozonation (Gilbert & Beyerle, 1992). There has been concern that EDTA mobilizes heavy metals in the environment. However, based on stoichiometry, 40 Jlg of EDTA per litre (the maximum concentration observed in the Rhine and Meuse rivers) would complex 4-15 Jlg of EDETICACID (EDTA) 113 metals per litre at most, and this would be likely to pose problems for drinking-water only with regard to cadmium. A further modifying factor is that the effect on cadmium leaching will be limited because the EDTA is primarily bound to other metals at these concentrations (van Dijk-Looyard et al., 1990). For the majority of fresh waters, EDTA will be associated largely with calcium, provided that the EDTA is not present in stoichiometric excess. For waters of pH lower than 6.0, however, competition from hydrogen ions for available ligand assumes greater importance.
2. ANALYTICAL METHODS
EDTA can be analysed by potentiometric stripping analysis (Fayyad et al., 1988). This method, which has been used to detect EDTA in a wide variety of wastewater and natural water samples, has a detection limit of l 1-1g/litre.
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Water
Most of EDTA's uses will result in its release to the aquatic environment. It has been estimated that concentrations of 50-500 1-1g/litre are present in wastewaters. Annual average concentrations of EDT A in European surface waters ranged between > 1 and >60 1-1g/litre, and a concentration of 900 1-1g/litre was found in the Zerka River in Jordan (van Dijk-Looyard et al., 1990). Measured concentrations in natural waters were also reported to range from 10 to 70 1-1g/litre, with a median value of 23 1-1g/litre (Frank & Rau, 1990). Mean EDTA concentrations at 45 different sampling points on 29 different rivers of Germany in 1993 ranged between almost 50 1-1g/litre (Lippe River at Wesel) and a few 1-1g/litre, with most annual mean values being between 5 and 15 1-1g/litre (EFA-Germany, 1995). EDTA has also been detected in surface waters and in drinking-water prepared from surface waters at concentrations of 10-30 1-1g/litre (van Dijk-Looyard et al., 1990).
3.2 Food
EDTA's use as a food additive has been limited. The maximum levels of EDTA in canned shrimp and prawns, canned mushrooms, and frozen french fries are 250, 200, and 100 mg/kg, respectively (Smith, 1990).
3.3 Estimated total exposure and relative contribution of drinking-water
Human exposure to EDT A arises directly from its use in food additives, disinfectants, medicines, and personal care and hygiene products. Exposure to EDTA from drinking-water is probably very small in comparison to exposure from other sources. 114 ORGANIC CONSTITUENTS
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
4.1 Metabolism in laboratory animals and humans
Foreman et al. (1953) examined the metabolism of 14C-labelled calcium disodium EDTA in the rat in two series of experiments involving the administration of doses of 50 mg/kg of body weight, intraperitoneally, intravenously, intramuscularly, or orally by intubation, to Sprague-Dawley rats. The studies indicated that calcium disodium EDT A was poorly absorbed from the gastrointestinal tract. About 80-95% of the dose appeared in the faeces after 24 hours. The amount absorbed in 24 hours, determined from the quantity found in the tissues and urine, ranged from 2 to 18%, with most of the values being between 2 and 4%. At the low pH of the stomach, the calcium chelate is dissociated with subsequent precipitation of the free acid, and this is only slowly redissolved as it passes through the alimentary tract. Increasing the dose is not necessarily a good way of increasing the amount absorbed, as the rats exhibited diarrhoea at higher doses. Srbrova & Teisinger (1957) confirmed the dissociation of the calcium chelate in the stomach. When a dose of 200 mg of calcium disodium EDTA was introduced into the duodenum of rats, the authors found absorption to be 6.5-26%. Experiments in humans also revealed poor absorption; only 2.5% of a 3-g dose of calcium disodium EDTA was excreted in the urine (Srbrova & Teisinger, 1957). Only 5% of a dose of 1.5 mg of 14C-labelled calcium disodium EDTA given in a gelatin capsule to normal healthy men was absorbed (Foreman & Trujillo, 1954). EDTA has also been shown to be rapidly excreted from the body. Intravenous doses of 3 g of radio labelled calcium disodium EDT A, given to two subjects, were almost entirely excreted within 12-16 hours (Srbrova & Teisinger, 1957). A summary of a 1956 Ph.D. thesis by Chan in Anonymous (1964) reported biochemical studies with disodium EDTA. In a study in rats, 32 hours following administration of a single oral dose of 95 mg of disodium EDTA per rat, 93% of the dose was recovered from the colon. After doses of 47.5, 95, and 142.5 mg of disodium EDTA, the amount of EDT A recovered in the urine was directly proportional to the dose given, suggesting that EDT A was absorbed from the gastrointestinal tract by passive diffusion.
4.2 Metal complexation with EDTA
EDT A is a hexadentate chelator capable of combining stoichiometrically with virtually every metal in the periodic table (Chaberck & Martell, 1959). With divalent or trivalent metal ions, a neutral or anionic metal chelate results. The metal is largely prevented from reacting with competing anions, and its solubility is greatly increased. The effectiveness ofEDTA as a chelate for a particular metal ion is given by its stability constant with the metal ion. Chelation potential is affected by pH, the molar ratio of chelate to metal ion, and the presence of competing metal ions capable of forming complexes with EDT A (Plumb et a! , 1950; Martell, 1960; Hart, 1984 ). The stability constants for different metal-EDTA complexes vary considerably, and EDETICACID (EDTA) 115 any metal that is capable of forming a strong complex with EDT A will at least partially displace another metal. Of the nutritionally important metals, Fe 3+ has the highest stability constant 2 2 (log k = 25.1 ), followed by Cu + with 18.4, Zn + with 16.1, Fe'+ with 14.6, Ca2+ with 10.6, Mg2+ with 8.7, and Na+ with 1.7 (West & Sykes, 1960). The situation is somewhat complicated by each metal having an optimum pH for chelate formation, 3 2 2 ranging from pH 1 for Fe + to pH 3 for Cu +, pH 4 for Zn \ pH 5 for Fe'+. pH 7.5 for Ca2+, and pH 10 for Mg2+ (West & Sykes, 1960). When sodium iron EDTA is ingested with foods, the Fe3+ ion would be expected to remain firmly bound to the EDTA moiety during passage through the gastric juice, but it could be exchanged for Cu2+, Zn2+. Fe2+, or Ca2+ in the duodenum (WHO, 1993). In biological systems, Ca2+ will usually be most accessible to EDT A. In general, zinc seems to be the next most accessible. About 80% of the zinc in liver is freely available to EDTA. The overall availability of the other physiologically important metals is probably in the order copper > iron > manganese > cobalt (Chenoweth, 1961 ). EDTA removes about 1.4% of the total iron from ferritin at pH 7.4 to form an iron chelate (Westerfield, 1961 ). Perry & Perry (1959) investigated the changes in normal concentrations of trace metals in human urine following administration of EDT A. Calcium disodium EDT A had been observed to lower the level of cholesterol in human plasma, and therefore this study investigated metal concentrations in consecutive 24-hour urine samples from hypercholesterolaemic patients before, during, and after the intravenous admiriistration of calcium disodium EDT A. The results indicated a 10-fold increase i~ urinary excretion of zinc during the administration of calcium disodium EDTA. A smaller effect on cadmium and manganese may have occurred, but the results were not clear, because some of the control concentrations were too low to quantify. There were also suggestive increases in the excretion of lead and vanadium. Foreman (1961) reported that EDTA enhanced the excretion of cobalt, mercury, manganese, nickel, lead, thallium, and tungsten. When EDTA is present in food, iron (primarily Fe3+) remains complexed with EDT A under the acidic conditions prevailing in the stomach. The chelate holds the iron in solution as the pH rises in the upper small intestine, but the strength of the complex is progressively reduced, allowing at least partial exchange with other metals and the release of some of the iron for absorption. There is convincing evidence that iron chelated by EDT A (sodium iron EDT A) is available for absorption via the physiologically regulated pathways responsible for iron uptake (Candela et al., 1984). The results of the absorption studies with sodium iron EDTA indicate that iron is dissociated from the EDT A moiety prior to absorption. Hurrell et al. (1994) examined the influence of sodium iron(III) EDT A, used as a food fortificant, on the metabolism of calcium, zinc, and copper in the rat. The results showed that changing the iron fortificant from iron sulfate to sodium iron(III) EDTA increased the apparent zinc absorption, retention, and excretion in the rats receiving the zinc-deficient diets. The authors suggest that the study appears to indicate a beneficial effect on zinc nutritional status following addition of sodium iron(III) EDTA as an iron food fortificant. However, the results are complicated by the fact that iron status is an important factor in zinc metabolism and a high level of 116 ORGANIC CONSTITUENTS iron inhibits the availability of zinc. In addition, the diets contained soy bean protein. which is high in phytate, an inhibitor of both iron and zinc absorption. These factors make it difficult to draw any definite conclusions from this study.
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
Acute toxicity studies have been carried out with disodium EDTA and calcium disodium EDT A in laboratory animals. LD 50 values (mg/kg of body weight) reported in these studies are summarized below:
Rat Oral: 2000-2200, Na2EDTA (Yang, 1952)
Oral: 10 000 ± 740, CaNa2EDTA (Oser et aL 1963)
Rabbit Oral: 2300, Na2EDTA (Shibata, 1956)
Oral: ~7000, CaNa2EDTA (Oser et al., 1963)
Intraperitoneal: ~500, CaNa2EDTA (Bauer et al., 1952)
Dog Oral: ~12 000, CaNa2EDTA (Oser et al., 1963)
Oser et al. (1963) reported that the oral LD 50 in rats was not affected by the presence of food in the stomach or by pre-existing deficiency in calcium, iron, copper, or manganese.
5.2 Short-term exposure
In a study involving the intraperitoneal administration of 250 or 500 mg of calcium disodiurn EDT A per kg of body weight per day to groups of five male rats for 3-21 days, it was reported that weight gain was satisfactory and that the histology of the lung, thymus, liver, spleen, adrenal, small gut, and heart was normal; there was a mild to moderate effect on the kidney (Reuber & Schmieller, 1962). Groups of five male rats were given 250, 400, or 500 mg of disodium EDT A per kg of body weight per day intraperitoneally for 3-31 days: some groups were observed for an additional 2 weeks. At 500 mg/kg of body weight per day, all rats became lethargic and died within 9 days; the kidneys were pale and swollen, and there was moderate dilation of the bowel and subserosa! haemorrhages. Histological examination of a number of organs showed lesions only in the kidneys. Animals at the 400 mg/kg of body weight per day dose level died within 14 days; kidney and bowel symptoms were similar to those seen at the high dose. One rat in the 250 mg/kg of body weight per day dose group showed haemorrhage of the thymus. All three groups exhibited varying degrees of damage to the kidney, with recovery occurring on withdrawal of the disodium EDT A (Reuber & Schmieller, 1962). Four groups of one male and three female mongrel dogs were fed diets containing 0, 50, 100, or 200 mg of calcium disodium EDTA per kg of body weight EDETIC ACID (EDTA) ll7 per day for 12 months. At the end of the study, all dogs appeared to be well, and there were no significant changes in blood or urine analysis. Gross and microscopic examinations of the major organs were normal (Oser et al., 1963).
5.3 Long-term exposure
The long-term toxicity of EDT A is complicated by its ability to chelate essential and toxic metals, both in water and in animals. Toxicity data are therefore equivocal and difficult to interpret. An early study by Krum (1948) demonstrated no adverse effect on weight gain, appetite, activity, and appearance in rats fed for 44-52 weeks on a diet containing 0.5% disodium EDTA. A 2-year study was carried out in which groups of rats were fed 0, 0.5, 1, or 5% disodium EDTA. The highest dose group showed a reduced food intake compared with the other groups and also suffered diarrhoea. No significant effects on weight gain were noted, nor were blood coagulation time, red blood cell counts, or bone ash adversely affected. Mortality of the animals could not be correlated with the level of disodium EDT A, as the highest mortality was observed in the control group and was due to pneumonia. Gross and microscopic examinations of the major organs did not reveal any significant differences between the groups (Yang, 1952). In another study, groups of 25 male and 25 female rats were fed diets containing 0, 50, 125, or 250 mg of calcium disodium EDTA per kg of body weight per day for 2 years, and the study was carried on through four successive generations. The rats were mated after 12 weeks of feeding and were allowed to lactate for 3 weeks, with 1 week's rest before producing a second litter. Ten male and 10 female rats from the F1 generation and similar F2 and F3 generations were allowed to produce two litters. With the second litter in the F 1, F 2, and F 3 generations, only the control and the 250 mg/kg of body weight per day dose groups were kept until the end of the 2-year study on the F0 generation. No significant abnormalities in appearance or behaviour were noted during the 12 weeks of the post-weaning period in all generations. The experiments showed no statistically significant differences in weight gain, food efficiency, haematopoiesis, blood sugar, non-protein nitrogen, serum calcium, urine, organ weights, and histopathology of the liver, kidney, spleen, heart, adrenals, thyroid, and gonads (Oser et al., 1963). Fifty weanling albino rats were fed a low-mineral diet (0.54% calcium and 0.013% iron) with the addition of 0, 0.5, or 1% disodium EDT A or 0.5 or 1% calcium disodium EDT A for 205 days. Diarrhoea was observed in the 1% disodium EDT A group, along with other abnormalities: growth retardation of the males, lowered erythrocyte and leukocyte counts, a prolonged blood coagulation time, slightly but significantly raised blood calcium level, a significantly lower ash content of the bone, and considerable erosion of the molars. Gross and histological examinations of the major organs revealed nothing abnormal. Rats fed for 220 days on an adequate mineral diet containing 1% disodium EDT A showed no evidence of dental erosion (Chan, 1956).
Groups of 50 male and 50 female B6C3F 1 mice received trisodium EDTA in the diet at concentrations of 3, 750, or 7500 mg/kg of feed for 103 weeks, followed 118 ORGANIC CONSTITUENTS by 1 week during which the mice were fed standard diet without EDTA. The animals were examined twice per day for signs of toxicity. Gross and histopathological examinations of major organs and tissues were performed on animals found dead or moribund and on those sacrificed at the end of the study. Body-weight gain was decreased in high-dose males during the second year of the study, although no statistical analysis was presented. No treatment-related tumours or non-neoplastic lesions were observed in this study (NCI, 1977).
5.4 Reproductive and developmental toxicity
Groups of six rats were maintained on diets contammg 0.5, 1, or 5% disodium EDTA for 12 weeks. The only toxic symptoms observed were diarrhoea and a reduction in food consumption at the 5% level. When the animals were 100 days old, mating was carried out and was repeated 10 days after weaning of the first litters. The animals given 5% disodium EDTA did not produce any litters. The other dose groups produced normal first and second litters (Yang, 1952). Oser et al. (1963) carried out a four-generation study in which groups of rats received calcium disodium EDTA at doses of 50, 125, or 250 mg/kg of body weight per day via the diet. There were no reproductive or teratogenic effects noted in any of the three generations of offspring. Groups of pregnant Sprague-Dawley rats were given diets containing 2 or 3% disodium EDTA from day 1 through day 21 of gestation. A further group of pregnant rats was fed diets containing 3% disodium EDTA from day 6 to day 14 of gestation, whereas a third group was given diets containing 3% disodium EDTA and 1000 mg of zinc per kg of feed from day 6 to day 21 of gestation. The control animals received a standard diet that contained 100 mg of zinc per kg of feed. In rats fed the 2% disodium EDT A, litter size was normal and fetuses were alive. Gross abnormalities were evident in 7% of the treated fetuses, compared with 0% in the control group. In rats fed 3% disodium EDTA, almost half of the implantation sites had dead fetuses or resorptions. Full-term young were significantly smaller than controls, and 100% of them were malformed. Maternal toxicity was indicated by diarrhoea and was observed in both the 2% and 3% dose groups. The malformations included severe brain malformations, cleft palate, malformed digits, clubbed legs. and malformed tails. Supplementation of the diet with 1000 mg of zinc per kg of feed prevented these detrimental effects, and it was suggested that the teratogenic effects of EDTA given to rats at very high levels were due to zinc deficiency (Swenerton & Hurley, 1971). Schardein et al. (1981) performed teratogenesis studies with EDT A and its salts in rats. EDTA (967 mg/kg of body weight), disodiurn EDTA (1243 mg/kg of body weight), trisodium EDTA (1245 mg/kg of body weight), tetrasodium EDTA (1374 mg/kg of body weight), and calcium disodium EDTA (1340 mg/kg of body weight) were administered orally (by intubation) to groups of 20 inseminated female rats during organogenesis. The dosing regimen was twice daily on days 7-14 of gestation. There were two control groups. One group received 1.0 m! of phosphate buffer per kg of body weight twice daily to serve as a vehicle control group, and the other remained untreated and served as an untreated control group. The dams were EDETIC ACID (EDTA) 119 killed on day 21 of gestation, and litter data for each dam were collected. The fetuses were then examined for gross external anomalies, visceral abnormalities, and skeletal anomalies. Diarrhoea was apparent in all the treated groups, and reduced activity was observed in a few of the dams. The diarrhoea generally occurred following treatment and disappeared on the last day of dosing or the day after. Three dams died during treatment with disodium EDT A. Examination of these did not indicate any gross abnormalities. Food intake was slightly reduced compared with controls during the treatment period of days 7-14 but was comparable to the controls during the pre-treatment period (gestation days 0-7) and post-treatment period (gestation days 14-21 ). None of the test compounds significantly affected litter size at term when compared with either control group. The mortality index of the offspring in all treated groups as measured by post-implantation loss was also comparable to both control groups. A total of 24 pups from the treated groups had abnormalities, including bifid vertebrae, agenesis of the ribs, inhibition of osteogenesis of the skull or ribs, and malformed ribs. There was, however. no pattern between treatment with any particular compound and the appearance of anomalies. In addition, the untreated control group had eight pups with some major defect. The results of these studies demonstrated that there was an absence of teratogenic effects even at doses that were maternally toxic. An important study was carried out by Brownie et al. (1986), who investigated the teratogenic effect of calcium EDT A in rats and the protective effect of zinc. Pregnant Long-Evans rats were randomly assigned to 11 treatment groups corresponding to differing doses of calcium EDTA (2, 4, 6, or 8 mmollm2 per day), zinc EDTA (8 or 20 mmol!m2 per day), and zinc calcium EDTA (8 or 20 mmol/m2 per day), plus controls. Each group contained 20 rats, except for the control group receiving 0.9% NaCl, which contained 30 rats. A further 12 animals per group were used for maternal plasma and liver and fetal zinc analysis. The rats were treated by subcutaneous injection of the chelating agent or saline solution on days 11 through 15 of gestation. Results showed increases in several abnormalities (e.g. submucous cleft, cleft palate, curly tail, abnormal rib and vertebrae) with increasing doses of calcium EDTA. No malformations were seen with zinc EDT A at either dose or with zinc calcium EDTA at the lower dose. However, submucous cleft was seen in 6 of 20 litters from the dams receiving the higher dose of zinc calcium EDT A. It was concluded that calcium EDTA is teratogenic in rats at concentrations that, except for decreased weight gain, produce no discernible toxicity to the dam, and that protection is afforded by incorporating zinc in the chelate. A study was carried out by Kimmel ( 1977) in which groups of pregnant CD rats were treated with disodium EDT A via a number of different routes of exposure. Forty-two rats received a dose of 954 mg/kg of body weight per day via the diet. Twenty-two received a dose of 1250 mg/kg of body weight per day, which was split into a dose of 625 mg/kg of body weight twice per day by gastric intubation. Eight rats received 1500 mg/kg of body weight per day as a split dose of 750 mg/kg of body weight twice per day by gastric intubation. Twenty-five rats received a dose of 375 mg/kg of body weight per day by subcutaneous injection. All the animals were dosed on gestation days 7-14. Fetuses were removed on day 21 of gestation and examined for gross and visceral abnormalities. The results showed that for the 120 ORGANIC CONSTITUENTS dietary group, there were no maternal deaths but a significant increase in fetal death compared with the controls, and 71% of the fetuses were. malformed. In the group that was administered 625 mg/kg of body weight per day by gavage, only 64% of the dams survived treatment. Those that survived exhibited fetal resorptions comparable to controls, and 20.5% of the fetuses were malformed. In the group administered 750 mg/kg of body weight per day, seven of the eight dams died. The group receiving subcutaneous injection of disodium EDTA showed a 76% survival of the dams. However, there was a significant increase in fetal resorptions compared with controls, although the proportion of malformed fetuses was similar to the controls. This study demonstrates that the route of exposure to EDTA is an important factor in the determination of the toxicity and teratogenic potential of EDTA.
5.5 Mutagenicity and related end-points
Trisodium EDTA was tested for its mutagenic potential in the L5178Y tk+ /tk mouse lymphoma cell forward mutation assay. Two experiments were performed with S9 metabolic activation system and three without S9 at concentrations of EDT A up to 5000 mg/litre. No mutagenicity was observed either with or without the S9 (McGregor et al., 1988). Trisodium EDTA was also tested in Salmonelt'a typhimurium TA 98, TA 100, TA1535, TA1537, and TA1538 and in Escherichia colz WP uvrA in the presence and absence of the S9 metabolic activation system. There was no evidence of any mutagenic potential in any of these bacterial systems (Dunkel et al., 1985).
5.6 Carcinogenicity
In an experiment in which groups of 50 male and 50 female B6C3F 1 mice received trisodium EDTA in the diet at concentrations of 3, 750, or 7500 mg/kg of feed for 103 weeks, followed by 1 week during which the mice were fed standard diet without EDTA, no treatment-related tumours were observed (NCI, 1977).
6. GUIDELINE VALUE
JECF A evaluated the toxicological studies available on sodium iron EDT A in 1993 (WHO, 1993), and there was no further information. compared with its 1973 evaluation (WHO, 1974), of noteworthy importance regarding the toxicity of EDTA and its calcium/sodium salts. Concern has been expressed over the ability of EDTA to complex zinc and therefore reduce its availability. However, this is only of significance at elevated doses substantially in excess of those encountered in the environment. The use of an additional uncertainty factor and the assumption of a 10-kg child were therefore considered inappropriate. A guideline value for EDT A in drinking-water can be derived by allocating 1% of the JECFA ADI (1.9 mg/kg of body weight as the free acid) to drinking-water (because of the potential for significant exposure from food owing to its use as a food additive). Therefore, assuming a 60-kg adult ingesting 2 litres of drinking-water EDETIC ACID (EDTA) 121 per day, the guideline value for EDTA (free acid) is 600 ~J.g/litre (rounded figure). This value is no longer considered to be provisional.
7. REFERENCES
Anonymous (1964) Summaries of toxicological data. Toxicology of EDTA. Food and cosmet1cs toxicology, 2:763-767. Bauer RO et al. (1952) Acute and subacute toxicity of ethylene diamme tetraacetiC acid (EDTA) salts Federatwn proceedmgs. 11:321. Brauch HS, Schullerer S ( 1987) Verhalten von ethylamintetraacetat (EDTA) und nitriloacetat (NTA) bei der trinkwasseraufbereitung. [Behaviour of ethylenediaminetetraacetic acid (EDTA) and nitrilotriacetic acid (NTA) in drinking water treatment.] Vom Wasser, 69·155-164. Brownie CF et al. (1986) Teratogenic effect of calcium edetate (CaEDTA) in rats and the protective effect of zinc. Toxicology and applied pharmacology, 82:426-443. Candela E et al. (1984) Iron absorption by humans and swine from Fe(lll)-EDTA. Further studies. Journal ofnutntwn, 114.2204-2211. Chaberck SA, Martell AE (1959) Orgamc sequestermg agents. New York, NY, John Wiley & Sons, Inc. Chan MS ( 1956) Some tox1colog1cal and physiological stud1es of ethylenedwnunetetraacet1c ac1d in the albino rat Ph.D. thesis dated March 1956, submitted to the Umversity of Massachusetts, Amherst, MA [summarized in Anonymous, 1964]. Chenoweth MB (1961) Known and suspected role of metal coordination in drug actions. Ill. Pharmacology and toxicology of chelating agents Federatwn proceedmgs, 20 (Suppl. I 0).125. Dunkel VC et a! ( 1985) Reproducibility of microbial mutagemcity assays: II. Testmg of carcinogens and noncarcinogens in Salmonella typh1murium and Eschenchw coli. Environmental mutagenesis, 7 (Suppl. 5:1):1-248 [cited m WHO, 1993]. EFA-Germany (1995) Report of the Environmental Federal Agency of Germany for the Federal Lande September 1995. Fayyad M, Tutunji M, Taha Z (1988) Indirect trace determination of EDTA in waters by potentiometric stripping analysis. Analyt1calletters, 21(8): 1425-1432. Foreman H ( 1961) Use of chelating agents in treatment of metal poisoning (with special emphasis on lead). Federation proceedmgs, 20 (Suppl. I 0): 191. Foreman H, Trujillo TT (1954) The metabolism of 14 C labeled ethylenediaminetetraacetic acid in human beings. Journal of laboratory and clinicalmed1cme, 43: I 045-1053. Foreman H, Vier M, Magee M (1953) The metabolism of C 14 labelled ethylenediaminetetraacetic acid in the rat. Journal ofbwlog1cal chemistry, 203·1045-1053 Frank R, Rau H (1990) Photochemical transformation in aqueous solution and possible environmental fate of ethylenediammetetraacetic acid (EDTA). Ecotox1cology and envtronmental safety, 19(1):55-63. Gilbert E, Beyerle M (1992) Elimination of nitrilotriacetic ac1d (NTA) and ethylenediamine tetraacetic acid (EDTA) in the ozonation step of the drinking water treatment process. Journal of water supp~v research and technology, Aqua, 41 (5).269-276. Hart JR (1984) EDTA-type chelating agents in everyday consumer products: some medicinal and personal care products. Journal of chemical educatwn, 61 (12): 1060-1061. Hurrell RF, Ribas S, Davidson L (1994) Sodium iron EDTA as a food fortlficant: influence on zinc, calcium and copper metabolism in the rat. Bnllsh ;ournal of nutntwn, 71.85-93. Kimmel CA ( 1977) Effect of route of administration on the toxicity and teratogenicity of EDTA m the rat To.ncology and applied pharmacology, 40:299-306. Krum JK (1948) Umversity of Massachusetts (thesis) [cited in WHO, 1993]. Martell AE (1960) Chelat10n· stability and selectivity. Annals of the New York Academy of Sciences, 88:284-292. McGregor DB et al. ( 1988) Responses of the L5178Y tk+/tk- mouse lymphoma cell forward mutation assay Ill 72 coded chemicals. Environmental and molecular mutagenesiS, 12:85-154. 122 ORGANIC CONSTITUENTS
NCI (1977) Bwassay of trisodrum ethylenedramrnetetraacetate trrhydrate (EDTA) for possible carcrnogenicity Washington, DC, US Department of Health, Education and Welfare, National Cancer Institute (Carcinogenesis Technical Report Series No. !I). Oser BL, Oser M, Spencer HC (1963) Safety evaluation studies of calcium EDTA. Toxicology and applred pharmacology, 5:142-162. Perry HM, Perry EF (1959) Normal concentrations of some trace metals in human urine: changes produced by EDTA Journal of clinical investigations, 38:1452-1463. Plumb RC, Martell AE, Bersworth FC (1950) Spectrophotometric determination of displacement series of metal complexes of the sodium salts of ethylene diaminetetraacetic acid. Journal of physical collordal chemistry, 54:1208-1215. Reuber MD, Schmieller GC (1962) Edetate lesions in rats. Archives ofenvironmental health, 5:430-436. Schardein JL et al. (1981) Teratogenesis studies with EDTA and its salts in rats. Toxicology and applred pharmacology, 61 :423-428 Shibata S (1956) Folio Pharmacologica Japonica, 52:113 (in Japanese). Smith BL, ed. (1990) Codex Alimentarius, abridged version. Rome, FAO/WHO, Codex Alimentarius Commission. Srbrova J, Teisinger (1957) Ober die Resorption des Calciumnatriumsalzes der Athylendiamintetraessigsaure bei der peroralen Verbreichung zur Therapie der Bleivergiftung. [On the resorption of the calcium disodium salt of EDTA after oral administration for treatment of lead poisoning.] Archiv fur Gewerbepathologie, 15:572. Svenson A, Lennart K, Bjorndal H ( 1989) Aqueous photolysis of the iron(III) complexes of NTA, EDTA and DPTA. Chemosphere, 18(9/10):1805-1808. Swenerton H, Hurley LS ( 1971) Teratogenic effects of a chelating agent and their prevention by zinc. Scwnce, 173:62--64. van Dijk-Looyard AM et al. (1990) EDTA m drrnk-en oppervlakwater. [EDTA in drrnking and surface water.] Bilthoven, Rijkinstitut voor Volksgezondheid en Milieuhygiene (National Institute of Public Health and Environmental Protection) (Report No. 718629006). West TS, Sykes AS (I 960) Diamino-ethane-tetra-acetic acid. In: Ana(vtical applrcations of diamino ethane-tetra-acetic aCid. Poole, The Bntish Drug Houses, Ltd., pp. 9-22. Westerfield WW (1961) Effects of metal binding agents on metalloproteins. FederatiOn proceedmgs, 20 (Suppl. I 0): 158. WHO (1974) Toxicologzcal evaluatiOn of certarn food addiflves including antrcaking agents, aniimicrobials, antiOxidants, emulsifiers and thickening agents. Geneva, Joint FAO/WHO Expert Committee on Food Additives (WHO Food Additives Series No. 5). WHO (1993) Toxicological evaluatwn of certarn food additives and contaminants Prepared by the 41st Meeting of the Joint FAO/WHO Expert Committee on Food Additives. Geneva, World Health Organization, International Programme on Chemical Safety (WHO Food Additives Series No. 32). Yang S-S (I 952) Toxicologzcal investigation of ethylenedzammetetraacetic aczd m the rat. Ph.D. thesis dated May 1952, submitted to the University of Massachusetts, Amherst, MA [summarized in Anonymous, 1964]. POLYNUCLEAR AROMATIC HYDROCARBONS
First draft prepared by J. Kielhorn and A. Boehncke Fraunhofer Institute for Toxicology and Aerosol Research Hannover, Germany
The 1993 WHO Guidelines for drinking-water quality recommended a guideline value for benzo[a]pyrene (BaP) in drinking-water of 0.7 J.l.g/litre, corresponding to an excess lifetime cancer risk of 1o-s. This guideline value was based on an increased incidence of forestomach tumours in mice fed BaP in the diet, quantified using the two-stage birth-death mutation model. There were insufficient data available to allow the derivation of drinking-water guideline values for other polynuclear aromatic hydrocarbons (P AHs ). The 1992 Final Task Group Meeting recommended that P AHs other than BaP be evaluated in the future, and a draft IPCS Environmental Health Criteria monograph on PAHs was finalized in 1995. The Coordinating Committee for the updating of the WHO Guidelines therefore recommended that a group of representative P AHs be selected for evaluation in the 1998 Addendum.
I. GENERAL DESCRIPTION
P AHs are a class of diverse organic compounds containing two or more fused aromatic rings of carbon and hydrogen atoms. They are ubiquitous pollutants formed from the combustion of fossil fuels and are always found as a mixture of individual compounds. Owing to their low solubility and high affinity for particulate matter, P AHs are not usually found in water in notable concentrations. Their presence in surface water or ground water is an indication of a source of pollution. P AHs are only slowly biodegradable under aerobic conditions and are stable to hydrolysis. The relative concentrations of PAHs in air, water, and food are usually the same, although this can change depending on certain sources of pollution. In drinking water, the P AHs detected in the highest concentrations are fluoranthene (FA), phenanthrene, pyrene (PY), and anthracene. Of the six PAHs usually measured in water for regulatory purposes, FA is the only P AH that is detected to any significant extent. Some P AHs are known carcinogens, but many of these have not been measured in drinking-water, have not been detected in drinking-water (e.g. dibenzo[ a,e ]pyrene, dibenzo[a,h ]pyrene, chrysene ), or have been found in relatively low concentrations in drinking-water. BaP, a carcinogen, is the most extensively studied P AH (WHO, 1996). Chlorinated P AHs may be formed where residual chlorine is present in drinking-water (Shiraishi et al., 1985). Little is known about the toxicity of these compounds compared with that of their parent compounds (Bhatia et al., 1987). 124 ORGANIC CONSTITUENTS
The P AHs discussed in this document were chosen on consideration of their solubility, their presence in surface water and drinking-water from water survey data, and their classification as known or suspected carcinogens.
1.1 Identity
Compound CAS no. Molecular No. of Abbreviation formula aromatic used rings Fluoranthene 206-44-0 CI6Hio 4 FA Pyrene 129-00-0 CI6HIO 4 py Benz[a ]anthracene 56-55-3 clsHI2 4 BaA Benzo(b ]fluoranthene 205-99-2 C2oH12 5 BbFA BenzoU]fluoranthene 205-82-3 C20H12 5 BjFA Benzo[k]fluoranthene 207-08-9 c2aHI2 5 BkFA Benzo[a] pyrene 50-32-8 c2aHI2 5 BaP Dibenz(a,h ]anthracene 53-70-3 C22H14 5 DBahA 191-24-2 Benzo[ghi]perylene C12H 12 6 BghiP Indeno [I ,2,3 -cd] pyrene 193-39-5 C22H12 6 IP
1.2 Physicochemical properties (WHO, 1997)
At ambient temperatures, P AHs are colourless to yellow solids. The general characteristics common to the class are high melting and boiling points, low vapour pressures, and low water solubilities; the latter tend to decrease with increasing molecular mass. PAHs are highly lipophilic.
PAH Melting Boding Vapour Water Henry's law n-Octanol- Orgamc pomt pomt pressure solubzlity constant at water carbon (°C) (oc; at 25°C at 25°C 25°C partition normalized (kPa) (fig/litre) (kPa m3! coefficzent sorptzon
mol) (log Ka 11) coefficient
(log K0 c)c FA 108.8 375 1.2x I o-6 260 6.5x I 0-4 5.22 4.0--6.4 py 150.4 393 6.0xlo-' 135 l.lxi0-3 5.18 4.0--6.5 BaA 160.7 400 2 8xlo-' 14 n.g' 5.61 4.5-7.3 BbFA 168.3 481 6 7xlo-s 1.2b 5 lx!0-5 6.12 n.g. BjFA 165 4 480 2.0x I o-'• 2.5b 4.4x I o-' 6.12' n.g. BkFA 215.7 480 1.3x!O-'' 0 76 n.g 6.84' 4.0-7.0 BaP 178.1 496 7.3x!0-'0 3.8 3.4xlo-s 6.50 4.0-8.3 DBahA 266 6 524 1.3xlo-" 0.5(27°C) 7x I o-6• 6.50 n.g. BghiP 278.3' 545 1.4x I o-" 0.26 2.7xlo-s 7.10 5.6-6.1 lP 163 6 536 1.3xlo-" 62b 2.9x I o-s 6.58 ng
Calculated. Temperature not g1ven. Range of measured data for sediments and soils. d Not given POLYNUCLEAR AROMATIC HYDROCARBONS 125
1.3 Major uses
Only a small number of PAHs are produced commercially, including FA and PY, which are used mainly as intermediates in the production of fluorescent dyes (FA) and perinon pigments (PY) (Franck & Stadelhofer, 1987; Griesbaum et al., 1989). In 1993, one of the greatest producers worldwide manufactured <50 t ofF A and <500 t of PY (WHO, 1997). Coal and crude oils contain PAHs in considerable concentrations owing to diagenetic formation in fossil fuels (IARC, 1985, 1989). As a consequence, the compounds are also found in coal and mineral oil products such as coke, bitumen, coal tar (and creosote), heating oils, vehicle fuels, lubricating and cutting oils, and printing colour oils (Grimmer et al., 1981; IARC, 1984; Tetzen, 1989; Menichini et al., 1990). PAHs in the environment are almost always derived from anthropogenic activities. The largest amount of P AHs enters the environment via the atmosphere from incomplete combustion processes, such as processing of coal and crude oil (e.g. refining, coal gasification, coking), industrial use of coal and mineral oil products (aluminium production, iron and steel production, foundries), heating (power plants and residential heating using wood, coal, and mineral oil), fires (e.g. forest, straw, agriculture, cooking), incineration of refuse, vehicle traffic, tobacco smoking, and volcanic activities (for quantitative data on the release of P AHs in the environment, see WHO, 1997).
1.4 Environmentalfate
P AHs are emitted mainly into the atmosphere and have been detected long distances from their source (Bjorseth & Sortland, 1983; McVeety & Hites, 1988). Because of their low vapour pressures, compounds with five or more aromatic rings will exist mainly adsorbed to airborne particulate matter, such as fly ash and soot. Those with four or fewer rings will occur both in the vapour phase and adsorbed to particles (Hoff & Chan, 1987; Baker & Eisenreich, 1990). P AHs reach the hydrosphere mainly by dry and wet deposition and road runoff but additionally from industrial wastes containing PAHs and leaching from creosote-impregnated wood. P AHs are adsorbed strongly to the organic fraction of sediments and soils. Leaching of P AHs from the soil surface layer to ground water is assumed to be negligible owing to the adsorption and to biodegradation in the aerobic soil surface layer, although their presence in groundwater has been reported, mainly at contaminated sites. The volatility of the compounds from water phases is low, with half-lives of 500 and 1550 hours for BaA and BaP, respectively (Southworth, 1979). The compounds are very slowly biodegradable under aerobic conditions in the aqueous compartment. The biodegradation rates decrease drastically with increasing number of aromatic rings. In laboratory experiments with soil samples, the calculated half-lives for the selected compounds vary widely, from about I 00 days to a couple of years (Bossert & Bartha, 1986; Coover & Sims, 1987; Park et al., 1990; Wild et al., 1991 ). P AHs are stable towards hydrolysis. 126 ORGANIC CONSTITUENTS
The most important degradation process for P AHs in air and water is indirect photolysis under the influence of hydroxyl radicals. Under laboratory conditions, the reaction of the compounds with airborne hydroxyl radicals shows maximum half lives between about 3 and 11 hours (Atkinson, 1987). For pure water, the photodegradation half-lives appear to be in the range of hours (Mill et al., 1981; Mill & Mabey, 1985), whereas the half-lives increase drastically when sediment/water partitioning is taken into account (Zepp & Schlotzhauer, 1979). Measured bioconcentration factors (BCFs) for the compounds in the aquatic environment vary widely owing to different measurement techniques and are especially high for some algae (BCF = 2398-55 800), crustaceans (BCF = 180-63 000), and molluscs (BCF = 58-8297). Bioconcentration factors in fish appear to be much lower than in these organisms because of rapid biotransformation processes (BCF = 10-4700) (WHO, 1997). In summary, it can be concluded that sediments and soils are the main sinks for P AHs in the environment and that P AHs with four or more aromatic rings are persistent in the environment (Mackay et al., 1992).
2. ANALYTICAL METHODS
A preconcentration step for sample enrichment may be necessary for the analysis of P AH levels in uncontaminated aqueous samples. Further, considerable adsorption losses during collection and storage of samples have to be taken into account. Apart from liquid/liquid extraction procedures (e.g. with dichloromethane), various solids have been successfully used for the preconcentration: Tenax-GC, XAD resins, open-pore polyurethane foam (PUF). and bonded-phase silica gel (van Noort & Wondergem, 1985; Basu et al., 1987). Detection is carried out by gas chromatography with a flame ionization or a mass selective detector and by high performance liquid chromatography with an ultraviolet or a fluorescence detector. The detection limits are between 0.01 and 200 ng/litre (Basu & Saxena, 1978a; Desideri et al., 1984).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
Numerous studies refer to the occurrence of P AHs in the environment. The compounds were detected in all environmental compartments. The concentrations differ widely owing to the widely differing sampling locations and conditions. In the matrices air, water, and food, the relative concentrations of the selected P AHs appear to be in the order ofF A =:: PY > BghiP =:: IP > BbF A =:: BkF A > BaA =:: BaP> DBahA. Very few data were available for BjFA; its concentration is probably in about the same range as those ofBbFA and BkFA.
3.1 Air
In the absence of industrial or other point sources of pollution, P AHs in the atmosphere are mainly from residential heating and vehicle traffic. The levels of individual substances vary over several orders of magnitude and are generally in the POLYNUCLEAR AROMATIC HYDROCARBONS 127 range between <0.1 and 100 ng/m3 (WHO, 1997). PAHs are mainly adsorbed to airborne particulate matter.
3.2 Water
Apart from highly industrially polluted rivers, the concentrations of individual P AHs in surface and coastal waters are generally :S:50 ng/litre (WHO, 1997). Concentrations above this level (sometimes into the 10 J-lg/litre range) indicate contamination by P AHs mainly through industrial point sources and shipyards, atmospheric deposition, and urban runoff. Ships for inland navigation are periodically treated with coal tar to prevent corrosive damage. The leaching/abrasion of this coating is a source ofPAHs (Berbee, 1992). In addition, wood preserved with creosote can leach PAHs into the environment, especially into waters where wood is used for bank protection or harbours and in the disposal of creosote-impregnated railway ties (Berbee, 1992; Sandell & Tuominen, 1996). P AH levels in uncontaminated ground water are usually in the range of 0-5 ng/litre. Leaching of P AHs from soils into groundwater is negligible, as the compounds tend to adsorb strongly to the soil organic matter (Woidich et al., 1976; Stuermer et al., 1982). Only at heavily contaminated sites do the P AHs reach the groundwater, giving concentrations above 10 J-lg/litre (Environment Canada, 1994). Elevated concentrations of PAHs (predominantly FA, BbFA, PY, lP, phenanthrene) were observed in rainwater and especially in snow and fog (WHO, 1997). This is probably a result of the adsorption of the compounds to air particulate matter, which is finely dispersed into the water during wet deposition. The typical concentration range for the sum of the selected PAHs in drinking-water is from about 1 ng/litre to worst cases of 11 J-lg/litre (see Table 1). Many individual P AHs are at concentrations below the detection limit. As an example, in 1988-1989, the sum of the six Borneff PAHs was below the detection limit of 5 ng/litre in 88% (5287 of 5975) of drinking-water samples from waterworks in Germany; the concentrations were below 40 ng/litre in 10% (588 samples); and concentrations above 200 ng/litre were detected in 0.08% (5 samples) (Dieter, 1994). The main source of PAH contamination in drinking-water is usually not the raw water sources but the coating of the drinking-water distribution pipes. At least in the past, coal tar was a common coating material for water pipes, used to give effective protection against corrosion. After the passage of drinking-water through those pipes or after repair work, significantly increased PAH levels have been detected in the water (Vu Due & Huynh, 1981; Basu et al., 1987; Davi et al., 1994); for example, a concentration of2.7 J-lg ofBorneffPAHs per litre was detected in one sample of such water (State Chemical Analysis Institute, 1995). Although WHO has called for a cessation of this practice (WHO, 1996), many countries still have a large amount of pipes lined with coal tar coating. If BaP is present at elevated concentrations in drinking-water, this is indicative of the presence of particulate matter (e.g. from the deterioration of the coal tar coating). Table I. Concentrations of the selected PAHs in drinking-water (ngllitre)
Location, year [reference] Source of water FA• py BaA BbFA• BjFA BkFA• BaP• DBahA BghiP• lP• Austria, 1976 [I] Spnng water and well-water 3.5-6 5 I 6-3 5 n.d-1 9 0 2-0 8 0 2-0.8 0.1-0 7 0 3-0 9 tr-0 7 USA, 1976-1977 [2,3] Treated water from polluted 2 4-24 0 2-1 2 0.1-0.7 0.2-1 6 0 4-4 0 0 7-2.2 source Norway, 1978-1981 [4.5] Tap-water 0 58-24 <0 3-15 0 1-5.5 0.05-4 0 0 03-0 14 0.02-0.10 <0 04-2.0 L2 0.4-1 I 0 4-1.2 Canada. 1987-1990 [6] Treatment plant water <5-623 40 <5-40 <5-40 <5 <5 <5 Poland. 1984 [7] Spa water 4-21 4-29 3-48 4-21 9-51 9-57 Switzerland, 1981 [8] Reservoir' 150-3400 9-14 1-5 tr-2 nd tr Tap-water 3 3 0 4-0 6 0 1-0.9 01-1 n d. tr Italy. 1991-1993 [9] Treatment plant <20 <10 <10 <20 <10 <10 <20 <20 <20 Fountam. new coal tar limng <20 max. 30 max 20 <20 <10 <10 <20 <20 <20 England & Wales, 1996 Tap-"ater (hard water) 585 20 6 14 12 8 [10] Tap water (soft water) 6520 1600 490 914 432 953
n d = not detected, tr = traces
The authors attnbute these high levels to the use of coal tar distributiOn pipes PAHs measured for regulation purposes m the EEC (BorneffPAHs)
References [I] WOJdich et al. (1976) [6] Environment Canada (1994) [2] Thruston ( 1978) [7] Babelek & Clezkowski ( 1989) [3] Basu & Saxena (1978b) [8] Vu Due & Huynh (1981) [4] Kveseth et a! (1982) [9] Davi et al. (1994) [5] Berghnd (1982) [10] Dnnkmg Water Inspectorate. personal commumcatwn (1997) POLYNUCLEAR AROMATIC HYDROCARBONS 129
In Canada, significantly increased levels of P AHs in drinking-water were reported for which the reason is not known (Environment Canada, 1994 ). Also, the P AH concentrations in spa waters from 10 different spas in the Sudetes region (Poland) are surprisingly high (Babelek & Ciezkowski, 1989). In most of the PAR-contaminated spas. groundwater, presumably polluted, also contributes to the spa water. In the majority of drinking-water samples taken in England and Wales, PAHs are not detected above the standard (EEC. 1980; CEC. 1995) for P AHs of 0.2 f.!g/litre. Only 5% of the reported samples fail to meet the standard. In practically every case where the P AH standard has been exceeded, the only P AH detected to any significant extent is FA. This is indicative of a coal tar pitch lining in good condition where the hard groundwater very slowly dissolves the lining. There are very few cases where other P AHs have been detected in significant concentrations, and these occur mainly where soft corrosive water is derived from surface water sources. This is probably indicative of physical deterioration of the lining, releasing particulate containing PAHs into the water supply (Drinking Water Inspectorate, personal communication, 1997).
3.3 Food
P AHs have been detected in fresh vegetables, fruits, and cereals as a result of the deposition of airborne P AHs, particularly near industrial sources or in areas with high traffic (Tuominen et al., 1988; de Vos et al., 1990; Dennis et al.. 1991). PAHs have also been found in mussels, snails, and fish from contaminated waters (Sirota & Uthe, 1981; Rostad & Pereira, 1987; Speer et al., 1990). PAHs are also present at elevated levels in some vegetable oils and margarine (Dennis et al., 1991; Thomson et al., 1996), probably formed during processing. PAHs are also formed during some methods of food preparation, such as char-broiling, grilling, roasting, frying, or baking. The highest levels were detected in smoked and grilled meat and fish samples (up to about 200 f.!g/kg) (WHO, 1997).
3.4 Estimated total exposure and relative contribution of drinking-water
For the general population, the major routes of exposure to PAHs are from inhalation via ambient and indoor air and ingestion via food. For ambient air, residential heating and vehicle traffic appear to be the main sources of exposure. In the direct vicinity of an emission source, a maximum intake of 1 f..lg of BaP per day may be reached (WHO, 1987; LAI, 1992). For the other selected compounds, maximum intakes of between 0.004 (DBahA) and 0.06 (BbF A) f.!g/day were estimated (Chen et al., 1980; Guicherit & Schulting, 1985). For indoor air, an important contribution is from smoking. In this case, the BaP intake may almost reach that for polluted ambient air. Especially in developing countries, the use of open fires for heating and cooking may further increase P AH exposure (Mumford et al., 1987; Raiyani et al., 1993). The main contributors of P AHs to the total dietary intake appear to be cereals, oils, and fats. The oil and fat group has high individual P AH levels, whereas 130 ORGANIC CONSTITUENTS the cereal group, although never containing high individual P AH concentrations, is a main contributor by weight to total intake in the diet. Smoked meat and fish products, although containing the highest P AH levels, appear to be low to modest contributors, as they are minor components of the usual diet (Larsson, 1982, 1986; Dennis et al., 1983, 1991; Maga, 1986). However, it should be noted that various countries and cultures have very different diets and methods of cooking, which may result in exposure to very different amounts ofPAHs. There are a few studies on daily intake of individual P AHs from food from western Europe (Dennis et al., 1983; Vaessen et al., 1984; de Vos et al., 1990; Pfannhauser, 1991; Turrio-Baldassarri et al., 1996) and Canada (WHO, 1996). The results for the individual PAHs were in the same range. BaP, BghiP, PY, and FA can each reach a maximum daily intake of :S:lO f.!g per person; for each of DBahA lP, BkF A, and BaA, the maximum daily intake is :0::0.5 11g per person. The maximum/median intake levels for the P AHs selected in this guideline, in f.!g/day per person, have been estimated to be as follows: FA (4.3/0.6); PY (4.0/0.6); BaA (0.14/0.02); BbF A ( 1.0/0.005); BjF A (0. 9/0.03); BkF A (0.3/0.04 ); BaP (0.36/0.05); DBahA (0.1 0/0.015); BghiP (7.6/0.12); and lP (0.31/0.025) (Pfannhauser, 1991 ). From the intake data for food and the drinking-water levels (see Table 1), it can be estimated that about 1% of the total dietary intake of P AHs is from drinking water, assuming a consumption of2 litres/day. Where there are elevated PAH levels from contamination by coal tar coatings, which would occur mainly during and after repair work, P AH intake from drinking-water could be equal to or even exceed other dietary intakes. Exposure via the oral and inhalation pathways varies considerably depending on diet and lifestyle, with inhalation exposure being of greater importance where indoor levels of P AHs are high because of smoking (Greenberg, 1996; lhme & Wichmann, 1996; Jansen et al., 1996).
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
4.1 Absorption
P AHs are absorbed in experimental animals and humans through the pulmonary tract, the gastrointestinal tract, and the skin. Absorption of BaP, DBahA, and PY was high (30-90%) following low and high oral doses in rats (Chang, 1943; Foth et al., 1988; Withey et al., 1991). Absorption from the gastrointestinal tract occurs rapidly. Oral administration of FA, PY, and BaA to rats caused peak concentrations of these compounds in the blood after 1-2 hours (Lipniak & Brandys, 1993). The intestinal absorption of the individual PAHs is highly dependent on their solubility, their lipidity, the presence of bile (Rahman et al., 1986), and the lipidity of the various PAH-containing foods ingested. Whereas oils enhanced the absorption of BaP, water and solid food suppressed the absorption (Kawamura et al., 1988). POLYNUCLEAR AROMATIC HYDROCARBONS 131
4.2 Distribution
In laboratory animals, P AHs become widely distributed in the body following administration by any one of a variety of routes and are found in almost all internal organs, particularly those rich in lipid (WHO, 1997). Maximum concentrations of BaA in perfused tissues (e.g. liver, blood, brain) were achieved within 1-2 hours after administration of high oral doses (76 and 152 mg/kg of body weight). In lesser perfused tissues (e.g. adipose and mammary tissue), maximum levels of this compound were reached in 3--4 hours (Bartosek et al., 1984). In male Wistar rats receiving a gavage dose of 2-15 mg of C4C]-pyrene per kg of body weight, the fat had the highest levels of radioactivity, followed by the kidney, liver, and lungs (Withey et al., 1991). Orally absorbed DBahA in rats was also widely distributed to several tissues. After continuous oral administration of 0.5 f!g of 3 [ H]BaP daily to male rats for up to 7 days, the radioactivity persisted in liver, kidney, lung, and testis (Yamazaki & Kakiuchi, 1989). Orally administered BaP (200 mg/kg of body weight) has been shown to cross the placental barrier and has been detected in fetal tissues (2.77 f!g/g) (Shendrikova & Aleksandrov, 1974). Using 14C-tagged BaP, a BaP concentration 1-2 orders of magnitude lower in embryonic than in maternal tissues was determined after oral administration in mice (Neubert & Tapken, 1988). Differences in concentrations in the fetus among the various P AHs appeared to be highly dependent on the gastrointestinal absorption of the compound.
4.3 Metabolism
The metabolism of PAHs is complex. Generally, the process involves epoxidation of double bonds, a reaction catalysed by the cytochrome P-450- dependent monooxygenase, the rearrangement or hydration of such epoxides to yield phenols or diols, respectively, and the conjugation of the hydroxylated derivatives. Reaction rates vary widely, and interindividual variations of up to 7 5-fold have been observed, for example, with human macrophages, mammary epithelial cells, and bronchial explants from different donors. Most metabolism results in detoxification, but some PAHs in some situations become activated to DNA-binding species, principally diol-epoxides, that can initiate tumours (WHO, 1997). Although the P AHs are similar, they have structural differences that are the basis for differences in metabolism and relative carcinogenicity. The metabolism of the more carcinogenic, alternant (equally distributed electron density) P AHs, such as BaP, BaA, and DBahA, seems to differ in some ways from that of non-alternant (uneven electron density distribution) PAHs, such as FA, BbFA, BkFA, BjFA, IP, BghiP, and PY (Phillips & Grover, 1994; ATSDR, 1995). In general, little is known about the metabolism of most P AHs, particularly in non-rodent species. It should be noted that there appear to be species differences in the enzymes that activate P AHs (Michel et al., 1995) and in the formation of DNA adducts (Kulkarni et al., 1986). 132 ORGANIC CONSTITUENTS
4.4 Excretion
P AH metabolites and their conjugates are excreted predominantly via the faeces and to a lesser extent in the urine. Conjugates excreted in the bile can be hydrolysed by enzymes of the gut flora and reabsorbed. It can be inferred from available data on total body burdens in humans that PAHs do not persist for long periods in the body and that turnover is rapid. This excludes those P AH moieties that become covalently bound to tissue constituents, in particular to nucleic acids, and are not removed by repair (WHO, 1997). The excretion of urinary metabolites is a method used to assess internal human exposure ofPAHs.
5. EFFECTS ON LABORATORY ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Toxicological effects of individual PAHs
The toxicological effects of the PAHs are summarized individually in order of molecular weight, with emphasis on oral studies where available. The toxicology of FA is described in the most detail because this is the P AH found in notable quantities in tap-water where there is contamination by· coal tar coatings and because of the uncertain classification of this P AH. Research on the toxicological effects of P AHs has been focused on the carcinogenicity of some selected compounds, but usually employing dermaL inhalation, or subcutaneous rather than oral exposure. The carcinogenic classification of the various P AHs is given in Table 2. There are only limited studies on non carcinogenic end-points.
5.1.1 Fluoranthene (FA)
5.1.1.1 Acute and short-term exposure
The oral LD 50 for FA in the rat is about 2000 mg/kg of body weight (range 1270-3130 mg/kg ofbody weight) (Smyth et al., 1962). Male and female CD-1 mice (20 per sex per group; 30 per sex for controls) were given FA by gavage for 13 weeks at 0, 125, 250, or 500 mg/kg of body weight per day and then sacrificed and autopsied (US EPA, 1988). All treated mice exhibited nephropathy, increased salivation, and increased liver enzyme levels in a dose-dependent manner. Mice given 500 mg/kg of body weight per day had increased food consumption and increased body weight. At doses of 250 and 500 mg/kg of body weight per day, statistically increased serum glutamate-pyruvate transaminase (SGPT) levels and increased absolute and relative liver weights were noted, as well as compound-related microscopic liver lesions (indicated by pigmentation) in 65 and 87.5% of the mice, respectively. Based on these increased POLYNUCLEAR AROMATIC HYDROCARBONS 133
Table 2. Evaluation of individual P AHs for carcinogenicity in animals and humans
PAH WHO, 1997• IARC, 1987b BaA positive 2A BbFA positive 28 BjFA positive 28 BkFA positive 28 BghiP negative 3 BaP positive 2A DBahA positive FA (positive)' 3 IP positive py (questionable)
Based on animal carcinogemcity studies only b 2A - probably carcinogenic to humans; 28 - possibly carcinogenic to humans; 3 -not classifiable as to human carcinogenicity. Recent data on FA since this meetmg could change the FA rating to questionable
SGPT levels, kidney and liver pathology, and clinical and haematological changes, the NOAEL is 125 mg/kg of body weight per day. 1
5.1.1.2 Mutagenicity and related end-points
In vitro genotoxicity studies for FA are mostly positive, but in vivo genotoxicity studies are mostly negative (IARC, 1983; US EP A 1992; ATSDR, 1995; WHO. 1997). FA tested positive with metabolic activation for gene mutation in Salmonella typhimurium, in the Escherichia coli SOS chromotest for DNA damage, and in in vitro tests in mammalian cells for DNA damage, mutation, and chromosomal effects. After oral administration ofFA at 750 mglkg of body weight, in vivo tests for sister chromatid exchange in mouse bone marrow were negative (Palitti et al., 1986). FA did not show any evidence of genotoxicity in the mouse bone marrow micronucleus or rat liver unscheduled DNA synthesis test systems following acute oral administration at levels of up to 2000 mg/kg of body weight ( Stocker et al.. 1996). A major FA-DNA adduct has been identified in most of the tissues examined (including liver, lung, heart, kidney, spleen, and thymus) in Sprague-Dawley rats chronically fed FA in the diet (Gorelick et al.. 1989). In BLU:Ha and CD-I mice
Source: Integrated Risk Information System (IRIS) Onhne. Cincinnati, OH, US Environmental Protection Agency, Environmental Cntena and Assessment Office, Office of Health and Environmental Assessment 134 ORGANIC CONSTITUENTS treated intraperitoneally with tumorigenic doses of FA (total of 3.5 mg over 2 weeks), highest levels of FA-DNA adduct were found in the lung (Wang et al., 1995a,b).
5.1.1.3 Dermal carcinogenicity studies
Dermal application of I% FA 3 times a week for I year to the backs of 20 female Swiss-Albino Ha/ICD/Mill mice did not induce skin tumours (Hoffmann et al., 1972), nor did 250 Jlg of FA applied to 15 male C3H mice for 82 weeks (Horton & Christian, 1974). Application of 40 Jlg of FA alone caused no tumours in 50 female Swiss mice treated for 440 days, but FA was a eo-carcinogen in a study in which the same dose of FA in combination with BaP induced a 2-fold increase in mouse skin tumours compared with BaP alone (van Duuren & Goldschmidt, 1976). FA did not exhibit tumour-initiating activity after 24 weeks in 30 female Swiss mice topically administered I 0 doses (0.1 mg per animal) followed by promotion with croton oil for 20 weeks (Hoffmann et al., 1972).
5.1.2 Pyrene (PY)
Male and female CD-I mice (20 per sex per group) given PY by gavage at doses of 0, 75, 125. or 250 mg/kg of body weight per day in corn oil for 13 weeks exhibited kidney effects (renal tubular pathology. decreased kidney weights) (US EPA, 1989). The low dose (75 mg/kg of body weight per day) was considered the NOAEL for nephropathy and decreased kidney weights. 2 Mutagenicity and related studies gave negative or equivocal results. There were no oral carcinogenicity studies. Skin painting assays in mice for complete carcinogenesis or initiating capacity have been negative or inconclusive. Mice injected intraperitoneally did not show significant elevated tumour rates. A 32P-postlabelling test for covalent DNA binding of PY to mouse skin in vivo gave negative results. PY was eo-carcinogenic with BaP in a mouse skin assay (WHO, 1997).
5.1.3 Benz[ajanthracene (BaA)
Male B6AFI!J newborn mice (40 per group) were administered 1.5 mg of BaA per day by oral gavage twice over 3 days (Klein, 1963). After 568 days of observation, increased incidences of hepatomas and pulmonary adenomas (80% and 85%, respectively) were noted, compared with the controls with solvent only (10% and 30%). No malignant tumours were observed. In a parallel study with the same dose 3 times weekly over 5 weeks with sacrifice at 547-600 days, 100% hepatomas and 95% pulmonary adenomas were noted (controls: 10% and 35%).
Source· Integrated Risk InformatiOn System (IRIS) Online. Cincinnati, OH, US Environmental Protection Agency, Environmental Criteria and Assessment Office, Office of Health and Environmental Assessment POLYNUCLEAR AROMATIC HYDROCARBONS 135
C57BL mice receiving a total dose of 0.5, 4, or 8 mg of BaA by gavage showed forestomach papillomas (0/13, 1110, and 1/8, respectively) after 16 months (Bock & King, 1959). BaA is genotoxic. It produces tumours in most assays in mice treated dermally, intraperitoneally, and subcutaneously. There are indications of immunotoxicity and fetotoxicity in subcutaneous studies. DNA adducts were detected in mouse skin after dermal application of BaA (US EP A, 1992; ATSDR, 1995; WHO, 1997).
5.1.4 Benzo[bffluoranthene (BbFA)
BbF A is genotoxic. Exposure of rats to BbF A by lung implantation resulted in tumour formation, as did intraperitoneal exposure of newborn mice. Skin painting and initiation/promotion studies in mice were positive. DNA adducts were detected in vitro and in vivo (US EPA, 1992; ATSDR, 1995; WHO, 1997).
5.1.5 BenzoOffluoranthene (BjFA)
From limited studies, there is evidence that BjF A is genotoxic and carcinogenic. BjFA showed tumorigenic activity in one skin painting assay, in initiation/promotion studies, and in the newborn mouse intraperitoneal bioassay. Exposure of rats by lung implantation did not result in tumour formation. DNA adducts were detected in vitro and in mouse skin in vivo after topical application of BjFA (US EPA, 1992; ATSDR, 1995; WHO, 1997).
5.1.6 Benzo[kffluoranthene (BkFA)
From the available evidence, BkF A is genotoxic and carcinogenic. Skin painting assays were not positive, but initiation/promotion studies resulted in increased tumour incidence. No significant tumorigenic activity was found in a lung adenoma bioassay in newborn mice. Lung implantation of BkF A produced tumours in rats. DNA adducts have been detected in vitro and in vivo (US EP A, 1992; ATSDR, 1995; WHO, 1997).
5.1. 7 Benzo[ajpyrene (BaP)
BaP is genotoxic in a variety of in vitro tests with metabolic activation and in in vivo studies. In mice, oral administration of BaP induces tumours of the forestomach. It induces mammary gland tumours after oral administration in rats. BaP has produced skin tumours after dermal application in mice, rats, rabbits, and guinea pigs. It produced lung and respiratory tumours when administered intratracheally to rats and hamsters. Lung implantation of BaP in rats caused pulmonary tumours. BaP administered intraperitoneally induced lung and hepatic tumours in mice. It was carcinogenic after subcutaneous administration to mice, rats, hamsters, guinea-pigs, and some primates. BaP binds to DNA and forms DNA adducts in target organs. 136 ORGANIC CONSTITUENTS
BaP was discussed in the 1993 WHO Guidelines for drinking-water quality (WHO, 1996). Only the oral studies on BaP and coal tar published since then are given here. It should be noted that further long-term oral studies are in progress. The tumorigenic activity of BaP (and coal tar; see section 5.3) after ingestion was investigated (Weyand et al., 1995). BaP (16 or 98 mg/kg of feed, equal to 41 and 257 JJ.g/day, respectively) in a basal gel diet was fed to female A/J mice (30 per group) for 260 days. After sacrifice, forestomach tumours and pulmonary adenomas were diagnosed and counted. The incidence of forestomach tumours after oral administration was 20% and 100%, respectively; tumour multiplicity was 0.24 and 4.22. The incidence of forestomach carcinomas in mice with forestomach tumours was 8% and 52%, respectively. Controls ingesting basal gel diet only showed no forestomach tumours. Whereas 16 mg/kg of feed did not induce a significant level of lung tumours (36%; 25 mice; 0.48 tumours per mouse), ingestion of 98 mg/kg of feed induced lung tumours in 52% of the mice (27 mice; 0.59 tumours per mouse). Nineteen per cent of the controls on basal diet showed pulmonary adenomas (21 mice; 0.19 tumours per mouse). It should be noted that A/1 mice are susceptible to pulmonary tumours.
In a 2-year carcinogenicity bioassay, female B6C3F1 mice were fed BaP at 0, 5, 25, or 100 mg/kg of feed (see also study with coal tar, section 5.3) (Culp et al., 1996). Forestomach tumours were induced in all groups of mice fed BaP. A 6% incidence was observed at the 5 mg/kg of feed dose (18.5 JJ.g/day), with the incidence increasing sharply to 78% at the 25 mg/kg of feed dose (90 J.lg/day) and then to 98% at 100 mglkg of feed (350 JJ.g/day). All of the mice fed BaP at 100 mg/kg of feed were removed by 80 weeks because of morbidity or death. A linear dose-response was observed between BaP dose and adduct levels in the forestomach of mice fed the same doses in a 4-week study.
5.1.8 Dihenz[a,fljanthracene (DBahA)
There are only limited studies on oral exposure to DBahA but they do provide some evidence for carcinogenicity through this route. In a lifetime study, mammary carcinomas were observed in 1120 female and 13/24 pseudo-pregnant BALB/c mice dosed with 0.5% DBahA after 15 weeks of dosing (Biancifiori & Caschera, 1962). No control group was included. DBA/2 mice (21 per sex) were given DBahA in an aqueous olive oil emulsion (0.8 mg of DBahA per day) for 200 days (Snell & Stewart, 1962). In general, the animals did not tolerate the vehicle well, and extensive dehydration and emaciation led to early death. At the end of the exposure, almost all surviving treated mice (14 males and 13 females), but only one of the control mice, showed pulmonary adenomatosis, alveologenic carcinoma, mammary carcinoma, and haemangioendotheliomas. DBahA is genotoxic in vitro and in vivo. It causes tumours in various organs in mice after oral administration and is a potent carcinogen in several species after various routes of administration. It forms DNA adducts in mouse skin in vitro and in vivo (US EPA, 1992; ATSDR, 1995; WHO, 1997). POLYNUCLEAR AROMATIC HYDROCARBONS 137
5.1.9 Benzo[ghijperylene (BghiP)
There are insufficient data on the genotoxic potential of BghiP, although the existing evidence is positive. BghiP tested negative for carcinogenicity activity and tumour-initiating activity in mouse skin. It was negative in the rat lung implantation assay. It has shown some co-carcinogenicity with BaP in mouse skin. BghiP binds to DNA in vitro and in vivo (US EPA, 1992; ATSDR, 1995; WHO, 1996, 1997).
5.1.10 Indeno[1,2,3-cdjpyrene (JP)
The limited data on the genotoxicity of IP are generally positive. IP has tumour-initiating activity in mouse skin and is carcinogenic in rat lungs. It bound to mouse skin with the formation of DNA adducts (US EPA, 1992; ATSDR, 1995; WHO, 1997).
5.2 Comparative studies
The following is a summary of the comparative studies on tumorigenic activity of individual P AHs, which have been used as the basis for comparative potency factors (see section 7). Full details and discussion of the adequacy of the databases are given in the original references and elsewhere (Clement Associates, 1988; US EP A, 1992). In general, it can be said that data from the skin painting and lung implantation studies have been used preferentially to those of initiation/promotion experiments and intraperitoneal studies for estimating comparative potencies. There are no comparative studies on oral administration. Exact comparative data are given only where this is possible (e.g. where single PAHs were tested in the same experiment at the same dose).
5.2.1 Carcinogen icily studies
5.2.1.1 Dermal
Skin-painting studies
Solutions of 0.5% BaP, BbFA, BjFA, or BkFA were applied dermally 3 times weekly to female Swiss Millerton mice (20 per group) through their lifetime, and the number of skin tumours was determined (Wynder & Hoffmann, 1959b ). The percentages of papillomas/carcinomas for these compounds after 4 months were 70/20, 95/10, 40/5, and 0/0, respectively. Minimal activity (10 papillomas) was found with BkF A after 11 months. BaP> BbFA > BjFA > BkFA. In a similar study regime, 0.01% solutions of BaP or DBahA applied dermally to mice (20 per group) showed 10%110% and 15%/5% papillomas/carcinomas after 6 months. A 0.1% solution of FA and 10% solution of PY showed no activity (Wynder & Hoffmann, 1959a). BaP= DBahA >> FA; PY. 138 ORGANIC CONSTITUENTS
A further study compared the carcinogenicity of BaP, BghiP, and lP applied dermally to mice 3 times weekly as above (Hoffmann & Wynder, 1966). A dose of 0.05% of BaP, BghiP, or IP produced 17/20, 0/20, and 0/20 tumour-bearing mice, showing that BaP is more potent than either of the other P AHs. There were no controls. BaP>> BghiP; lP. In a lifetime skin painting assay with female NMRI mice, BaP and BbFA were carcinogenic, BjF A was weakly carcinogenic, and BkF A and lP had no cancer inducing effects (Habs et al., 1980). BaP>> BbFA > BjFA > BkFA; lP.
Initiation/promotion assay
Ten doses of BaP, BghiP, or lP at a total dose of 0.25 mg per mouse were applied every second day to the backs of Swiss Millerton mice followed by promotion with 2.5% croton oil in acetone. Tumour-bearing animals were reported as 24/30, 2/27, and 5/30, respectively (Hoffmann & Wynder, 1966). BaP>> lP > BghiP. In an initiation/promotion assay in CD-I mice, four PAHs (BaP, BbFA, BjF A, and BkF A) were each applied at a total tlose of 30 f.!g in 10 subdoses over 20 days to the shaved backs of 20 mice per group (LaVoie et al., 1982). Ten days after completion of the initiation, promotion was begun by thrice-weekly application of 12-0-tetradecanoylphorbol-13-acetate in 0.1 m! of acetone. The skin tumours were predominantly squamous cell papillomas. After 20 weeks (10 weeks for BaP), the percentage of skin tumour-bearing animals was 85, 45, 30, and 5, respectively. The vehicle controls had no tumours. BaP> BbF A> BjF A> BkF A.
5.2.1.2 Other routes
Intraperitoneal injection in newborn mice
The tumorigenic activity of the non-altemant P AHs (BbF A, BjF A, BkF A, and lP) as well as BaP was evaluated by injecting intraperitoneally a total of 0.5, 1.1, 2.1, 2.1, or 0.5 !J.mol of each compound, respectively, in dimethyl sulfoxide in aliquots of 5, 10, or 20 f.!l on days 1, 8, and 15 of life, respectively, to CD-1 mice (LaVoie et al., 1987). A direct comparison was not possible owing to differences in the total amount injected; however, both BbFA and BjFA exhibited significant tumorigenic activity, whereas neither BkF A nor lP was tumorigenic under these conditions. There were problems with the solubility ofiP. BaP> BbFA = BjFA > BkFA; lP.
Lung implantation
Deutsch-Wenzel et al. (1983) and Wenzel-Hartung (1990) investigated the carcinogenic effects of P AHs after intrapulmonary injection and assessed the relative potencies with respect to epidermoid carcinomas and pleomorphic sarcomas. A rank order was based on BaP as reference substance: DBahA (1.91) BaP (1.00) - BbFA (0.11) - lP (0.08) - BkFA (0.03) - BjFA (0.03). BghiP showed no tumour-producing effects. POLYNUCLEAR AROMATIC HYDROCARBONS 139
Subcutaneous injection
Dose-response curves for BaP and DBahA were established following a single subcutaneous injection of the PAHs in tricaprylin into the right axilla of male C3H mice (Bryan & Shimkin, 1943). Ninety-nine per cent of the tumours detected were spindle-cell sarcomas. Vehicle control response levels were not included. Under the conditions in this experiment, the potency of DBahA was estimated to be 4.5 times that of BaP. DBahA >>BaP. In a study with C57 black mice, 8/10 males and 6/10 females had injection site tumours 60-80 weeks after 10 weekly subcutaneous injections of 1 mg of BaA (Boyland & Sims, 1967). After a dose of 1 mg of DBahA, 20/20 males and 17/20 females had tumours. DBahA > BaA.
5.2.1.3 Further evidence
Sebaceous gland assay
Application of carcinogenic P AHs to mouse skin leads to the destruction of sebaceous glands, hyperplasia, hyperkeratosis, and even ulceration (Bock, 1964). A sebaceous gland assay has been used as a screening method for the tumorigenic potential of PAHs. Acute topical application of BaP, BaA, or DBahA was reported to suppress sebaceous glands (Bock & Mund, 1958). BaP = DBahA > BaA. In a further sebaceous gland assay using other PAHs, it was found that, compared with BaP, the activity was BaP > BbF A = BjF A = BkF A= lP (Habs et al., 1980).
DNA adduct formation
In a 32P-postlabelling test for covalent DNA binding of P AHs to mouse skin in vivo following a single topical application, relative DNA adduct levels were BaP> BaA = DBahA = BghiP (Reddy et al., 1984). DNA adducts were not detected with PY. In a similar study, the relative covalent binding of PAHs to DNA was BbFA > BjFA > BkF A> lP (Weyand et al., 1987). In an in vitro study, the relative covalent binding of P AHs to DNA was reported as BaP > DBahA > BaA > PY (Grover & Sims, 1968).
5.2.2 Summary
The results of these carcinogenicity and other studies, although not always giving the identical order, can be summarized as follows: BaP = DBahP > BaA > BbFA > BjFA > BkFA > lP >FA> BghiP > PY. FA has been included in very few comparative studies, but the above placing is probably correct (see recent studies in section 5.1.1). A quantitative evaluation of comparative studies in this section has been attempted by several authors, leading to near agreement on values of relative potencies using BaP as 1 (see Table 3). 140 ORGANIC CONSTITUENTS
Table 3. Relative potencies of PAHs considered in this evaluation
Compound Ref. Ref. Ref. Ref. Ref. Ref. Summary• 1 2 3 4 5 6 Benz[ a ]anthracene 0.145 0.1 0.1 0.1 0.1 0.1 0.1 Benzo[a]pyrene 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Benzo[b ]fluoranthene 0.141 0 I 0.1 0.1 0 I 0.1 0.1 Benzo[ghi]perylene 0.022 0.01 0.01 0.01 BenzoU]fluoranthene 0.1 0.1 0.1 Benzo[ k]fluoranthene 0.061 0.1 0.1 0 I 0.01 0.1 0.1 Dibenz[ a,h ]anthracene 1.11 5 1.0 1.0 1.0 1.0 1.0 F1uoranthene 0.001 0.001 0.01 Indeno[ I ,2,3-cdjpyrene 0.232 0 I 0.1 0.1 0.1 0.1 0.1 Pyrene 0.81 0 001 0.001 0.001 a BghiP, FA, and PY are not included owing to their negative or uncertain rating as carcinogens.
References: I. Krewski et al. ( 1989) 2. Nisbet & LaGoy (1992) 3. Ma1co1m & Dobson (1994) 4. Ka1ber1ah et al. ( 1995) 5. US EPA (1993) 6. McC1ure & Schoeny ( 1995)
5.3 Studies with coal tar
Contamination of drinking-water with P AHs occurs mostly from the leaching of these compounds from coal tar coated distribution pipes. Some studies relevant to toxicity resulting from the presence of coal tar in drinking-water are therefore mentioned here, although they are not directly applicable. It should be remembered that the relative amounts of PAHs (and other compounds) in drinking-water depend on their solubility in water (e.g. FA is very soluble), and the chemical profile and concentrations will be different from that of coal tar itself. Coal tars, also known as manufactured gas plant residue (MGP), are complex mixtures containing over 1000 compounds, of which at least 30 are PAHs. The chemical composition varies with changes in feedstocks and processing temperatures. Coal tars are known skin carcinogens when applied topically to experimental animals, and this carcinogenicity correlates with their high P AH content (Wall cave et al., 1971 ). There are comparatively few studies on the carcinogenic potential of coal tars after chronic ingestion. Only those studies relevant to oral toxicity are mentioned here. POLYNUCLEAR AROMATIC HYDROCARBONS 141
5.3.1 Mutagenicity and related end-points
Coal tar paints (CTP) used in potable supply systems have been found to be mutagenic in the Ames test with metabolic activation (Robinson et al., 1984; Silvano & Meier, 1984). In a mutagenicity study on water from water distribution pipes before and after the water treatment process, the mutagenic activity did not correlate with the levels of P AHs in the water (Basu et al., 1987).
5.3.2 Carcinogenicity
CTP was positive in a dermal initiation/promotion assay with SENCAR mice, and one coal tar product was positive when tested as a complete carcinogen in the mouse at 2 J.!l per dermal application once weekly for 30 weeks (Robinson et al., 1984). The biological responses to the products were greater than expected from their P AH content. In a further study, a suspension of CTP particulate was administered to groups of 40 female A/J mice by gavage over 8 weeks (Robinson et al., 1987). Total doses of 24, 240. or 1320 mg of CTP particulate resulted in 35%. 97%. and 72% of the mice developing lung tumours (tumour multiplicity: 0.46. 4.27, 4.33). Twenty nine per cent of the control mice had lung tumours (tumour multiplicity: 0.32). Forestomach tumours were induced only at the highest dose of 1320 mg of CTP particulate. Controls had no forestomach tumours. Female A/J mice (30 per group) were fed a basal gel diet for 260 days with 0.1% or 0.25% coal tar (MGP; 7 and 16.3 J.!g of BaP per day, as MGP contained 2.76 mg of BaP per g) (Weyand et al., 1995). Seventy per cent and 100% of the mice developed lung tumours. with a multiplicity of 1.19 and 12.17 tumours per mouse, respectively. Nineteen per cent of the controls had tumours. with a tumour multiplicity of 0.19 tumours per mouse. No forestomach tumours were found. Comparing these results with those reported in the same study with pure BaP (41 and 257 f.!g/day; see section 5.1.7), MGP produced a considerably higher lung tumour rate than would be expected from its BaP content. In contrast. pure BaP produced forestomach tumours, which was not the case with MGP at the given concentrations. In a follow-up study using the same dose and administration regimen (i.e. basal gel diet) in female A/J mice for 2 weeks, DNA adducts induced by MGP and BaP in mouse lung and forestomach were characterized (Weyand & Wu, 1995). The major adduct in forestomach was attributable to BaP. Three adducts were detected in mouse lung. two of which could be contributed by BbFA and BaP. respectively, but the major DNA adduct could not be attributed to any of the P AHs identified as constituents of MGP.
In a 2-year carcinogenicity bioassay, female B6C3F 1 mice were fed 0, 0.01. 0.03, 0.1. 0.3, 0.6, or 1.0% coal tar containing 2.24 mg ofBaP per g (Culp & Beland. 1994; Culp et al.. 1996). Forestomach tumours were found in each dose group. with the incidence increasing sharply from 6% in mice fed 0.1% coal tar to 30% at the 0.3% coal tar dose (equivalent to 8.4 J.!g and 19.1 f.!g of BaP per day. respectively). 142 ORGANIC CONSTITUENTS
The incidence of forestomach tumours was approximately the same at 0.3% and 0.6% coal tar but declined at the 1.0% dose, apparently as a result of mortality from a high incidence of small intestinal adenocarcinomas in mice fed 0.6% or 1.0% coal tar. Lung tumour incidence was not reported. A parallel study with 18.5, 90, or 350 )lg of BaP per day resulted in a tumour incidence of 6%, 78%, and 98%, respectively. In BaP treated mice, one major DNA adduct was observed; this adduct accounted for 7-15% of the forestomach adducts in mice fed coal tar. A dose-related increase was observed in adduct levels in the forestomachs of BaP- and coal tar-fed mice. From the above study, it can be seen for comparison that the same (6%) forestomach tumour incidence was noted at an oral dose of 18.5 )lg of BaP per day and a 0.1% dose of coal tar containing 8.4 )lg of BaP per day (i.e. coal tar has more than twice the tumorigenic potency as BaP). Therefore, it seems that the effects of complex mixtures may be different from those of the P AHs alone. Interaction of P AHs and other compounds in coal tar may cause higher or lower tumour rates than can be expected from their content of known carcinogenic P AHs.
6. EFFECTS ON HUMANS
Human exposure to P AHs is not to individual compounds but to a mixture of these compounds in either occupational or environmental situations. There are no reports on the effects of oral ingestion by humans of the PAHs selected for evaluation, although people who consume grilled or smoked food do ingest these compounds. A high lung cancer mortality in Xuan Wei, China, has been linked to PAH exposure from unvented coal combustion (Mumford et al., 1987; Lewtas et al., 1993). PAHs present in tobacco smoke (mainstream and sidestream) are implicated as contributing to lung and other cancers (IARC, 1986; Grimmer et al., 1987, 1988). Most available human data are from inhalation and percutaneous absorption of P AHs from a large range of occupational exposures. In earlier times, following high dermal exposure, chimney sweeps developed skin cancers, especially scrotal cancer. Epidemiological studies are available for workers exposed at coke ovens in coal coking and coal gasification, in asphalt works, in foundries, in aluminium production plants, and from diesel exhaust (Verma et al., 1992; Armstrong et al., 1994; Partanen & Boffetta, 1994; Costantino et al., 1995). In all these occupations, there is also exposure to other chemicals, making a direct correlation of cause to increased levels in lung cancer more problematic. There is additionally the confounding factor of smoking. Evaluation of these studies shows, however, that it is plausible that the increased risk of lung cancer occurring in several of these occupations can be attributed at least in part to P AHs (WHO, 1997). Biomarkers have been developed to assess internal PAH exposure (WHO, 1997). Most studies focus on measurement of P AH metabolites in urine, of which 1-hydroxypyrene is the most widely used (Levin, 1995). Pyrene is normally abundant in environmental P AH mixtures. Increased urinary levels of 1-hydroxypyrene have been found, for example, in patients cutaneously treated with coal tar, in workers exposed to creosote oil, in coal tar distillery workers, in road POLYNUCLEAR AROMATIC HYDROCARBONS 143 paving workers, in coke oven workers, and in workers exposed to bitumen fumes (Jongeneelen et al., 1986, 1988a,b; Clonfero et al., 1989; Burgaz et al., 1992; Jongeneelen, 1992; Ny et al., 1993; Levin et al., 1995). Significant correlations were obtained between urinary 1-hydroxypyrene of coke oven workers or city residents and levels of PY or BaP in the ambient air (Zhao et al., 1990, 1992; Sherson et al., 1992). A controlled human exposure study showed that a 100- to 250-fold increase in a dietary dose paralleled a 4- to 12-fold increase in urinary 1-hydroxypyrene elimination (Buckley & Lioy, 1992). Trial studies suggest that urinary 1-hydroxypyrene may be a useful marker of PAH pollution in the environment (Kanoh et al., 1993). Background levels amount to 0.06-0.23 f.!mol/mol of creatine in non-smokers. Smokers have about double that level (WHO, 1997).
7. GUIDELINE VALUES
Evidence that mixtures of P AHs are carcinogenic in humans comes primarily from occupational studies of workers. Cancer associated with exposure to P AH containing mixtures in humans occurs predominantly in the lung and skin following inhalation and dermal exposure, respectively. There are no data available for humans for the oral route. There are only a few animal carcinogenicity studies on oral administration of PAHs. BaA, BaP, DBahA, and mixtures of PAHs (coal tar) were tested orally and were carcinogenic. Most studies found forestomach tumours in rodents. The best data are from the BaP study by Neal & Rigdon (1967), described in WHO (1996), although this study is inadequate in many respects (Rugen et al., 1989; Collins et al., 1991 ). Results from recent studies with BaP by Weyand et al. (1995) and Culp et al. (1996), although limited, are in agreement with risk calculations based on this older study. Further information is available on the carcinogenicity of single P AHs from experiments with dermal application. Following dermal exposure, BaA, BaP, BbFA, BjF A, BkFA, DBahA, and IP are tumorigenic in mice. FA was not positive in the mouse skin assay but was found to be tumorigenic in the intraperitoneal lung adenoma assay in newborn mice. The relevance of these types of short-term cancer bioassays is under discussion. From the limited data available, BghiP and PY are not carcinogenic. With the exception ofPY (equivocal results), all PAHs discussed here are genotoxic at least in vitro (ATSDR, 1995; WHO, 1997). Table 2 compares evaluations of individual PAHs for carcinogenicity in animals and humans. It is not possible to assess directly the risk of P AHs to humans for the oral route owing to a lack of human data. One must rely on animal data to estimate the risk of exposure to individual P AHs, not forgetting that humans are exposed to mixtures of P AHs and not to pure individual P AHs. The extrapolation of risk to humans from animal data is complicated: the relevance of forestomach tumours in rodents when considering extrapolation to humans is not clear. There is some indication that there are interspecies differences in the enzymes that activate PAHs (Michel et al., 1995); further, intraspecies differences in susceptibility in humans may be due to differences in cytochrome P-450 enzymes (Guengerich & Shimada, 1991). 144 ORGANIC CONSTITUENTS
7.1 Guideline value for BaP
The guideline value for BaP, one of the most carcinogenic PAHs, in drinking-water corresponding to an excess lifetime cancer risk of 1o-s was estimated as 0.7 f.!g/litre (WHO, 1996). This is based on the oral carcinogenicity study ofNeal & Rigdon (1967) and calculated using a two-stage birth-death mutation modeL which incorporates variable dosing patterns and time of sacrifice (Thorslund & Farrar, 1990). The data of Weyand et al. (1995) and Culp et al. (1996) on forestomach tumour incidence in mice give nearly identical results, giving support for the validity of the Neal & Rigdon (1967) study. If BaP is present in drinking-water at significant concentrations, this indicates the presence of coal tar particles, which may arise from seriously deteriorating coal tar linings.
7.2 Guideline value for FA
FA is the P AH most commonly detected in drinking-water, primarily in association with coal tar linings of cast or ductile iron distribution pipes. A guideline value for this PAH was estimated from a 13-week oral gavage study in mice with a NOAEL of 125 mg/kg of body weight per day, based on increased SGPT levels, kidney and liver pathology, and clinical and haematological changes. An uncertainty factor of 10 000 (1 00 for inter- and intraspecies variation, 10 for the use of a subchronic study and inadequate database, and 10 because of clear evidence of co-carcinogenicity with BaP in mouse skin painting studies) gives a TDI of 0.0125 mg/kg of body weight per day. Assuming a 60-kg adult drinking 2 litres of water per day with an allocation of 1% of the TDI to water, because there is significant exposure from food, a health-based value of 4 f.!g/litre (rounded figure) can be calculated. This health-based value is significantly above the concentrations normally found in drinking-water. Under usual conditions, therefore, the presence of FA in drinking-water does not represent a hazard to human health. For this reason, the establishment of a numerical guideline value for FA is not deemed necessary.
7.3 Relative potency of other PAHs compared with BaP
Attempts have been made to compare the carcinogenicity of individual P AHs using BaP as a standard. Although there are no comparative studies on the oral toxicity of P AHs, there are several comparative studies based on mouse skin carcinogenesis, initiation/promotion on mouse skin, intrapulmonary administration to rats, subcutaneous injection in mice, and intraperitoneal injection in newborn mice. There have been various attempts to rank selected P AHs in order of potential potencies based on these studies (see Table 3). Owing to the fact that the relative potencies of the individual P AHs were comparable in these studies, although the route of application was different, it is assumed that the relative carcinogenicity of these compounds is also similar for the oral and other routes of application. POLYNUCLEAR AROMATIC HYDROCARBONS 145
7.4 Complex mixtures
It cannot be assumed that the carcinogenic effects of individual P AHs are additive or that PAHs present in a mixture (e.g. coal tar) act in the same way as each PAH individually. There is ample evidence for enhancement or inhibition of carcinogenicity by other PAHs (see Warshawsky et al., 1993; ATSDR, 1995).
7.5 Recommendations
Although WHO (1996) called for the use of coal-tar-based pipe linings to be discontinued, it is apparent from reports in the recent literature that coal tar linings are still being used in new as well as in existing pipes. Furthermore, monitoring studies in areas where these coal tar linings are still in existence show that, depending on the conditions (particularly where soft corrosive water is being carried), the linings seem to be deteriorating, releasing particulate matter containing PAHs into the water supply. Such particulate matter is also released during repair work on water pipes with coal-tar-based linings. This particulate matter is likely to contain the more carcinogenic PAHs (e.g. BaP). It is recommended, as before, that: • the use of coal-tar-based and similar materials for pipe linings and coatings on storage tanks be discontinued; and • the monitoring of levels of individual indicator PAHs (including FA and BaP) and not just total P AHs in drinking-water continue, with the objective of detecting where coal-tar-based linings are deteriorating, so that they can be replaced in a timely manner by new pipes.
8. REFERENCES
Armstrong B et al. (1994) Lung cancer mortality and polynuclear aromatic hydrocarbons: a case cohort study of aluminum production workers in Arvida, Quebec, Canada. American journal ofepidemiology, 139:250-262. Atkinson R (1987) Structure-activity relationship for the estimation of rate constants for the gas phase reactions of OH radicals with organic compounds. International journal of chem1cal kmeflcs, 19:799-828. A TSDR (1995) Toxicological profile for polycyclic aromatic hydrocarbons (PAHs). Atlanta, GA, US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry. Babelek T, Ciezkowski W (1989) Polycyclic aromatic hydrocarbons as an indicator of contamination of medicinal waters in the spas in the Sudetes Mountains of southwestern Poland. Environmental geology and water sc1ences, 14:93-97. Baker JE, Eisenreich SJ (1990) Concentrations and fluxes of polycyclic aromatic hydrocarbons and polychlorinated biphenyls across the air-water interface of Lake Superior. Environmental science and technology, 24:342-352. Bartosek I et al. (1984) Comparative kinetics of oral benz(a)anthracene, chrysene and triphenylene in rats. study with hydrocarbon mixtures. Toxicology letters, 23:333-339. Basu DK, Saxena J (1978a) Monitoring of polycyclic aromatic hydrocarbons in water. 11. Extraction and recovery of six representative compounds with polyurethane foams. Environmental science and technology, 12:791-795. 146 ORGANIC CONSTITUENTS
Basu DK, Sax en a J (1978b) Polynuclear aromatic hydrocarbons in selected US drinking water and their raw water sources. Environmental sc1ence and technology, 12:795-798. Basu DK et al. (1987) Comparison of drinking water mutagenictty with leaching of polycyclic aromatic hydrocarbons from water distribution pipes. Chemosphere, 16:2595-2612. Berbee RPM (1992) PAH in the aquatic environment: sources and emissions. Summary. In: Proceedings: Workshop on polycyclic aromatic hydrocarbons (PAH), Oslo, 11-IJ November I99I. Norwegian State Pollution Control Authority (SFT), Norwegian Food Control Authority (SNT). Paris Commission (Report No. TA-816 1992). Berglind L (1982) Determination of polycyclic aromatic hydrocarbons in mdustrial discharges and other aqueous effluents. Oslo, Central Institute for Industrial Research, 21 pp. (Nordic PAH Project, Report No. 16). Bhatia AL, Tausch H, Stehlik G (1987) Mutagenicity of chlorinated polycyclic aromatic compounds. Ecotoxicology and environmental safety, 14:48-55. Biancifiori C, Caschera F (1962) The relation between pseudopregnancy and the chemical induction by four carcinogens of mammary and ovarian tumours in BALB/C mice. British journal of cancer, 16:722-730. Bjorseth A, Sortland 0 (1983) Long-range transport of polycyclic aromatic hydrocarbons. In: Bjorseth A, ed. Polycyclic aromatic hydrocarbons. New York, NY, Marcel Dekker, pp. 507-524. Bock FG (1964) Early effects of hydrocarbons on mammalian skin. Progress m experimental tumor research, 4:126-168. Bock FG, King DW (1959) A study of the sensitivity of the mouse forestomach toward certain polycyclic hydrocarbons. Journal of the National Cancer Institute, 23:833-838. Bock FG, Mund R (1958) A survey of compounds for activity in suppression of mouse sebaceous glands. Cancer research, 18:887-892. Bossert ID, Bartha R (1986) Structure-biodegradability relationships of polycyclic aromatic hydrocarbons in soil. Bulletin ofenvironmental contamination and toxicology, 37:490-495. Boyland E. Sims P (1967) The carcinogenic activities in mice of compounds related to benz[a]anthracene. International ;ournal ofcancer, 2:500-504. Bryan WR, Shimkin MB (1943) Quantitative analysis of dose-response data obtained with three carcinogenic hydrocarbons in strain C3H male mice. Journal of the Natwnal Cancer Institute, 3:503-531. Buckley TJ, Lioy PJ (1992) An examination of the time course from human dietary exposure to polycyclic aromatic hydrocarbons to urinary elimination of 1-hydroxypyrene. Bntish journal of mdustrial medic me, 49: 113-124. Burgaz S, Borm PJA, Jongeneelen F (1992) Evaluation of urinary excretion of 1-hydroxypyrene and thioethers in workers exposed to bitumen fumes. International archives of occupational and environmental health, 63:397-40 I. CEC (1995) Proposal for a council directive concermng the quality of water intended for human consumptiOn. Presented by the Commission of the European Communities, 83 pp. (COM(94) final). Chang LH (1943) The fetal excretion of polycyclic hydrocarbons following their administration to the rat. Journal of bwlogical chemistry, 151 :93-99. Chen PH, Shieh HH, Gaw JM (1980) Determination of polycyclic aromatic hydrocarbons in airborne particulates at various sites in Taipei city by GC/MS and glass capillary GC. Proceedmgs of the National Science Council, 4:280-284. Clement Associates (1988) Comparative potency approach for est1matmg the cancer nsk associated with exposure to m1xtures of polycyclic aromatic hydrocarbons Fairfax, VA, ICF Clement Associates, 125 pp. (Interim Final Report 68-02-4403). Clonfero E et al. (1989) Biological monitoring of human exposure to coal tar. Urinary excretion of total polycyclic aromatic hydrocarbons, 1-hydroxypyrene and mutagens in psoriatic patients. International archives of occupc!lional and environmental health, 61:363-368. Collins JF et al. (1991) Risk assessment for benzo[a]pyrene. Regulatory toxicology and pharmacology, 13:170-184. POLYNUCLEAR AROMATIC HYDROCARBONS 147
Coover MP, Sims RC (1987) The effect of temperature on polycyclic aromatic hydrocarbon persistence in an unacclimated agricultural soil. Hazardous waste and hazardous materials, 4:69-82. Costantino JP, Redmond CK, Bearden A (1995) Occupationally related cancer risk among coke oven workers: 30 years of follow-up. Journal of occupational and environmental med1cme, 37:597-604. Culp SJ, Beland FA (1994) Comparison of DNA adduct formation in m1ce fed coal tar or benzo[a]pyrene. Carcinogenesis, 15:247-252. Culp SJ et al. (1996) DNA adduct measurements in relation to tumor incidence during the chronic feeding of coal tar or benzo(a)pyrene to mice. Polycyclic aromatic compounds, JJ: 161-168. Davi ML et al. (1994) Determination of polycyclic aromatic hydrocarbons in drinking water by mass spectrometry. Life chemistry reports, 10:181-188. Dennis MJ et al. (1983) Analysis of polycyclic aromatic hydrocarbons in UK total diets. Food and chemical toxicology, 21:569-574 Dennis MJ et al. (199 I) Factors affecting the polycyclic aromatic hydrocarbons content of cereals, fats and other food products. Food additives and contaminants, 8:517-530. Desiden PG et al. (1984) Concentration, separation and determination of hydrocarbons in sea water. Journal of chromatography, 284:167-178. Deutsch-Wenzel RP et al. (1983) Experimental studies in rat lungs on the carcinogenicity and dose response relationships of eight frequently occurring environmental polycyclic aromatic hydrocarbons. Journal of the National Cancer lnstllute, 71:539-544. de V os RH et al. ( 1990) Polycyclic aromatic hydrocarbons in Dutch total diet samples (1984-1986). Food and chemical toxicology, 28:263-268. Dieter HH (1994) [Drinking water.]Jn: Wichmann HE, Schlipkoter H-W, Ftilgraff G, eds. Handbuch der Umweltmedizin, 5th ed. Landsberg, Ecomed Fachverlag, pp. 1-53 (in German). EEC (1980) Council Directive of 15 July 1980 relating to the quality of water intended for human consumption (801778/EEC). Official JOUrnal of the European Commission, No. L 229, 30.8.1980, pp. 11-29. Environment Canada (1994) Canadian Environmental Protection Act Prionty Substances List assessment report polycyclic aromatic hydrocarbons Ottawa, Ontario, Supply and Services Canada, 61 pp. Foth H, Kahl R, Kahl GF (1988) Pharmacokinetics of low doses of benzo(a)pyrene in the rat. Food and chemical toxicology, 26:45-51. Franck HG, Stadelhofer JW (1987) [Industrial aromatic chemistry - raw products, processes, products) Berlin, Springer-Verlag, pp. 308-380 (in German). Gorelick NJ et al. (1989) Formation of DNA and hemoglobin adducts offluoranthene after single and multiple exposures. Carcinogenesis, 10:1579-1587. Greenberg A (1996) Measurement of benzo[a]pyrene as a surrogate for total human exposure to PAH. Polycyclic aromatic compounds, 11:153-160. Griesbaum K et al. ( 1989) Hydrocarbons. In: Ullmann "s encyclopedia of industnal chemistry, 5th ed. Vol. AJJ. High-performancefibers to Imidazole and derivatives. Weinheim, Verlag Chemie, pp. 227-281. Grimmer G, Jacob J, Naujack KW (1981) Profile of the polycyclic aromatic hydrocarbons from lubricating oils. Inventory by GCGC/MS - PAH in environmental materials, part I. Fresenius' Zeilschrift fiir Analyt1sche Chemie, 306:347-355. Grimmer G, Naujack KW, Dettbarn G (1987) Gas chromatographic determination of polycyclic aromatic hydrocarbons, aza-arenes, aromatic amines in the particle and vapor phase of mainstream and sidestream smoke of cigarettes. Toxicology letters, 35:117-124. Grimmer G et al. (1988) Contribution of polycyclic aromatic compounds to the carcinogenicity of sidestream smoke of cigarettes evaluated by implantation into the lungs of rats. Cancer letters, 43:173-177. Grover PL, Sims P (1968) Enzyme-catalysed reactions of polycyclic hydrocarbons with deoxyribonucleic acid and protein m vitro. Bwchemical JOUrnal, 110:159-160. Guengerich FP, Shimada T (1991) Oxidation of toxic and carcinogenic chemicals by human cytochrome P-450 enzymes. Chemical research in toxicology, 4:391-407. 148 ORGANIC CONSTITUENTS
Guicherit R, Schulting FL (1985) The occurrence of organic chemicals in the atmosphere of the Netherlands. Sc1ence of the total enVIronment, 43:193-219. Habs M, Schmahl D, Misfeld J (1980) Local carcinogenicity of some environmentally relevant polycyclic aromatic hydrocarbons after lifelong topical application to mouse skin. Arch1v ji1r Geschwulstforschung, 50.266-274. Hoff RM, Chan KW ( 1987) Measurement of polycyclic aromatic hydrocarbons in the air along the Niagara River. Environmental sc1ence and technology, 21:556-561. Hoffmann D, Wynder EL (1966) (Carcinogenic effect of dibenzopyrenes.] Zeztschrift fur Krebsforschung, 68:137-149 (in German). Hoffmann D et al. (1972) Fluoranthenes: quantitative determination in cigarette smoke, formation by pyrolysis, and tumor-initiating activity. Journal of the Natwnal Cancer Institute, 49:1165-1175. Horton AW, Christian GM (1974) Cocarcinogenic versus incomplete carcinogenic activity among aromatic hydrocarbons: contrast between chrysene and benzo[b ]tnphenylene. Journal of the Natwnal Cancer Institute, 53.1017-1020. !ARC (I 983) Po(vnuclear aromatzc compounds, part I· Chemzcal, environmental and expenmental data Lyon, International Agency for Research on Cancer, 477 pp. (!ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 32). !ARC (1984) Polynuclear aromatic hydrocarbons, part 2 Carbon blacks, mmeral ozls and some nitroarenes. Lyon, International Agency for Research on Cancer, 365 pp. (!ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 33). !ARC (1985) Polynuclear aromatic compounds, part 4. Bztumens, coal-tars and denved products, shale-oils and soots. Lyon, International Agency for Research on Cancer, 271 pp. (!ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 35). !ARC (1986) Tobacco smoking. Lyon, International Agency for Research on Cancer, 139 pp. (!ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 38). !ARC (1987) Overall evaluations ofcarcmogemclty an updatmg ofiARC Monographs volumes 1- 42 Lyon, International Agency for Research on Cancer, 403 pp. (!ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Suppl 7). !ARC (I 989) Occupational exposures in petroleum refining, crude ad and ma;or petroleum fuels Lyon, International Agency for Research on Cancer, 283 pp. (!ARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Man, Vol. 45). Ihme W, Wichmann HE (1996) [Exposure assessment via model calculations and simulations- oral human exposure to PAH via drinking water and other sources.] Umweltwissenschajien und Schadsto.ffForschung, 8(6):343-354 (in German). Jansen EH et al. (I 996) Route-specific urinary biomarkers in the risk assessment of PAH exposure. Polycyclzc aromatic compounds, 11:185-192. Jongeneelen FJ (1992) BIOlogical exposure limit for occupational exposure to coal tar pitch volatiles at coke ovens. International archives of occupational and environmental health, 63:511-516. Jongeneelen FJ et al. (1986) Biological monitoring of polycyclic aromatic hydrocarbons. Metabolites in urine. Scandmavwn ;ournal of work, environment & health, 12: 13 7-143. Jongeneelen FJ et al (1988a) Airborne concentrations, skin contaminatiOn, and urinary metabohte excretion of polycyclic aromatic hydrocarbons among paving workers exposed to coal tat derived road tars. American Industnal Hyg1ene Association;ournal, 49:600--607. Jongeneelen FJ et al (1988b) 1-Hydroxypyrene in urine as a biological mdicator of exposure to polycyclic aromatic hydrocarbons in several work environments. Annals of occupatwnal hygzene, 32:35-43. Kalberlah F, Frijus-Plessen N, Hassauer M (1995) [Toxicological criteria for the risk assessment of polyaromatic hydrocarbons (PAH) in existing chemicals. Part I: The use of equivalency factors.] Altlasten-Spektrum, 5·231-237 (in German). Kanoh T et al. (1993) Unnary 1-hydroxypyrene as a marker of exposure to polycyclic aromatic hydrocarbons in environment. Env1romnental research, 62:230-241. Kawamura Y et al. (1988) The effect of various foods on the intestinal absorption of benzo(a)pyrene in rats. Journal ofthe Food and Hygzene Society ofJapan, 29·21-25. POLYNUCLEAR AROMATIC HYDROCARBONS 149
Klein M (1963) Susceptibility of strain B6AF/J hybrid infant mice to tumorigenesis with 1,2- benzanthracene, deoxycholic acid, and 3-methylcholanthrene. Il Tumours called forth by painting the skin with dibenzpyrene. Cancer research, 23:1701-1707. Krewski D, Thorslund T, Withey J (1989) Carcinogenic risk assessment of complex mixtures. Toxicology and mduslrlal health, 5:851-867. Kulkarni MS et al. (1986) Species differences in the formation of benzo(a)pyrene-DNA adducts in rodent and human endometrium. Cancer research, 46:2888-2891. Kveseth K, Sortland B, Bokn T (1982) Polycyclic aromatic hydrocarbons in sewage, mussels and tap water. Chemosphere, 11:623--639. LA! (1992) [Cancer nsk from Glr pollution.] Entwicklung von "BeurteilungsmaBstaben flir kanzerogene Luftverunreinigungen" im Auftrage der Umweltministerkonferenz. Dlisseldorf, Ministerium flir Umwelt, Raumordnung und Landwirtschaft des Landes Nordrhein Westfalen, LanderausschuB flir lmmissionsschutz (Federal Committee for Control of Ambient Air Levels), 71 pp. (in German). Larsson BK (1982) Polycyclic aromatic hydrocarbons in smoked fish. Zelfschrift fur Lebensmittei Untersuchung und -Forschung, 174: I 01-107. Larsson BK (1986) Po~vcyclic aromatic hydrocarbons in Swedish foods: Aspects on analysis, occurrence and mtake. Uppsala, Swedish Food Research Department, The National Food Administration, 60 pp. La Voie EJ et al. (1982) Tumour initiating activity of hydrodiols of benzo[b ]fluoranthene, benzoU]fluoranthene, and benzo[k]fluoranthene. Carcmogenesis, 3:49-53. La Voie EJ et al. (1987) Tumorigenic activity of non-alternant polynuclear aromatic hydrocarbons in newborn mice. Cancer letters, 34:15-20. Levin JO ( 1995) First international workshop on hydroxypyrene as a biomarker for PAH exposure in man~ summary and conclusions. Science of the total environment, 163:164-168. Levin JO, Rhen M, Sikstrom E (1995) OccupatiOnal PAH exposure: urinary 1-hydroxypyrene levels of coke oven workers, aluminium smelter pot-room workers, road pavers, and occupationally non-exposed persons in Sweden. Sc1ence ofthe total enVIronment, 163:169-177. Lewtas Jet al. (1993) Comparison of DNA adducts from exposure to complex mixtures in various human tissues and experimental systems. Environmental health perspecllves, 99:89-97. Lipniak M, Brandys J (1993) Toxicokinetics of fluoranthene, pyrene and benz(a)anthracene in the rat Polycyclic aromatic compounds, 3:111-119. Mackay D, Shiu WY, Ma KC (1992) Illustrated handbook of physical-chemical properties and env~ronmental fate for orgamc chemicals Vol. 11. Polynuclear aromatic hydrocarbons, polychlorinated dioxins and diben=ofurans. Boca Raton, FL, Lewis Publishers. Maga JA (1986) Polycyclic aromatic hydrocarbon (PAH) composition of mesquite (Prosopis fuliflora) smoke and grilled beef. Journal ofagricultural and food chenustry, 34:249-251. Malcolm HM, Dobson S (1994) The calculation of an environmental assessment level (EAL) for atmospheric PAHs using relative potencies. London, Department of the Environment, 34 pp. (Report No. DoE/HMIP/RR/94/041 ). McClure P, Schoeny R ( 1995) Evaluation of a component-based relat1ve potency approach to cancer risk assessment for exposure to PAH. Presented at 15th International Symposium on Polycyclic Aromatic Compounds, Belgirate (Italy), 19-22 September 1995. International Society for Polycyclic Aromatic Compounds, European Chemicals Bureau, and European Centre for the Validation of Alternative Methods of the Joint Research Centre of the European Commission, Ispra. McVeety BD, Hites RA (1988) Atmospheric deposition of polycyclic aromatic hydrocarbons to water surfaces: a mass balance approach. Atmospheric enVIronment, 22:511-536. Menichini E, Bonanni L, Merit F (1990) Determination of polycyclic aromatic hydrocarbons in mmeral oils and oil aerosols in glass manufacturing. Toxicological and environmental chemistry, 28:37-51. Michel X et al. (1995) Regw-selecilve metabolism of ben=o[a]pyrene by microsomes from 5 vertebrate species Presented at 15th International Symposium on Polycyclic Aromatic Compounds m Belgirate (Italy), 19-22 September 1995. 150 ORGANIC CONSTITUENTS
Mill T, Mabey W (1985) Photochemical transformations. In: Neely WB, Blau GE, eds. Environmental exposure from chemicals. Vol. I. Boca Raton, FL, CRC Press, pp. 175-216. Mill T et al. (1981) Photolysis of polycyclic aromatic hydrocarbons in water. Chemosphere, 10:1281-1290. Mumford JL et al. (1987) Lung cancer and indoor air pollution in Xuan Wei, China. Science, 235:217-220. Neat J, Rigdon RH (1967) Gastric tumors in mice fed benzo(a)pyrene: a quantitative study. Texas reports on bwlogy and med1cine, 25:553-557. Neubert D, Tapken S ( 1988) Transfer of benzo(a)pyrene into mouse embryos and fetuses. Archives of toxicology, 62:236-239. Nisbet IC, LaGoy PK (1992) Toxic equivalency factors (TEFs) for polycyclic aromatic hydrocarbons (PAHs). Regulatory toxicology and pharmacology, 16:290-300. Ny ET et al. (1993) The relationship between polycyclic aromatic hydrocarbons in air and in urine of workers in a S6derberg potroom. Amencan Industrial Hygiene Association journal, 54:277-284. Palitti F et al. ( 1986) An in wtro and m vivo study on mutagenic activity of fluoranthene: comparison between cytogenetic studies and HPLC analysis. Mutation research, 174:125-130. Park KS et al. (1990) Fate of PAH compounds in two soil types: influence of volatilization, abiotic loss and biological activity. Environmental toxicology and chemistry, 9:187-195. Partanen T, Boffetta P (1994) Cancer risk in asphalt workers and roofers: review and meta-analysis of epidemiologic studies. American journal of industrial medicine, 26:721-740. Pfannhauser W (1991) [Polycyclic aromatic hydrocarbons (PAH) in food and on samples of selected vegetables in Austria {1991 ).] Mitteilungen aus de m Gebiete der Lebensmitteluntersuchung und Hygiene, 82:66-79. Phillips DH, Grover PL (1994) Polycyclic hydrocarbon activation: bay regions and beyond. Drug metabolism rev1ews, 26:443-467. Rahman A, Barrowman JA, Rahimtula A (1986) The influence of bile on the bioavailability of polynuclear aromatic hydrocarbons from the rat intestine. Canadian journal of physwlogy and pharmacology, 64:1214-1218. Raiyani CV et al. (1993) Assessment of indoor exposure to polycyclic aromatic hydrocarbons for urban poor using various types of cooking fuels. Bulletin of environmental contamination and tOXICOlogy, 50:757-763. Reddy MY et al. (1984) 32P-postlabeling test for covalent DNA binding of chemicals in vivo: application to a variety of aromatic carcinogens and methylating agents. Carcmogenes1s, 5:231-243. Robinson M et al. (1984) Comparative carcinogenic and mutagenic activity of a coal tar and petroleum asphalt paints used in potable water supply systems. Journal of applied tOXICOlogy, 4:49-56. Robinson M et al. ( 1987) Carcinogenic effects in A/J mice of particulate of a coal tar paint used in potable water systems. Cancer letters, 34:49-54. Rostad CE, Pereira WE (1987) Creosote compounds in snails obtamed from Pensacola Bay, Florida, near an onshore hazardous-waste site. Chemosphere, 16:2397-2404. Rugen PJ, Stern CD, Lamm SH (1989) Comparative carcinogenicity of the PAHs as a basis for acceptable exposure levels (AELs) in drinking water. Regulatory toxicology and pharmacology, 9:273-283. Sandell E, Tuominen J (1996) The impact of the use and disposal of creosote impregnated railway ties on a freshwater supply area in southern Finland. Polycyclic aromatic compounds, 11:83-90. Shendrikova lA, Aleksandrov VA (1974) Comparative penetration of polycyclic hydrocarbons through the rat placenta into the fetus in rats. Bulletm of experimental biology and medicine, 77:169-171. Sherson D et al. (1992) Interaction of smoking, PAH uptake and cytochrome P450IA2 activity among foundry workers. British journal of mdustrial medicine, 49:197-202. Shiraishi N et al. ( 1985) Occurrence of chlorinated polynuclear aromatic hydrocarbons in tap water Environmental sc1ence and technology, 19:585-590. POLYNUCLEAR AROMATIC HYDROCARBONS 151
Silvano M, Meier JR (1984) Mutagenicity of coal tar paints used in drinking water distribution systems. Science ofthe total environment, 39:251-263. Sirota GR, Uthe JF (1981) Polynuclear aromatic hydrocarbon contamination in marine shellfish. In: Cooke M, Dennis AJ, eds. Polynuclear aromatic hydrocarbons. chemical analysis and biological fate. 5th International Symposium on Polynuclear Aromatic Hydrocarbons, Columbus, OH, October 1980. Columbus, OH, Battelle Press, pp. 329-341. Smyth HF et al. (1962) Range-finding toxicity data: List VI. Industrial hygiene journal, March April:95-107. Snell KC, Stewart HL (1962) Pulmonary adenomatosis induced in DBA/2 mice by oral administration of dibenz[a,h]anthracene. Journal ofthe National Cancer Institute, 28:1043-1049. Southworth GR (1979) The role ofvolatilization in removing polycyclic aromatic hydrocarbons from aquatic environments. Bulletin ofenvironmental contammatwn and toxicology, 21:507-514. Speer K et al. (1990) Determination and distribution of PAH in native vegetable oils, smoked fish products, mussels and oysters, and bream from the river Elbe. Journal of high resolutwn chromatography, 13:104-111. State Chemical Analysis Institute (1995) Annual report for 1994 Food surveillance and environmental control. Freiburg, 6 pp. Stocker KJ et al. ( 1996) Assessment of the potential in vivo genotoxicity of fluoranthene. Mutagenesis, 11 :493--466. Stuermer DH, Ng DJ, Morris CJ (1982) Organic contaminants in groundwater near underground coal gasification site in northeastern Wyoming. Enwronmental science and technology, 16:582-587. Tetzen D (1989) [Environmentally relevant classification of PKWF and Printosol. Classification of printing ink oils.] Verjkroniek, 62:469--472 (in German). Thomson B, Lake R, Lill R (1996) The contribution of margarine to cancer risk from polycyclic aromatic hydrocarbons in the New Zealand diet. Polycychc aromatic compounds, 11:177-184. Thorslund TW, Farrar D (1990) Ingestwn dose-response model for benzo[a]pyrene. Unpublished report prepared for the US Environmental Protection Agency, 22 pp. Thruston AD Jr ( 1978) High pressure liquid chromatography techniques for the isolation and identification of organics in drinking water extracts. Journal of chromatographic science, 16:254-259 Tuominen JP, Pyysalo HS, Sauri M (1988) Cereal products as a source of polycyclic aromatic hydrocarbo.ns. Journal of agricultural and food chemistry, 36:118-120. Turrio-Baldassarri L et al. (1996) Polycyclic aromatic hydrocarbons in Italian national and regional diets. Polycychc aromatic compounds, 10:343-349. US EPA (1988) 13-week mouse oral subchronic toxicity study (jluoranthene) Muskegon, M!, Toxicity Research Laboratories, I 04 pp. (TRL Study #042-008). US EPA (1989) Mouse oral subchronic toxicity of pyrene Muskegon, M!, Toxicity Research Laboratories, 102 pp. (TRL Study #042-012). US EPA (1992) Drinking water cntena document for polycyclic aromatic hydrocarbons (PAHs) Prepared by the Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office, Cincinnati, OH, for the Office of Water, Washington, DC, 442 pp. US EPA (1993) ProviSional guidance for quanfl/at1ve nsk assessment of polycyclic aromatic hydrocarbons Office of Health and Environmental Assessment (EP A/600/R-93/089). Vaessen HA et al. (1984) Polycyclic aromatic hydrocarbons in selected foods; analysis and occurrence. Toxicological and environmental chemist!)', 7:297-324. van Duuren BL, Goldschmidt BM ( 1976) Cocarcinogenic and tumor-promoting agents in tobacco carcinogenesis. Journal of the Nat1onal Cancer Institute, 56·1237-1242. van Noort PC, Wondergem E (1985) The isolation of some polynuclear aromatiC hydrocarbons from aqueous samples by means of reversed-phase concentrator columns. Analytica Chimica Acta, 172:335-340. Verma DK et al (1992) Polycyclic aromatic hydrocarbons (PAHs): a possible cause of lung cancer mortality among nickel/copper smelter and refinery workers Amencan lndustnal Hygiene Associatwnjournal, 53:317-324. 152 ORGANIC CONSTITUENTS
Vu DueT, Huynh CK (1981) (Organic micropollutants in water. Preliminary results on haloforms and polycyclic aromatic hydrocarbons.] Medecine sociale et preventive, 26:315-316 (in French). Wallcave L et al. (1971) Skin tumorigenesis in mice by petroleum asphalts and coal-tar pitches of known polynuclear aromatic hydrocarbon content. Toxicology and applied pharmacology, 18:41-52. Wang JS, Busby WF, Wogan GN (1995a) Formation and persistence of DNA adducts in organs of CD-I mice treated with a tumorigenic dose offluoranthene. Carcmogenes1s, 16:2609-2616. Wang JS, Busby WF, Wogan GN (1995b) Tissue distribution of DNA adducts in preweanling BLU: Ha mice treated with a tumorigenic dose offluoranthene. Cancer letters, 92:9-19. Warshawsky D, Barkley W, Bingham E (1993) Factors affecting carcinogenic potential of mixtures. Fundamental and applied toxicology, 20:376-382. Wenzel-Hartung R et al. (1990) Evaluation of the carcinogenic potency of four environmental polycyclic aromatic compounds following intrapulmonary application in rats. Expenmental pathology, 40:221-227. Weyand EH, Wu Y (1995) Covalent binding of polycyclic aromatic hydrocarbon components of manufactured gas plant residue to mouse lung and forestomach DNA Chemical research m toxicology, 8:955-962. Weyand EH, Rice JE, LaVoie EJ (1987) 32P-postlabeling analysis of DNA adducts from non alternant PAH using thin-layer and high performance liquid chromatography. Cancer letters, 37:257-266. Weyand EH et al. (1995) Differences in the tumorigenic activity of a pure hydrocarbon and a complex mixture following ingestion: benzo(a)pyrene vs manufactured gas plant residue. Chen11ca/ research m tox1cology, 8:949-954. WHO (1987) Polynuclear aromatic hydrocarbons (PAH). In: Air quahty gu1delines for Europe. Copenhagen, WHO Regional Office for Europe, pp. I 05-117 (WHO Regional Publications, European Series No. 23). WHO (1996) Polynuclear aromatic hydrocarbons. In: Guidelmes for drinkmg-water quality, 2nd ed Vol 2 Health cntena and other supportmg informatwn Geneva, World Health Organization, pp. 495-505. WHO (1997) Non-heterocychc po(vcyclic aromallc hydrocarbons Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria 202) (in press). Wild SR, Berrow ML, Jones KC (1991) The persistence of polynuclear aromatic hydrocarbons (PAH's) in sewage sludge amended agricultural soils. Enwronmental pollution, 72:141-157. Withey JR, Law FC, Endrenyi L (1991) Pharmacokinetics and bioavailability of pyrene in the rat. Journal of toxicology and environmental health, 32:429-447. Woidich W et al. (1976) (Analysis of polycyclic aromatic hydrocarbons m drinking and utilitary water.] Lebensmittelchenue und Genchtliche Chemie, 30:141-146 (in German). Wynder EL, Hoffmann D (1959a) A study of tobacco carcinogenesis. VII. The role of higher polycyclic hydrocarbons. Cancer, 12: I 079-1086. Wynder EL, Hoffmann D (1959b) The carcinogenicity ofbenzofluoranthenes. Cancer, 12:1194-1199. Yamazaki H, Kakiuchi Y (1989) The uptake and distribution of benzo(a)pyrene in rat after continuous oral administration Toxicological and environmental chem1stry, 24:95-104. Zepp RG, Schlotzhauer PF (1979) Photoreactivity of selected aromatic hydrocarbons in water. In· Jones PW, Leber P, eds. Polynuclear aromatic hydrocarbons Thwd lnternatwnal Symposwm on Chemistry and Bwlogy- Carcinogenesis and Mutagenesis Ann Arbor, MJ, Ann Arbor Science Publishers, pp. 141-158. Zhao Z-H, Quan W-Y, Tian D (1990) Urinary 1-hydroxypyrene as an indicator of human exposure to ambient polycyclic aromatic hydrocarbons in a coal-burnmg environment Science of the total enl'lronment, 92:145-154. Zhao Z-H, Quan W-Y, Tian D (1992) Experiments on the effects of several factors on the 1- hydroxypyrene level in human urine as an indicator of exposure to polycyclic aromatic hydrocarbons. Science ofthe total environment, 113:197-207. BENTAZONE
First draft prepared by H. Ga/al-Gorchev International Programme on Chemical Safety World Health Organization, Geneva, Switzerland
Bentazone was evaluated in the second edition of the WHO Guidelines for drinking-water quality. A guideline value for bentazone of 30 J..tg/litre was based on an ADI of 0-0.1 mg/kg of body weight established by JMPR in 1991 (based upon haematologica1 effects observed in a 2-year dietary study in rats) assuming a 60-kg person consuming 2 litres of drinking-water per day and allocating 1% of the ADI to drinking-water "to allow for uncertainties regarding dietary exposure." Bentazone is "moderately persistent in the environment" and therefore has the potential of being present in food. In 1993, a manufacturer of bentazone (BASF, Germany) sent information to IPCS concerning the environmental behaviour ofbentazone, which, in the opinion of BASF, would justify an allocation of 10% of the ADI to drinking-water, rather than 1%. The Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality therefore requested JMP to evaluate the environmental behaviour and persistence of bentazone so that a decision could be made concerning the per cent allocation of the ADI to drinking-water.
1. SUMMARY OF INFORMATION SUBMITTED BY BASF
Crop plants are able to metabolize bentazone quickly to 6- and 8-hydroxy bentazone and conjugate these with sugars. Further degradation to small fragments
(C 1-C3) occurs; the fragments are subsequently incorporated into natural plant products, such as proteins, lignin, and starch. Degradation trials show that from the time of treatment, the plant residues diminish by a factor of 100-1000 within a few days to a few weeks. Residues in raw agricultural commodities ranged from 0.1 mg/kg to approximately 1 mg/kg in straw and other "leftover" plant parts. In the upper soil layer, bentazone is quickly degraded, microbially and aerobically, via the intermediary and unstable products 6- and 8-hydroxy-bentazone and 2-amino-N-isopropyl benzamide. These are immediately bound biotically and abiotically to the mostly non-bioavailable soil organic matter fraction. Additionally, 24-50% of the bentazone is mineralized to carbon dioxide. Half-lives in field soils range from 3 to 21 days, with an average of 12 days. Abiotic degradation processes predominantly involve photolysis on plants and soil surfaces and in surface water. The average half-life in laboratory soil tests is about 45 days, considerably higher than in field tests. The difference bet.ween laboratory and field half-lives is due to an enhanced microbial activity under field conditions, which is important to bentazone degradation. 154 PESTICIDES
The high water solubility of bentazone, its low adsorption on organic soil material, and the results of soil column laboratory studies would seem, at first glance, to predict a leaching potential. However, several field lysimeter studies showed that bentazone does not leach under field conditions and thus does not pose a potential risk to groundwater. Annual average concentrations in the leachates were always less than 0.1 J.lg/litre. Bentazone has been reported to occur in groundwater and drinking-water, predominantly in filtrate collected along the banks of the Rhine River several years ago. Until 1988, effluents from the bentazone production plant were discharged into the Rhine River. Since then, bentazone levels in the Rhine River have been below or around 0.1 J.lg/litre. A few isolated point source contaminations, arising from accidents and negligence, for example, are marked by sporadic findings of concentrations above I J.lg/litre. The low octanol-water partition coefficient of bentazone precludes its bioaccumulation in food. In summary, recent studies on the environmental behaviour of bentazone have shown that the use of bentazone according to good agricultural practice does not result in the contamination of groundwater and drinking-water because of its rapid degradation in plants and soils (Huber et al., 1993; Huber & Otto, 1994).
2. JOINT MEETING ON PESTICIDES: ASSESSMENT OF THE ENVIRONMENTAL BEHAVIOUR OF BENTAZONE
The Core Assessment Group (CAG) of JMP (Environment) evaluated the environmental behaviour of bentazone. CAG concluded that the degradation rate of bentazone was significantly faster in the field than in laboratory studies. However, use of the estimated average field degradation half-time (DT50) of 12 days in simple models rather than the laboratory DT50 of 45 days, while reducing the observed potential leaching from "substantial" to "marginal," still leaves cause for concern regarding leaching. Under many field conditions, degradation of the compound will be complete in the upper soil layers. This is particularly true for dry conditions and will still hold for conditions of non-extreme rainfall. However, compounds with this high water solubility and these low soil adsorption characteristics are liable to leach under conditions of extreme rainfall (such as storms shortly after application). Bentazone will be expected to pass both through the soil profile and via cracks to the underlying aquifer. Once outside the zone of biological action, there is no abiotic mechanism for its degradation. Some contamination of the groundwater would be expected to occur under these circumstances. This potential has been confirmed by some reports of bentazone in groundwater (IPCS, in press).
3. GUIDELINE VALUE
Based on the BASF and JMP evaluations of new data on bentazone's environmental behaviour, bentazone does not seem to accumulate in the BENTAZONE 155 environment, and exposure from food is unlikely to be high. As a result, a 10% allocation of the ADI to drinking-water is appropriate for bentazone. The new guideline value, based on this 10% allocation, is therefore 300 )-lg/litre.
4. REFERENCES
Huber R, Otto S (1994) Environmental behavior of bentazon herbicide. Reviews of environmental contamination and toxicology, 137:111-134. Huber R et al. (1993) Bentazon - Assessment of environmental behavior Unpublished report submitted to WHO by BASF, Limburgerhof, Germany, August. IPCS (in press) Report of the 1997 meeting of the Core Assessment Group. Geneva, World Health Organization, International Programme on Chemical Safety, Joint Meeting on Pesticides.
CARBOFURAN
First draft prepared by H. Galal-Gorchev International Programme on Chemical Safety World Health Organization, Geneva, Switzerland
In the second edition of the WHO Guidelines for drinking-water quality, a guideline value for carbofuran of 5 f..lg/litre was recommended based on a NOAEL of 0.05 mg/kg of body weight per day, identified in a study in which carbofuran was administered orally to men (2 men/dose) at two doses; at the higher dose (0.1 mg/kg of body weight per day), symptoms of acetylcholinesterase depression were observed. A TDI of 1.67 f..lg/kg of body weight was derived by applying an uncertainty factor of30 (10 for intraspecies variation and 3 for the steep dose-response curve) to the NOAEL, and a 60-kg body weight, a 2 litres/day drinking-water consumption, and a 10% allocation of the TDI to drinking-water were assumed. In 1996, carbofuran was re-evaluated by JMPR. The Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality considered it desirable to update the evaluation of carbofuran in drinking-water on the basis of the recent JMPR evaluation.
1. GENERAL DESCRIPTION l.I Identity
CAS no.: 1563-66-2
Molecular formula: C12H 15N03
The IUPAC name for carbofuran is 2,3-dihydro-2,2-dimethylbenzofuran-7-yl methylcarbamate.
1.2 Physicochemical properties
Carbofuran is a white crystalline solid. It is stable under neutral or acidic conditions but unstable in alkaline media (FAO/WHO, 1985; Health and Welfare Canada, 1991).
Property Value Vapour pressure 2.7 mPa at 33°C Melting point 150°C Log octanol-water partition coefficient 1.6-2.3 Water solubility 350 mg/litre at 25°C Specific gravity 1.18 g/cm3 158 PESTICIDES
1.3 Major uses
Carbofuran is used worldwide for many crops. It is a broad-spectrum carbamate insecticide, acaricide, and nematicide. The technical product contains a minimum of95% carbofuran (FAO/WHO, 1977).
1.4 Environmental/ate
Carbofuran is rapidly taken up by plants through the roots from soil and water and is translocated mainly into the leaves. The main metabolite in plants has been identified as 3-hydroxycarbofuran. Carbofuran is degraded in soil by hydrolysis, microbial action, and, to a lesser extent, photo-decomposition. Its persistence is dependent upon pH, soil type, temperature, moisture content, and the microbial population. Degradation products in soil include carbofuran phenol, 3-hydroxycarbofuran, and 3-ketocarbofuran; field studies have indicated a half-life of 26-110 days in soil. Carbofuran may leach significantly, although leaching may not occur in highly organic soils. Carbofuran is degraded in water by hydrolysis, microbial decomposition, and photolysis. Hydrolysis half-lives in water at 25°C of 690, 8.2, and 1.0 weeks have been reported at pH levels of 6.0, 7.0, and 8.0, respectively (Health and Welfare Canada, 1991).
2. ANALYTICAL METHODS
The concentration of carbofuran in water may be determined by separation by high-performance liquid chromatography, hydrolysis with sodium hydroxide, extraction of the resulting methylamine with a-phthalaldehyde, and fluorescence detection of the derivative (detection limit 0.9 !J.g/litre). The concentration of carbofuran may also be quantified by acidification of the sample, extraction with dichloromethane, and separation by gas chromatography with a nitrogen-phosphorus detector (detection limit 0.1 !J.g/litre) (Health and Welfare Canada, 1991).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
In a study designed to evaluate human exposure to carbofuran following aerial applications, it was estimated that maximum inhaled doses were in the range 0. 7-2.0 mg/day (Draper et al., 1981 ).
3.2 Water
Carbofuran was detected only once (at 3.0 !J.g/litre) in 678 samples from surveys of Canadian municipal and private water supplies conducted from 1971 to 1986 (Health and Welfare Canada, 1991). CARBOFURAN 159
A maximum carbofuran concentration of 1 11g/litre was found in streams in the USA (Kimbrough & Litke, 1996). Levels found in groundwater in the USA ranged from 1 to 30 llg/litre (WHO, 1996a).
3.3 Food
Carbofuran does not bioaccumulate in food. Residues in treated crops are generally very low or not detectable (FAO/WHO, 1980; US EP A, 1987).
3.4 Estimated total exposure and relative contribution of drinking-water
Based upon the physical and chemical properties of carbofuran and the few data on occurrence, drinking-water from both groundwater and surface water sources has the potential of being the major route of exposure (US EP A, 1987).
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
Carbofuran is rapidly absorbed, metabolized, and eliminated, mainly in the urine, after oral administration to mice, rats, hens, and goats. After oral administration of C4CJphenyl carbofuran to rats, 92% of the radiolabel was eliminated in the urine and 3% in faeces. Most of the radiolabel was eliminated within 24 hours after treatment. With a [14C]carbonyl-labelled compound, 45% was 14 eliminated as C02• The metabolic pathway consists of hydroxylation, oxidation, hydrolysis, and conjugation (FAO/WHO, 1997).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
Carbofuran is highly toxic after acute oral administration. The oral LD 50 values in cats, rabbits, guinea-pigs, rats, mice, and dogs ranged from 3 to 19 mg/kg of body weight. Toxic signs observed were typical for cholinesterase inhibition: salivation, cramps, trembling, and sedation were observed within minutes after administration and lasted for up to 3 days (FAO/WHO, 1997). WHO (1996b) has classified carbofuran as "highly hazardous."
5.2 Short-term exposure
In a 13-week study, carbofuran (purity 99.6%) was fed to groups of beagle dogs (four per sex per group) at dietary concentrations of 0, 10, 70, or 250 mg/kg of feed (equal to 0, 0.43, 3.1, and 10.6 mg/kg of body weight per day); the highest dose was reduced from 500 mg/kg of feed because of marked toxicity. Hyperaemia, increased salivation, and inhibition of erythrocyte acetylcholinesterase activity were 160 PESTICIDES observed at the lowest dose. A NOAEL was not identified in this study (Bloch et al., 1987a; FAO/WHO, 1997). In a subsequent 4-week study, groups of male beagle dogs (four per group) were fed carbofuran (purity 99.6%) at 0 or 5 mg/kg of feed (equal to 0 and 0.22 mg/kg of body weight per day). Clinical signs, mortality, body weight, food consumption, and cholinesterase activity in plasma and erythrocytes were unaffected by treatment. The NOAEL in this study was 0.22 mg/kg of body weight per day, the only dose tested (Bloch et al., 1987b; FAO/WHO, 1997). In a 1-year study, groups of beagle dogs (six per sex per group) were fed carbofuran (96.1% purity) at dietary concentrations of 0, 10, 20, or 500 mg/kg of feed (equal to 0, 0.3, 0.6, and 13 mg/kg of body weight per day). Plasma cholinesterase inhibition was observed in most males at 10 and 20 mg/kg of feed and markedly (77%) in all animals at 500 mg/kg of feed. Erythrocyte and brain acetylcholinesterase were not inhibited at 10 or 20 mg/kg of feed. Histopathological testicular changes were observed at 500 mg/kg of feed and in a single male at 20 mg/kg of feed. The plasma cholinesterase inhibition was considered non significant, and attention was focused on the testicular changes. The NOAEL in this study was 10 mglkg offeed, equal to 0.3 mglkg of body weight per day (Taylor, 1983). The overall NOAEL in these short-term studies in dogs was 5 mg/kg of feed, equal to 0.22 mg/kg of body weight per day (FAO/WHO, 1997).
5.3 Long-term exposure
In a 2-year study, Charles River mice (100 per sex per group) were fed dietary carbofuran (purity 95.6%) concentrations of 0, 20, 125, or 500 mglkg offeed (equal to 0, 2.8, 18, and 70 mg/kg of body weight per day). Mice receiving the highest dose showed a decrease in body-weight gain. Cholinesterase activities were not measured in erythrocytes or plasma. At the two highest doses, a statistically significant depression of brain acetylcholinesterase activity was observed. The NOAEL was 20 mg/kg of feed, equal to 2.8 mglkg of body weight per day (Goldenthal, 1982a; FAO/WHO, 1997). Carbofuran (purity 95.6%) was fed to groups of Charles River rats (90 per sex per group) at concentrations ofO, 10, 20, or 100 mg/kg of feed for 2 years. Body weight gain and plasma, erythrocyte, and brain acetylcholinesterase activities were reduced at 100 mglkg of feed. The NOAEL was 20 mg/kg of feed, equivalent to 1 mglkg of body weight per day (Goldenthal, 1982b; FAO/WHO, 1997).
5.4 Reproductive and developmental toxicity
In a three-generation reproductive toxicity study, Charles River rats were fed carbofuran (purity 95.6%) at concentrations ofO, 20, or 100 mglkg of feed (equal to 0, 1.2, and 6 mg/kg of body weight per day for males and 0, 1.9, and 9.7 mg/kg of body weight per day for females). The NOAEL was 20 mg/kg of feed, equal to 1.2 mg/kg of body weight per day, on the basis of reductions in body-weight gain in the parental generation and reductions in the growth and survival of pup generations at 100 mg/kg of feed (Goldenthal, 1982c; FAO/WHO, 1997). CARBOFURAN 161
In an early study of developmental toxicity, Charles River rats (24 per group) were given carbofuran (purity 95.6%) by gavage at doses ofO, 0.1, 0.3, or 1 mg/kg of body weight per day on days 6-15 of gestation. Dose-dependent, transient clinical signs (chewing motions) were observed at all dose levels for a short period after treatment. Overt signs of toxicity in animals at 0.3 mg/kg of body weight per day included rough coats and lethargy. At the highest dose, lacrimation, increased salivation, trembling, and convulsions were also seen. A NOAEL was not identified in this study (Barron et al., 1978; FAO/WHO, 1997). In a later study, carbofuran (purity 95.6%) was given orally by gavage to groups of Charles River rats (25 per group) at doses of 0, 0.25, 0.5, or 1.2 mg/kg of body weight per day from day 6 to day 15 of gestation. The NOAEL for maternal and fetal toxicity was 1.2 mg/kg of body weight per day, the highest dose tested (Rodwell, 1980; FAO/WHO, 1997). In a further study of teratogenicity, groups of Charles River rats (40 females per group) were fed carbofuran (purity 95.6%) at dietary concentrations ofO, 20, 60, or 160 mg/kg of feed (equal to 0, 1.5, 4.4, and 11 mg/kg of body weight per day) on days 6-19 of gestation. The NOAEL for pup toxicity was 60 mg/kg of feed (equal to 4.4 mg/kg of body weight per day), based on reduced pup weight at 160 mg/kg offeed. The NOAEL for maternal toxicity was 20 mg/kg of feed (equal to 1.5 mg/kg of body weight per day), based on reduced body-weight gain at the two highest doses (Rodwell, 1981; FAO/WHO, 1997). None of these studies showed teratogenic potential. In an early teratogenicity study, New Zealand white rabbits (17 animals per group) were given carbofuran (purity 95.6%) by gavage at doses of 0, 0.2, 0.6. or 2 mg/kg of body weight per day on gestation days 6-18. Maternal toxicity was observed at 2 mg/kg of body weight per day and included trembling, salivation, sneezing, chewing motions, and reduced food and water consumption. The NOAEL in this study was 0.6 mg/kg of body weight per day (Rao, 1978; FAO/WHO, 1997). In a subsequent study, New Zealand white rabbits (20 animals per group) were given carbofuran (purity 95.6%) by gavage at doses ofO, 0.12, 0.5, or 2 mg/kg of body weight per day on days 6-18 of gestation. The NOAEL in this study was 0.5 mg/kg of body weight per day, on the basis of slightly reduced body-weight gain in dams and a slightly increased incidence of skeletal variations in pups at 2 mg/kg of body weight per day (Laveglia, 1981; FAO/WHO, 1997). These studies provided no evidence ofteratogenicity.
5.5 Neurotoxicity studies
In a 90-day study in Sprague-Dawley rats (1 0 per sex per group) at dietary carbofuran (purity 98.6%) concentrations ofO, 50, 500, or 1000 mg/kg of feed (equal to 0, 2.4, 27.3, and 55.3 mg/kg of body weight per day in males and 0, 3.1, 35.3, and 64.4 mg/kg of body weight per day in females), systemic toxicity (reduction in body weight gain) was observed at all doses. Clinical signs of neurotoxicity were observed at 500 and 1000 mg/kg of feed. No histopathological lesions were found in the nervous system. The NOAEL for neurotoxicity was thus 50 mg/kg of feed, equal to 162 PESTICIDES
2.4 mg/kg of body weight per day. There was no NOAEL for systemic toxicity (Freeman, 1994; FAO/WHO, 1997). In a study of developmental neurotoxicity, carbofuran (purity 98.1 %) was administered in the diet of female Sprague-Dawley rats (24 per group) at concentrations ofO, 20, 75, or 300 mg/kg of feed (equal to 0, 1.7, 5, and 11 mg/kg of body weight per day) on gestation day 6 through lactation day 10. Reductions in the body-weight gain of dams and pups and in pup survival and some evidence of delayed pup development were found at 75 and 300 mg/kg of feed. The NOAEL was 20 mg/kg of feed, equal to 1.7 mg/kg of body weight per day (Ponnock, 1994; FAO/WHO, 1997).
5.6 Mutagenicity and related end-points
Carbofuran has been tested for genotoxicity in a wide range of tests in vivo and in vitro. JMPR concluded that it was not genotoxic (FAO/WHO, 1997).
5. 7 Carcinogenicity
No evidence of tumorigenicity was found in the 2-year dietary studies on mice and rats described above.
6. EFFECTS ON HUMANS
Carbofuran poisoning was reported in three female farm workers who were not wearing any protective clothing and were throwing carbofuran granules in a coffee plantation in Jamaica. Signs of poisoning included vomiting, lassitude, nausea, and hypersalivation. Cholinesterase activity was not determined in these patients (FAO/WHO, 1997).
7. GUIDELINE VALUE
In the 1996 JMPR re-evaluation, an ADI of 0-0.002 mg/kg of body weight was determined based on a NOAEL of 0.22 mg/kg of body weight per day in a short-term ( 4-week) study of acute (reversible) effects in the dog, the most sensitive species, using an uncertainty factor of 100. This 4-week study was conducted as an adjunct to a 13-week study in which inhibition of erythrocyte acetylcholinesterase activity was observed at the lowest dose. Use of a 4-week study was considered appropriate because the NOAEL is based on a reversible acute effect. This NOAEL will also be protective for chronic effects (FAO/WHO, 1997). On the basis of the JMPR ADI (2.2 J.Lg/kg of body weight, if not rounded) and assuming a 60-kg body weight, drinking-water consumption of 2 litres/day, and an allocation of 10% of the ADI to drinking-water, a guideline value of 7 J.Lg/litre (rounded figure) can be calculated for carbofuran. CARBOFURAN 163
8. REFERENCES
Barron P, Giesler P, Ras GN (1978) Teratogemcity of carbofuran m rats. Unpublished report prepared by Warf Institute, Inc., Wl. Submitted to WHO by FMC Corp., Philadelphia, PA (Act No. 184.33). Bloch I et al. (1987a) 13-week oral toxicity feedzng study with carbofuran (D1221) zn the dog Unpublished report prepared by Research & Consulting Company AG, Itingen. Submitted to WHO by FMC Corp., Philadelphia, PA (RCC No. 077837; FMC Study No. A95-4249). Bloch I et al. (1987b) 4-week oral tox1city (feeding) study wtth carbofuran "D 1221" in male dogs. Unpublished report prepared by Research & Consulting Company AG, ltingen. Submitted to WHO by FMC Corp., Philadelphia, PA (RCC No. 087963; FMC Study No. A95-4248). Draper WM, Gibson RD, Street JC (1981) Drift from and transport subsequent to a commercial, aerial application of carbofuran: an estimation of potential human exposure. Bulletin of environmental contamination and toxicology, 26:537-543. FAO/WHO (1977) 1976 evaluations of some pesticide residues m food. Joint Meetmg of the FAO Panel of Experts on Pesticide Res1dues in Food and the Environment and the WHO Expert Group on Pesticide Residues Rome, Food and Agriculture Orgamzation of the United Nations (AGP: 1976/M/14). FAO/WHO (1980) Pest1c1de residues m food. 1979 evaluations Joint Meetmg of the FAO Panel of Experts on Pesticide Residues in Food and the Environment and the WHO Expert Group on Pesticide Res1dues. Rome, Food and Agriculture OrganizatiOn of the United Nations (FAO Plant Production and Protection Paper 20, Supplement). FAO/WHO (1985) Data Sheet on Pesticides No. 56- Carbofuran (VBC/PDS/DS/85.56). FAO/WHO (1997) Pesticide residues zn food- 1996 Jomt FAO/WHO Meeting on Pest1c1de Res1dues. Evaluations 1996. Part 11- Toxicological Geneva, World Health Organization, International Programme on Chemical Safety (WHO/PCS/97 .I). Freeman C (1994) Subchronic neurotoxicity screen in rats with carbofuran techmca/. Unpublished report prepared by FMC Corporation Toxicology Laboratory, NJ. Submitted to WHO by FMC Corp., Philadelphia, PA (FMC Study No. 92-3705). Goldenthal El (1982a) 2-year d1etary toxtcity and carcinogemcity study m mice Unpublished report prepared by International Research and Development Corporation, Ml. Submitted to WHO by FMC Corp., Philadelphia, PA (FMC Study No. Act No. 150.52, IRDC No. 167-116). Goldenthal El (1982b) 2-year dietary toxicity and carcmogenicity study m rats. Unpublished report prepared by International Research and Development Corporation, Ml. Submitted to WHO by FMC Corp., Philadelphia, PA (FMC Study No. Act No. 130.51; IRDC No. 167-115). Goldenthal El (1982c) Three generatton reproduction study m rats. Unpublished report prepared by International Research and Development Corporation, MI. Submitted to WHO by FMC Corp., Philadelphia, PA (FMC Study No. Act No. 131.53; IRDC No. 167-114). Health and Welfare Canada ( 1991) Guidelines for Canadian drinkmg water quality. Ottawa, Ontario. Kimbrough RA, Litke DW ( 1996) Pesticides in streams draining agricultural and urban areas in Colorado. Environmental science and technology, 30:908-916. Laveglia J ( 1981) Teratology study m the rabbit wtth carboforan Unpublished report prepared by International Research and Development Corporation, MI. Submitted to WHO by FMC Corp., Philadelphia, PA (FMC Study No. A 80-452; IRDC No. 167-156). Ponnock KS (1994) A developmental neurotox1ctty study of carboforan in the rat vta d1etary admimstration. Unpublished report prepared by Pharmaco LSR, Inc., Toxicology Services North America, NJ. Submitted to WHO by FMC Corp., Philadelphia, PA (Study No. 93- 4506; FMC Study No. A 93-3746). Rao GN (1978) Teratology study of carbofuran in rabbits. Unpublished report prepared by Warf Institute, Inc., WI. Submitted to WHO by FMC Corp., Philadelphia, PA (Study No. 185-33). Rodwell DE (1980) Teratology study m the rat with carbofuran. Unpublished report prepared by International Research and Development Corporation, MI. Submitted to WHO by FMC Corp., Philadelphia, PA (FMC Study No. A 80-453; IRDC No. 167-155). 164 PESTICIDES
Rodwell DE (I 981, with amendment) Teratology and postnatal study m the rat wzth carbofuran. Unpublished report prepared by International Research and Development Corporation, Ml. Submitted to WHO by FMC Corp., Philadelphia, PA (FMC Study No. A 80-444; IRDC No. 167-154), Taylor GD (I 983) One-year chronic oral toxzcity study in beagle dogs wzth carbofuran. Unpublished report prepared by ToxiGenics, Inc., IL. Submitted to WHO by FMC Corp., Philadelphia, PA (FMC Study No. A 81-605; Toxicogenics No. 410-0715). US EPA (1987) Carbofuran. Health advisory. Washington, DC, US Environmental Protection Agency, Office of Drinking Water. WHO (1996a) Guidelines for drinking-water quality, 2nd ed Vol. 2. Health criteria and other supporting information Geneva, World Health Organization. WHO (1996b) The WHO recommended classification of peslzczdes by ha=ards and guzdelines to classification. Geneva, World Health Organization, International Programme on Chemical Safety (WHO/PCS/96.3). CYANAZINE
First draft prepared by A.M. Mahfouz Office of Water, Health and Ecological Criteria Division United States Environmental Protection Agency, Washington, DC, USA
Cyanazine has not previously been evaluated in the WHO Guidelines for drinking-water quality or by JMPR. The Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality considered cyanazine to be of high priority because it has often been found in water.
1. GENERAL DESCRIPTION
1.1 Identity
CAS no: 21725-46-2
Molecular formula: C9H13ClN6
Cyanazine is a member of the triazine family of herbicides. The IUP AC name for cyanazine is 2-( 4-chloro-6-ethylamino-1 ,3,5-triazin-2-yl)amino-2-methyl propionitrile.
1.2 Physicochemical properties1
Property Value Physical state White solid at 25°C Melting point 167.5-169°C Density 0.35 (fluffed), 0.45 (packed) g/cm3 Vapour pressure 2.1 x 10-7 to 10.0 x 10-7 Pa at 20°C Water solubility 171 mg/litre at 25°C Log octanol-water partition coefficient 2.24
1.3 Major uses
Cyanazine is used as a pre- and post-emergence herbicide for the control of annual grasses and broadleaf weeds (Meister, 1983).
1.4 Environmental/ate
Cyanazine can be degraded in soil and water by microorganisms by N-dealkylation and hydrolysis. Cyanazine readily leaches from soil. In laboratory tests, cyanazine (5-10 mg/litre) had a half-life of 2-14 weeks in four kinds of soils at
1 Source: CHEMLAB's Chemical Information System. Bethesda, MD, CIS, Inc. (1985). 166 PESTICIDES
22°C (Osgerby et al., 1968). Four degradation products can be identified for cyanazine - the amide, two acids, and the amine. Aerobically and anaerobically aged cyanazine residues, primarily the amide degradation product, are intermediately mobile to mobile on sandy clay loam soil (Eadsforth, 1984). The amide degradation product is predominant in the leachate from sandy soil; the acid degradation products predominate in leachate from loamy sand and sandy loam soils. Unaltered cyanazine can also be identified in soilleachate.
2. ANALYTICAL METHODS
Cyanazine in water samples can be analysed using a high-performance liquid chromatographic (HPLC) method (Method #4; US EPA, 1986b). In this method, 1litre of sample is extracted with methylene chloride using a separatory funnel. The methylene chloride extract is dried and concentrated to a volume of 10 ml or less. HPLC is used to separate compounds, and measurement is conducted with an ultraviolet detector. Using this method, the estimated detection limit for cyanazine is 0.3 J..Lg/litre.
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Water
In the USA, cyanazine was detected in surface water and groundwater at maximum concentrations of 1300 and 3500 J..lg/litre, respectively. Cyanazine was identified in drinking-water in New Orleans, Louisiana, at concentrations ranging from 0.01 to 0.35 J..Lg/litre. 2 Cyanazine was also found in surface water in Ohio River basins (Datta, 1984) and in groundwater in Iowa and Pennsylvania; typical detectable concentrations ranged from 0.1 to 1.0 J..Lg/litre (Cohen et al., 1986). Monitoring data in a reservoir on the Des Moines River in Iowa from September 1977 through November 1978 indicated that agricultural runoff (from corn and soybean) was a major source of cyanazine in the river: levels of 71-457 and 2-151 ng/litre were detected during the active months of May through August and during September through December, respectively; cyanazine was not detected from January through April (NAS, 1977; US EPA, 1984a). Cyanazine was detected (detection limits 0.025-1 J..lg/litre) in 9 of 1128 samples of municipal and private water supplies in three Canadian provinces between 1978 and 1986; concentrations ranged from <0.1 to 4.0 J..lg/litre (Hiebsch, 1988). It was also detected (detection limit 0.02 J..Lg/litre) in 45 of 440 surface water samples from three Ontario, Canada, river basins surveyed between 1981 and 1985; mean detectable concentrations were 0.8, 0.1, and 1.5 J..lg/litre in the three basins (Frank & Logan, 1988). Cyanazine has been found in groundwater in the Netherlands at concentrations above 0.1 J..Lgllitre. It was not detected in surface water used as a source for drinking-water (Council of Europe, 1993).
Source: STORET water quality file. Washington, DC, US Environmental Protection Agency, Office of Water (data file search conducted in May 1988). CYANAZINE 167
3.2 Food
Data on levels of cyanazine residues in food are not available. It is expected that the intake of cyanazine from food is very low, because no residues of cyanazine or its degradation products have been detected in crops following application (Department ofNational Health and Welfare, 1986).
3.3 Estimated total exposure and relative contribution of drinking-water
Based on Canadian monitoring data and the data from the Netherlands, the daily intake of cyanazine from drinking-water can be estimated to fall in the range of0.2-3 Jlg. As there is no information available on concentrations of cyanazine in food or air, the relative contribution of drinking-water to total daily intake cannot be estimated.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
Cyanazine is rapidly absorbed from the gastrointestinal tract of rats, dogs, and cows (Shell Chemical Company, 1969; Hutson et al., 1970; Cmyford & Hutson, 1972). Measurements of urinary, faecal, and biliary excretion indicated that 80-88% of the administered dose was eliminated within 4 days in rats and dogs and within 21 days in cows. In rats, elimination in urine was almost equal to elimination in faeces. In dogs and cows, approximately one-half of the dose was eliminated in the urine, and about one-third was eliminated in the faeces. In cows, the amount of residues excreted daily was constant throughout the study period. Cyanazine was also detected in cows' milk. The degradation of cyanazine follows metabolic pathways involving dealkylation and conjugation with glutathione. Seven metabolites were identified in rats: five in urine and two in faeces. N-De-ethylation was the major route of degradation of cyanazine; 47% of the leaving ethyl group was eliminated by exhalation (Shell Chemical Company, 1969; Crayford & Hutson, 1972). Crayford et al. (1970) studied the metabolism of two of the major plant metabolites of cyanazine - 2-hydroxy-4-ethylamino-6-(1-carboxy-1-methylamino ) s-triazine and 2-hydroxy-4-amino-6-(1-carboxy-1-methylethylamino )-s-triazine - in rats. Approximately 91% of the first compound and 84% of the second compound were recovered unchanged from urine and faeces.
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
WHO (1996) has classified cyanazine as "moderately hazardous." Acute oral
LD50s in rats ranged from 149 to 835 mg/kg of body weight (SRI, 1967; Young & Adamik, 1979b; Meister, 1983; NIOSH, 1987). In these studies, the percentage of active ingredient (a.i.) in the tested products was not clearly identified. However, 168 PESTICIDES
studies with technical cyanazine (97% a.i.) in rats, mice, and rabbits showed LD50s of 182, 380, and 141 mg/kg of body weight, respectively (Walker et al., 1974). Cyanazine caused mild eye irritation at 100 mg (Young & Adamik, 1979a) and slight skin irritation at 2000 mg (Young & Adamik, 1979c) in rabbits. A skin sensitization test in guinea-pigs was negative (Walker et al., 1974; Young & Adamik, 1979d).
The acute dermal LD50 in rabbits treated with technical cyanazine (purity unspecified) was >2000 mg/kg of body weight (SRI, 1967; Young & Adamik, 1979c); in rats, the LD50 was> 1200 mg/kg of body weight (97% a.i.) (Walker et al., 1974).
5.2 Short-term exposure
In a 13-week feeding study in mice at levels of 0, 10, 50, 500, 1000, or 1500 mg/kg of feed (equivalent to 0, 1.3, 6.5, 65, 130, and 195 mg/kg of body weight per day; US EP A, 1986a), reduction in body-weight gain and statistically significant increases in relative liver weights were observed in both sexes at 65 mg/kg of body weight per day and above. The NOAEL was 6.5 mg/kg of body weight per day (Fish et al., 1979). In a study by Walker et al. ( 1968), groups of 10 female CFE rats were treated by gavage with single oral doses of 1, 5, or 25 mg/kg· of body weight of a wettable powder formulation (75% a.i.); the control group received water. No diuretic effects were produced; however, serum protein and potassium concentrations increased at the high dose, and serum osmolality increased at 5 mg/kg of body weight. The NOAEL appeared to be 1 mg/kg of body weight; however, this study did not provide enough information to allow the presence or absence of more significant effects at this dosage level to be determined. In a 4-week oral toxicity study, groups of 10 male and 10 female CFE rats received diets containing cyanazine (75% or 97% a.i.) at 1, 10, or 100 mg/kg of feed (equivalent to 0.05, 0.5, and 5 mg/kg of body weight per day; US EPA, 1986a). A control group of 20 animals per sex was used. Reductions in body weight and food intake were noted at the high dose level. The LOAEL appeared to be 0.05 mg/kg of body weight per day based on kidney function tests, although additional information was not available to determine if any other significant adverse effects were noted at this level. In 13-week feeding studies in rats given cyanazine doses of 0, 1, 1.5, 3, 6, 12, 25, 50, or 100 mg/kg of feed (equivalent to 0, 0.05, 0.075, 0.15, 0.30, 0.60. 1.25, 2.5, and 5 mg/kg of body weight per day), decreased body-weight gain was noted in males at 0.075 mg/kg of body weight per day and above and in females at 2.5 mg/kg of body weight per day and above. The NOAEL ranged from 0.05 to 1.25 mg/kg of body weight per day (Walker & Stevenson, 1968b). In a 13-week oral study, beagle dogs were given cyanazine in gelatin capsules at 0, 1.5, 5, or 15 mg/kg of body weight per day. Emesis was noted within the first hour of dosing in all of the high-dose males. Reduced body-weight gain was also noted in the high-dose group during the second half of the study period, as well as increased kidney and liver weights in the females of this group. The NOAEL was 5 mg/kg of body weight per day (Walker & Stevenson, 1968a; Walker et al., 1974). CYANAZINE 169
5.3 Long-term exposure
In a 2-year study in mice given cyanazine (96.4% pure) at doses ofO, 10, 25, 250, or 1000 mg/kg of feed (equivalent to 0, 1.3, 3.3, 33, and 130 mg/kg of body weight per day), survival ranged from 46 to 58% in males and from 38 to 50% in females; survival was slightly reduced in females but not in males at the two highest doses. Mean body-weight gains for the entire study were reduced 4% from controls in both sexes at 25 mg/kg of feed and 18% and 25% in males and females, respectively, receiving 250 mg/kg of feed. At 1000 mg/kg, weight gains were reduced 25% and 32% in males and females, respectively; the MTD was exceeded at this dose. Non-neoplastic histological findings at 250 mg/kg of feed and above included increased incidences of parenchyma! atrophy of the liver in females, kidney toxicity, and skin ulceration. Based on decreased weight gains and histological changes, the NOAEL for systemic toxicity was 25 mg/kg of feed (3.3 mg/kg of body weight per day) (Gellatly, 1981). In a 2-year study, Sprague-Dawley rats (50 per sex per dose) were exposed to cyanazine (96% pure) at 0, 1, 5, 25, or 50 mg/kg of feed (Bogdanfy, 1990). An additional 10 animals per sex per dose were used as satellite animals for biochemical testing, then sacrificed after 12 months. The mean daily intakes were 0, 0.040, 0.198, 0.985, or 2.06 mg/kg of body weight for the males and 0, 0.053, 0.259, 1.37, or 2.81 mg/kg of body weight for the females. No adverse effects on survival were observed. A significant increase in the incidence of hyperactivity was observed in males receiving 25 mg/kg of feed (40%) and 50 mg/kg of feed (58%) compared with controls (20%), but no hyperactivity was observed in females. The incidence of palpable masses was significantly (p < 0.05) increased for females at the highest dose (51% vs 38% for controls), and the median time to first observed mass was decreased (343 days) compared with controls (406 days). Mean body weights and body-weight gains were significantly depressed during the first year of the study in males receiving 50 mg/kg of feed (up to 18%) and females receiving 25 or 50 mg/kg of feed (about 20%). A significant (p < 0.05) increase in the incidence of extramedullary haematopoiesis of the spleen was observed in males in the highest dose group (56% vs 39% for controls), and there was a significant trend (p < 0.02) for granulocytic hyperplasia of the bone marrow (p = 0.05). In highest-dose females, sciatic nerve demyelinization was increased (18% vs 8% for controls), with a positive dose trend (p = 0.0 13). The NOAEL was 5 mg/kg of feed (0.198 mg/kg of body weight per day for males; 0.259 mg/kg of body weight per day for females) based on hyperactivity in male rats and decreased body-weight gain in females. In a 2-year study in beagle dogs with technical cyanazine (97% a.i. in gelatin capsules) at dose levels of 0, 0.625, 1.25, or 5 mg/kg of body weight per day, frequent emesis was noted within 1 hour of dosing in the high-dose group; this effect was associated with reduction of growth rate and serum protein. The NOAEL appeared to be 1.25 mg/kg of body weight per day (Walker et al., 1970). Because of inadequate histopathology and data reporting in this study, a 1-year feeding study in beagle dogs was later performed by Dickie (1986) at cyanazine (98% pure) levels of 0, 10, 25, 100, or 200 mg/kg of feed (equal to 0, 0.27, 0.68, 3.20, and 6.11 mg/kg of body weight per day for males and 0, 0.28, 0.72, 3.02, and 6.39 mg/kg of body 170 PESTICIDES weight per day for females). No systemic toxicity was noted at 10 or 25 mg/kg of feed. At 100 and 200 mg/kg of feed, dose-related decreases in body weight and body-weight gains were observed, platelet counts were occasionally elevated (non significantly), liver-to-body-weight ratios were slightly increased, and kidney-to body-weight ratios were slightly increased (p < 0.05) in females. Decreased serum levels of total protein albumin and calcium were consistently noted in both sexes at 200 mg/kg of feed. The average NOAEL for systemic toxicity in males and females was 0.7 mg/kg of body weight per day.
5.4 Reproductive and developmental toxicity
No significant effects on reproductive parameters were found in a three generation study in Long-Evans rats using technical cyanazine (unknown percentage a.i.) at levels of 0, 3, 9, 27, or 81 mg/kg of feed (equivalent to 0, 0.15, 0.45, 1.35, and 4.05 mg/kg of body weight per day; US EP A, 1986a). The NOAEL in this study appeared to be 1.35 mg/kg of body weight per day based on reduced body-weight gain in parental animals and increased brain weight and decreased relative kidney weight in F 3b female weanlings (Eisenlord et al., 1969). In a two-generation reproductive study in Sprague-Dawley rats, cyanazine (100% a.i.) was administered at 0, 25, 75, 150, or 250 mg/kg of feed (equal to intake in dams of 0, 1.8, 5.3, 11.1, and 18.5 mg/kg of body weight per day; intake changed during lactation to 0, 3.8, 11.2, 23.0, or 37.1 mg/kg body weight per day). There were no compound-related effects on the number of females producing litters or on litter size. Based on decreased pup viability and decreased mean pup body weight during lactation, the NOAEL for reproductive toxicity was 3.8 mg/kg of body weight per day. Dose-related parental toxicity was observed at the lowest dose tested, as body weights ofF0 and F 1 adults of both sexes decreased throughout the study period (WIL Research Laboratories, 1987; Dapson, 1990). In a study by Lu et al. (1981, 1982), Fischer 344 rats (30 dams per group) were administered cyanazine (98.5% a.i.) by gavage (suspended in a 0.2% Methocel emulsion) at dose levels of 0, 1.0, 2.5, 10, or 25 mg/kg of body weight per day on gestation days 6-15. Maternal body-weight reductions were noted at 10 and 25 mg/kg of body weight per day. Diaphragmatic hernia associated with liver protrusion, microphthalmia, and anophthalmia were observed at 25 mg/kg of body weight per day. Lochry et al. (1985) repeated the above study, administering cyanazine (98% a.i.) to dams (70 per dose group) by gavage in an aqueous suspension of 0.25% (w/v) methyl cellulose at dose levels ofO, 5, 25, or 75 mg/kg of body weight per day on days 6-15 of gestation. Maternal body-weight reductions were noted in all dosage groups and appeared to be partly associated with lower food intake during the dosing period. Alteration in skeletal ossification sites was observed in the fetuses at all dose levels. Teratogenic effects - anophthalmia/microphthalmia, dilated brain ventricles and cleft palate in the fetuses, arid abnormalities of the diaphragm (associated with liver protrusion) in pups sacrificed at time of weaning - were demonstrated at 25 and 75 mg/kg of body weight per day. Maternal and developmental toxicity were CYANAZINE 171 observed at 5 mg/kg of body weight per day (lowest dose tested), and the NOAEL for teratogenic effects was 5 mg/kg of body weight per day (Bui, 1985). In an additional study in Sprague-Dawley rats, oral administration of cyanazine at 30 mg/kg of body weight per day (the highest dose tested) resulted in maternal body-weight reductions and increased incidence of piloerection; no developmental toxicity was observed. The maternal systemic NOAEL was 3 mg/kg of body weight per day (Shell Chemical Company, 1983). New Zealand white rabbits (22 dams per dose) were orally dosed with cyanazine (98% a.i.) in gelatin capsules at levels of 0, 1, 2, or 4 mg/kg of body weight per day on gestation days 6-18. At 2 and 4 mg/kg of body weight per day, maternal toxic effects included anorexia, weight loss, death, and abortion; alterations in skeletal ossification sites, decreased litter size, and increased post-implantation loss were also observed. Malformations at 4 mg/kg of body weight per day included anophthalmia/microphthalmia, dilated brain ventricles, domed cranium, and thoracoschisis; however, these responses were observed at levels in excess of maternal toxicity. The NOAEL for both maternal and developmental toxicity was 1 mg/kg of body weight per day (Shell Toxicology Laboratory [Tunstall], 1982).
5.5 Mutagenicity and related end-points
The mutagenicity studies for cyanazine provide equivocal evidence for genotoxicity. Cyanazine induced dose-related forward mutation with and without metabolic activation in the mouse lymphoma L5178Y/TK cell gene mutation assay (Jannasch & Sawin, 1986). Cyanazine was positive for in vitro unscheduled DNA synthesis in repeat assays using rat primary hepatocytes (Vincent, 1987). Cyanazine was negative in an in vivo unscheduled DNA synthesis assay in rat spermatocytes to examine possible interaction with germ cells, a Salmonella assay, a Chinese hamster ovary/hprt gene mutation assay, and an in vitro human lymphocyte/aberrations assay (Stahl, 1987). Cyanazine was found to be negative in a Salmonella assay with rodent metabolic activation, but positive with a plant-derived activation system (Plewa eta!., 1984).
5. 6 Carcinogenicity
Cyanazine was not carcinogenic in mice (Gellatly, 1981; Shell Chemical Company, 1981). However, in Sprague-Dawley rats, dietary administration ofO, 1, 5, 25, or 50 mg/kg of feed for 2 years caused statistically significant increases in malignant mammary gland tumours (adenocarcinoma and carcinosarcoma) in females at incidences of 5/58 (8%), 7/61 (11 %), 12/60 (20%), 20/62 (32%), and 15/62 (24%), respectively. with a significant positive trend (p = 0.0049). Statistical comparison excluded rats that died before week 48 when the first tumour occurred. The incidences of these tumours in dosed rats were outside the historical control data. Atrazine, which has a chemical structure similar to that of cyanazine, has been found to increase the incidence of mammary tumours in rats and has been classified by IARC (1991) in Group 2B (agent is possibly carcinogenic to humans). The hypothesis that a hormonal mechanism of action may be involved in the 172 PESTICIDES manifestation of mammary gland tumours m rats upon exposure to triazine herbicides is currently under investigation.
6. EFFECTS ON HUMANS
No information was found in the available literature on the health effects of cyanazine in humans.
7. GUIDELINE VALUE
On the basis of the available mutagenicity data on cyanazine, evidence for genotoxicity is equivocal. Cyanazine causes mammary gland tumours in Sprague Dawley rats but not in mice. The mechanism of mammary gland tumour development in Sprague-Dawley rats is currently under investigation, and a hormonal mechanism of action may be involved. Cyanazine is also teratogenic in Fischer 344 rats at dose levels of 25 mg/kg of body weight per day and higher. Based on a 2-year toxicity/carcinogenicity study in rats (Bogdanfy, 1990), a NOAEL of 0.198 mglkg of body weight per day has been identified, based on hyperactivity in male rats. By applying an uncertainty factor of 1000 (1 00 for inter and intraspecies variation and 10 for limited evidence of carcinogenicity), a TDI of 0.198 f.!g/kg of body weight can be calculated. With an allocation of 10% of the TDI to drinking-water and assuming a 60-kg adult consuming 2 litres of drinking-water per day, the guideline value is 0.6 f.!g/litre (rounded figure).
8. REFERENCES
Bogdanfy MS (1990) Combmed chrome tox1c1tyloncogemcity study wtth cyanazme m rats Unpublished report prepared by Haskell Laboratory and submitted to the US Envtronmental Protection Agency by E.I. duPont de Nemours Company (Study No. 23-90, 11 May 1990; MRID No. 415099-02). Bui QQ (1985) Review of a developmental tox1c1ty study (teratology and postnatal study). Internal memorandum to Robert J. Taylor [reviewmg study cited m Shell Development Company (1985) Report No. 619-002, Accession No. 257867], US Environmental Protection Agency. Cohen SZ, Eiden C, Lorber MN (1986) Monitonng ground water for pesticides m the U S.A. In: Evaluatwn of pesticides in ground water, pp. 170-196 (American Chemical Society Symposium Series 315). Council of Europe (1993) Pesticides and ground water Strasbourg, Council of Europe Press Crayford JV, Hutson DH (1972) Metabolism of the herbicide 2-chloro-4-(ethylamino)-6-(1-cyano-1- methylethylamino)-s-triazine in the rat [MRID No. 00022856] PesttCLde bwchenustry and physiology, 2:295-307. Crayford JV, Hoadley EC, Pikering BA (1970) The metabolism of the maJor plant metabohtes of Bladex (DW -1385 and DW -139-1) m the rat Group research report TLGR 0081 70 Unpublished study prepared by Shell Research, Ltd (MRID No. 000223871) Dapson SC (1990) Review of add!lional data submitted in support of a multigeneration study m rats with Bladex Memorandum to R. Taylor Barnes, Office of Pesttctde Programs, US Environmental Protection Agency, 7 March 1990 (MRID No. 411111 0-04; HED No. 9-636). Datta PR (I 984) Rev1ew of six documents regardmg momtonng of pesticides 111 northwestern Oh to rivers. Internal memorandum, US Environmental Protection Agency, Washington, DC. Department ofNat10nal Health and Welfare (1986) Natwnal pest1c1de res1due linuts 111 foods Ottawa, Ontario, Department of National Health and Welfare, Food Directorate. CYANAZINE 173
Dickie BC (1986) One-year oral feeding study 111 dogs with the tnazine herbic1de cyanazine Unpublished study performed by Hazleton Laboratones for E.I. duPont de Nemours and Company (Study No. 6160-104 and Addendum No. 107F; MRID Nos 40081901 and 40229001). Eadsforth CV (1984) The leaching behavwr of Bladex and its degradatwn products in German sods under laboratory condrtions. Expenment No. 2994. Unpublished study submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC. Eisenlord G, Loquvam GS, Leung S (1969) Results ofreproductwn study of rats fed d1ets contammg SD 15418 over three generatwns Report No 47 Unpublished study received on unknown date under 9G0844; prepared by Hine Laboratories, Inc.; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:095023-D; MRID No. 00032346). Fish A, Hend R W, Clay CE (1979) Toxicity studies on the herbicide Bladex a three-month feeding study in mice: TLGR 0021.79. Unpublished study received 19 July 1979 under 20 1-279; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:09835-C). Frank R, Logan L (1988) Pesticide and industrial chemical residues at the mouth of the Grand, Saugeen and Thames rivers, Ontario, Canada, 1981-85. Archives of envrronmental contammation and toxicology, 17:741. Gellatly J ( 1981) A two-year feeding study of Bladex 111 the mouse Unpublished report prepared by Shell Toxicology Laboratory; submitted to the US Environmental Protection Agency by Shell Chemical Company (Report No. 1493, December 1981; EPA Accession Nos. 247295 through 298). Hiebsch SC ( 1988) The occurrence of thirty-five pestic1des m Canad1an dnnking water and surface water Unpublished report prepared for the Environmental Health Directorate, Department of National Health and Welfare, Ottawa, Ontario. Hutson DH et al. (1970) Mercapturic acid formation in the metabolism of 2-chloro-4-ethylamino-6- (1-methyl-1-cyanoethylamino )-s-triazine in the rat. Journal of agncultural and food chemistry, 18:507-512. !ARC (1991) Atrazine. In: Occupational exposures in msect1cide applicatwn, and some pest1c1des. Lyon, International Agency for Research on Cancer, pp. 441-446 (!ARC Monographs on the Evaluation ofCarcmogenic Risks to Humans, Vol 53). Jannasch MG, Sawin VL (1986) Genetic toxicity assay of Bladex herb1crde· gene mutatwn assay in mammalian cells in culture, L5178Y mouse lymphoma cells Unpublished study No. 61287 (12 August 1986), performed by Westhollow Research Center, Houston, TX, for Shell, Houston, TX (MRID No. 00165051). Lochry EA, Hoberman AM, Christian MS ( 1985) Study on the developmental tox1c1ty of techmcal Bladex herb1c1de (SD 15418) 111 F1scher 344 rats. Unpublished report submitted to the US Environmental Protection Agency by Shell Oil Company; prepared by Argus Research Laboratory, Inc., Horsham, PA (Report No. 619-002, 18 Aprill985). Lu CC, Tang BS, Chai EY ( 1981) Techmcal Bladex (SD 1541 8) teratology study 111 rats. Project No. 61230. Unpublished study received 4 January 1982 under 201-179; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:070584-A; MRID No. 00091020). Lu CC, Tank BS, Chai EY (1982) Teratogenicity evaluations of technical Bladex in Fischer 344 rats. Teratology, 25(2):59A-60A. Meister R, ed. (1983) Farm chem1cals handbook. Willoughby, OH, Meister Publishing Company. NAS (1977) Drinking water and health Washington, DC, National Academy of Sciences, National Academy Press. NIOSH (1987) Registry of toxic effects of chemrcal substances Rockville, MD, US Department of Health, Education and Welfare, National Institute for Occupational Safety and Health. Osgerby JM, Clarke DF, Woodbum AT (1968) The decomposition and absorptwn of DW 3418 (WL 19,805) in soils. Unpublished study submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC. 174 PESTICIDES
Plewa MJ et al. ( 1984) An evaluation of the genotoxic properties of herbicides following plant and animal activation. Mutation research, 136:233-245. Shell Chemical Company ( 1969) Metabolism of cyanazme Unpublished study submitted to the US Environmental Protection Agency by Shell Chemical Company (MRID No. 00032348). Shell Chemical Company (1981) Two-year oncogemcity study m the mouse. Unpublished report submitted under Pesticide Petition No. 9F2232 (EPA Accession Nos. 247295 to 247298) Shell Chemical Company (1983) Teratogenic evaluation of Bladex in SD CD rats Unpublished report submitted by Shell Development Company; prepared by Research Triangle Institute (Project No. 31 T-2564; 16 May 1983); submitted to the US Environmental Protection Agency on 6 July 1983 (EPA Accession No. 071738). Shell Toxicology Laboratory (Tunstall) (1982) A teratology study m New Zealand whue rabbits given Bladex orally. Unpublished report prepared by Sittingbourne Research Centre, England (Project No. 221/81, Experiment No. AHB-2321, November 1982). Submitted to the US Environmental Protection Agency on I February 1983, as document SBGR 82.357, by Shell Oil Company, Washington, DC, under Accession No. 071382. SRI (1967)Acute oral tox1c1ty ofSD 15418 (technical cyanazine). Stanford Research Institute Project 55 868 (Report No. 43, 26 May 1967). Unpublished report submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (Pesticide Petition No. 9G0844, Accession No. 91460). Stahl R (1987) In vitro evaluation of cyana=ine (INR-1957) for chromosome aberratwns m human lymphocytes: Haskell Laboratory Report No 328-87· MR 4581-460 Unpublished study prepared by E.l. duPont de Nemours and Company, Inc., 23 pp. (MRID No 40304705). US EPA (1984a) Draft health and environmental effects profile for cyanazine Cincinnati, OH, US Environmental Protection Agency, Environmental Criteria and Assessment Office. US EPA (1984b) Cyana=ine toxicology data review for registration standard. Washington, DC, US Environmental Protection Agency, Office of Pesticide Programs. US EPA (1985) Code offederal regulations 40 CFR 180.307. Washington, DC, US Environmental Protection Agency. US EPA (1986a) Reference values for nsk assessment. Prepared by the Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office, US Environmental Protection Agency, Cincinnati, OH, for the Office of Solid Waste, US Environmental Protection Agency, Washington, DC. US EPA (1986b) US. EPA Method No. 4: Determination ofpesticides in ground water by HPLCIUV (January 1986 draft). Available from Environmental Protection Monitoring and Support Laboratory, US Environmental Protection Agency, Cincinnati, OH. Vincent OR (1987) Assessment of cyanazine in the m VItro unscheduled DNA synthesis assay in rat primary hepatocytes Unpublished study No. HLR-347-87, performed by Haskell Laboratory for E.!. duPont de Nemours and Company, Wilmington, DE (MRID No. 403047-02; 14 July 1987). Walker AIT, Stevenson DE (1968a) The toxicity of the s-triazine herbicide (DW 34/8): 13-week oral experiment in dogs: Group research report TLGR.0016.68 Unpublished study received 17 October 1969 under 9G0844; prepared by Shell Research, Ltd., England; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:091460-G; MRID No. 00093199). Walker AIT, Stevenson DE (!968b) The toxicity of the s-tnazme herbicide (DW 34/8). 13-week oral experiment m rats: Group research report TLGR. 0017.68. Unpublished study received 17 October 1969 under 9G0844; prepared by Shell Research, Ltd., England; submitted by Shell Chemical Company, Washington, DC (MRID No. 00093198). Walker AIT, Kampjes R, Hunter GG (1968) Toxicity stud1es in rats on the s-triazine herbicide (DW 3418): (a) 13-week oral expenments; and (b) The effect on kidney function. Group research report TLGR. 0007.69. Unpublished study received 17 October 1969 under 9G0844; prepared by Shell Research, Ltd., England; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:091460-H; MRID No. 00093200). CYANAZINE 175
Walker AIT, Thorpe E, Hunter CG (1970) Toxicity studies on the s-trwzme herbtctde Blade (DW 3418). Two-year oral experiment with dogs: Group research report TLGR 0065 70 Unpublished study received 4 December 1970 under OF0998; prepared by Shell Research, Ltd .. England; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:091724-R; MRID No. 00065483). Walker AIT et al. (1974) Toxicological studies with the I ,3,5-triazine herbicide cyanazine. Pesticide sctence, 5(2): 153-159. WHO (1996) The WHO recommended classification of pesticides by hazards and gutdelines to classification. Geneva, World Health Organization, International Programme on Chemical Safety (WHO/PCS/96.3). WIL Research Laboratories (1987) Two-generation rat reproduction study. WIL No. 93001, 12 August 1987. Unpublished study submitted to the US Environmental Protection Agency by E. I. duPont de Nemours and Company (MRID No. 403600-01). Young SM, Adamik ER (1979a) Acute eye tmtation study in rabbits wzth SD 15418 (technical Bladex herbtcide) Code 16-8-0-0. Project No. WIL-1223-78. Unpublished study received 10 January 1980 under 201-281; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:099198-E; MRID No. 00026427). Young SM, Adamik ER (1979b) Acute oral toxicity study in rats with SD 15418 (techmcal Bladex herbicide): Code 16-8-0-0: Project No W1L-1223-78. Unpublished study received 10 January 1980 under 201-281; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:099198-C; MRID No. 00026424). Young SM, Adamik ER (1979c) Acute dermal toxicity study 111 rabbits with SD 15418 (techmca/ Bladex herbicide): Code 16-8-0-0: Project No. WIL-1223-78 Unpublished study received 10 January 1980 under 201-281; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:099!98-C; MRID No. 00026425). Young SM, Adamik ER (1979d) Delayed contact in hypersensitivity study m guinea pigs with SD 15418 (technical Bladex herbicide): Code 16-8-0-0: Project No. W1L-1223-78 Unpublished study received 10 January 1980 under 201-281; submitted to the US Environmental Protection Agency by Shell Chemical Company, Washington, DC (CDL:099198-F; MRID No. 00026428).
1,2-DIBROMOETHANE
First draft prepared by K. Hughes Bureau of Chemical Hazards, Environmental Health Directorate Health Protection Branch, Health Canada, Ottawa, Ontario, Canada
A guideline value for 1,2-dibromoethane was not derived in the 1993 WHO Guidelines for drinking-water quality because, although I ,2-dibromoethane appears to be a genotoxic carcinogen, the studies available at that time were inadequate for mathematical risk extrapolation. It was recommended that I ,2-dibromoethane be re-evaluated as soon as new data became available. 1,2-Dibromoethane was evaluated by JMP in 1994, and an IPCS Environmental Health Criteria monograph was published in 1996. It was considered desirable to determine if these new evaluations contained data suitable for the quantitative assessment of the carcinogenic risk of 1,2-dibromoethane. The Coordinating Committee for the updating of the WHO Guidelines therefore recommended that 1,2-dibromoethane be evaluated for the 1998 Addendum.
I. GENERAL DESCRIPTION
1.1 Identity
CAS no.: I 06-93-4
Molecular formula: C2H4Br2
1,2-Dibromoethane IS also known as ethylene dibromide and as 1,2-ethylene dibromide.
1.2 Physicochemical properties (WHO, 1996) 1
Property Value Physical state Colourless liquid Melting point 9.9°C Boiling point 131.4°C Vapour pressure 1.4 7 kPa at 25°C Water solubility 4.3 g/litre at 30°C
1.3 Organoleptic properties
1,2-Dibromoethane has a chloroform-like odour. No data on taste or odour thresholds have been identified.
1 Conversion factor in air: I mglm' = 0.13 ppm. 178 PESTICIDES
1.4 Major uses
1,2-Dibromoethane is used as a lead scavenger in tetra-alkyl lead petrol and antiknock preparations and as a fumigant for soils, grains, and fruits. However. with the phase-out of leaded petrol and the cancellation of the use of 1,2-dibromoethane in agricultural applications in many countries, use of this substance for these purposes has declined significantly. In addition to its continued use as a petrol additive in some countries, 1,2-dibromoethane is currently used principally as a solvent and as an intermediate in the chemical industry.
1.5 Environmentalfate
Because of its volatility, 1,2-dibromoethane's principal environmental sink is the atmosphere. The half-life for volatilization from surface waters is about 1-5 days under typical conditions (Mackay et al., 1982). 1,2-Dibromoethane is moderately mobile in soil, with the rate of penetration being greatest in sandy soils with low organic content (Townshend et al., 1980), although a small fraction may persist in the top layers for several years. For example, application of 1,2-dibromoethane into fine sandy loam at a "standard" rate of 70 kg/ha resulted in a concentration nearly 1 year later of 130 J..lg/kg; in another case, levels of up to 200 J..lg/kg were measured in soil 19 years after the last known application (rate unknown) (Steinberg et al., 1987). Diffusion of residues from soil to water is slow, with a diffusion coefficient of 10-16 cm2/s (Pignatello et al., 1987). 1,2-Dibromoethane is biodegraded within days in surface soils, whereas it may persist for months in aquifer materials (Pignatello & Cohen, 1990). It is not expected to bioconcentrate or biomagnify in terrestrial or aquatic food-chains (ATSDR, 1992).
2. ANALYTICAL METHODS
1,2-Dibromoethane is usually recovered from water samples by purge-and trap methods, direct absorption and thermal desorption, solvent extraction, or headspace collection, followed by analysis by gas chromatography (often in combination with mass spectrometry) with electron capture detection or flame or chemical ionization detection. Reported levels of detection for 1,2-dibromoethane in water range from 0.001 to 300 J..lg/litre (Keough et al., 1984; Koida et al., 1986; Stottmeister et al., 1986; Woodrow et al., 1986).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Mean levels of 1,2-dibromoethane in ambient air in 12 cities across Canada 3 surveyed between 1989 and 1992 ranged from non-detectable (i.e. <0.1 J..lg/m ) to 3 0.13 J..lg/m ; 1,2-dibromoethane was detected in 144 of 4298 samples analysed (Environment Canada, 1994). Concentrations ranging from 0.008 to 0.515 J..lg/m3 were measured in 71 samples from five urban and five mountainous locations in Japan in 1,2-DIBROMOETHANE 179
1983 (Environment Agency Japan, 1985). In seven sites in the USA, average levels ranged from 0.122 to 0.450 J.!g/m3 (Singh et al., 1982). Mean ambient concentrations in and around London, England, in 1982 and 1983 ranged from 0.15 to 1.2 J.!g/m3 (Clark et al., 1984a,b ). In the vicinity of waste and landfill sites in the USA, mean concentrations of 1,2-dibromoethane ranged from <0.038 to 5.4 J.!glm3 (Harkov et al., 1983, 1984).
3.2 Water
Few data on levels of 1,2-dibromoethane in drinking-water supplies have been identified. Based on a review of available surveys, Pignatello & Cohen (1990) reported that "typical" concentrations of 1,2-dibromoethane in samples from the wells in which it was detected (detected in 1959 of 15 450 wells surveyed) ranged from 0.01 to 15 J.!g/litre in the USA, from 1 to 2 J.!g/litre in Israel, and from 0.03 to 0.2 J.!g/litre in Australia. A maximum concentration of 94 J.!gllitre was measured in 15 samples from three irrigation wells surveyed between 1981 and 1983 in an area of Georgia, USA, in which 1,2-dibromoethane had been used extensively as a fumigant (Martl et al., 1984). 1,2-Dibromoethane was detected (detection limit not specified) in 34 of 421 samples of groundwater and in 11 of 175 samples of surface water from New Jersey; USA, during 1977-1979; the highest concentrations were 48.8 and 0.2 J.!g/litre in groundwater and surface water, respectively (Page. 1981 ). Concentrations of 0.06-0.55 J.!gllitre were detected between 1983 and 1984 in samples of groundwater from nine sites in an area of Japan where 1,2-dibromoethane had been applied to the soil (Terao et al., 1985). 1,2-Dibromoethane was not detected (detection limit 0.3-2 J.!g/litre) in 27 samples of surface water collected in Japan in 1982 (Environment Agency Japan, 1985).
3.3 Food
Few recent data on levels of 1,2-dibromoethane in foodstuffs have been identified. Daft (1989) detected (detection limit 1 ng/g) 1,2-dibromoethane in 2 of 549 samples of 234 foodstuffs collected in market basket surveys in the USA in the mid-1980s; 1,2-dibromoethane was detected in 1 sample each of whiskey and peanut butter at concentrations of 2 and 11 ng/g, respectively. Suzuki et al. ( 1989) measured concentrations of <0.02-0.51, 16.6-61.6, and 22.6-179 J.!g/kg in respective samples of mango (n = 9), papaya (n = 10), and grapefruit pulp (n = 30) imported into Japan in 1989. 1,2-Dibromoethane has been detected at levels of 0.33-0.47 mg/kg in bread made from wheat or flour that had been fumigated with the substance (FAO/WHO, 1972). Although 1,2-dibromoethane was detected at concentrations up to 0.4 mg/kg in flour, bran, and middlings of wheat that had been fumigated with a formulation containing the substance in Canada, no residues were detected (detection limit not specified) in 24 loaves of bread made from the exposed wheat (Berck, 1974). 180 PESTICIDES
3.4 Estimated total exposure and relative contribution of drinking-water
3 Based on a daily inhalation volume for adults of 20 m , a mean body weight of 60 kg, the assumption that concentrations of 1,2-dibromoethane are similar indoors and outdoors (as no data were identified on levels in indoor air), and the range of mean concentrations in ambient air determined in Canada (the most recent survey available is considered most appropriate, as use of this substance has declined 3 significantly) of <0.1-0.13 J.lg/m , the mean intake of 1,2-dibromoethane from air is estimated to range from <0.03 to 0.04 J.lg/kg of body weight per day. Although monitoring data are limited, intake of 1,2-dibromoethane in drinking-water is estimated to range from <0.0003 to 0.5 J.lg/kg of body weight per day, based on the daily consumption of 2 litres of drinking-water, a mean body weight of 60 kg, and the range of "typical" concentrations reported in drinking-water supplies in the USA, Israel, Australia, and Japan of <0.01-15 J.lg/litre. It should be noted, however, that these estimates for drinking-water are largely based on data collected before the use of 1,2-dibromoethane in agricultural practices was discontinued in many countries. Although 1,2-dibromoethane was not detected in >99% of samples of prepared foodstuffs in the USA in the mid-1980s, which would suggest that intake in food is negligible, food may be a significant source of exposure in areas where the substance is still being used as a grain fumigant; however, the lack of adequate monitoring data for such areas precludes quantitative estimation. Available data are inadequate to allow a quantitative estimation of the relative contribution of various environmental media to total exposure to 1,2-dibromoethane; however, air likely represents a principal source for the general population, with drinking-water being a major source in areas with significant groundwater contamination and food being a major source where the substance is still used as a grain fumigant.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
The concentration of 1,2-dibromoethane in the blood of guinea-pigs dermally exposed to the undiluted substance for 6 hours increased rapidly within the first hour and then slowly declined (Jakobson et al., 1982). Ingested 1,2-dibromoethane is rapidly and completely absorbed by the gastrointestinal tract, based on the recovery within 24 hours of radioactivity in the excreta and tissues of rats administered 15 mg of the C4C]-labelled compound per kg of body weight. The majority (72%) of the radioactivity was recovered in the urine, whereas, of the tissues examined, the liver contained the greatest amount of the label (1.8%) (Plotnick et al., 1979). 1,2-Dibromoethane can be metabolized by an oxidative pathway (cytochrome P-450 system) and a conjugation pathway (glutathione S-transferase system) (ATSDR, 1992). In the first pathway, 1,2-dibromoethane is oxidized via mixed function oxidases to 2-bromoacetaldehyde, which may subsequently be converted directly by aldehyde dehydrogenase to 2-bromoethanol or 2-bromoacetic acid. In addition, 2-bromoacetaldehyde may be conjugated with glutathione, ultimately resulting in the formation of thioglycolic acid and mercapturic acids (the primary 1,2-DIBROMOETHANE 181 urinary metabolites). In the second pathway, 1,2-dibromoethane may be conjugated with glutathione via glutathione S-transferase to form S-(2-bromoethyl)glutathione. This highly reactive intermediate may be detoxified by further conjugation to form ethylene, which is exhaled, and glutathione disulfide, which is eliminated in the faeces, or it may bind to DNA through direct nucleophilic substitution or through the ethylene-S-glutathionyl-episulfonium ion. The reactive metabolite bromoacetaldehyde, formed via the oxidation pathway, is believed to be responsible for non-genotoxic tissue damage (via covalent binding to proteins) (van Duuren et al., 1985), whereas S-(2-bromoethyl)glutathione, formed via the conjugation pathway, has been associated with genotoxicity and perhaps carcinogenicity (via interaction with DNA) in experimental systems (van Bladeren, 1983). Simula et al. (1993) demonstrated that human glutathione was capable of converting 1,2-dibromoethane to reactive metabolites, based on an increased mutagenicity in Salmonella typhimurium expressing human glutathione Al-l. Recently, Ploemen et al. (1995) reported genetic polymorphism in the ability of human erythrocytes to conjugate 1,2-dibromoethane via the class theta glutathione-S-transferases.
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
Ingested 1,2-dibromoethane is moderately acutely toxic in experimental animals. Reported oral LD 50s for rats, mice, guinea-pigs, rabbits, and chickens range from 55 to 420 mg/kg of body weight; rabbits appear to be the most sensitive species, whereas mice are the least sensitive (Rowe et al., 1952; McCollister et al.,
1956). The dermal LD 50 in rabbits is 450 mg/kg of body weight; dermal and eye contact with 1% 1,2-dibromoethane in solution also causes irritation in animals (Rowe et al., 1952).
5.2 Short-term exposure
Information on the short-term toxicity of ingested 1,2-dibromoethane is limited. In short-term studies preliminary to longer-term carcinogenicity bioassays, groups of five male or female Osborne-Mendel rats or B6C3F 1 mice were administered 1,2-dibromoethane in corn oil by gavage at doses of 0, 40, 63, 100, 163, or 251 mg/kg of body weight per day, 5 days/week, for 6 weeks, followed by a 2-week observation period (NCI, 1978). In rats, there were mortalities at 100 mg/kg of body weight per day, and decreased body-weight gain was observed at 100 mg/kg of body weight per day and above; in mice, deaths were observed at 100 and 251 mg/kg of body weight per day, and body weight gain was depressed at the highest dose. No other end-points appear to have been examined. 182 PESTICIDES
5.3 Long-term exposure and carcinogenicity
In a long-term bioassay (NCI, I978), groups of 50 male or female Osbome-Mendel rats (n = 20 in controls) were administered technical-grade I ,2-dibromoethane in corn oil by gavage at time-weighted average doses of 0, 38, or 41 mg/kg of body weight per day (males) or 0, 37, or 39 mg/kg of body weight per day (females), 5 days/week. [Rats in the high-dose groups were initially administered 80 mg/kg of body weight per day; however, owing to high mortality after 16 weeks (I8 males and 20 females), exposure was suspended for 13 weeks and commenced again at 40 mg/kg of body weight per day.] Surviving male and female rats were sacrificed after 49 and 6I weeks, respectively. Body-weight depression was apparent in the exposed rats after the first IO weeks. Hyperkeratosis and acanthosis of the forestomach were observed in I2/50 males and I8/50 females administered the high dose, in 4/50 females at the low dose, and in I/20 female controls. Other non-neoplastic lesions observed to occur with increased frequency in exposed rats included degenerative changes in the liver in males and in the adrenal gland in both sexes. Testicular atrophy was noted in I4/49 and I8/50 rats administered the low and high doses, respectively, but not in controls. Although early mortality was high in exposed rats, there were significant increases in the incidence of squamous cell carcinomas of the forestomach in both sexes; these were observed as early as week 12 of exposure and were locally invasive and eventually metastasized. The incidence of hepatocellular carcinomas or neoplastic nodules was significantly increased in females at the higher dose compared with controls. There were also increases in the incidence of haemangiosarcomas; the increase was significant at the lower dose in males. When early mortality was considered, the incidence of adenomas or carcinomas of the adrenal cortex was significantly elevated in females administered the high dose. The NCI (I978) also conducted a similar study in which groups of 50
B6C3F1 mice of either sex (n = 20 in controls) were administered time-weighted average I ,2-dibromoethane doses of 62 or I 07 mg/kg of body weight per day in corn oil by gavage, 5 days/week, for 53 weeks. [The actual doses administered ranged from 60 to I 00 mg/kg of body weight per day and from 60 to 200 mg/kg of body weight per day for the low- and high-dose groups, respectively.] Surviving animals were sacrificed after 78 weeks, except for females in the lower dose group, which were killed after 90 weeks. In both sexes, there were significant dose-related decreases in body-weight gain and survival in exposed mice and significant increases in the incidence of acanthosis and hyperkeratosis of the forestomach at the higher dose. Testicular atrophy was noted in 10/4 7 males administered I 07 mg/kg of body weight per day, compared with none in controls or the low-dose group. The incidence of squamous cell carcinomas of the forestomach was significantly increased in exposed mice of both sexes; as in rats, metastasis to distant sites was reported. Squamous cell papillomas were also observed in 2/49 males in the high dose group and in 1149 females in the low-dose group. There was also a significant increase in the incidence of alveolar/bronchiolar adenomas in both sexes; an alveolar/bronchiolar carcinoma was observed in one female mouse administered 62 mg/kg of body weight per day. 1,2-DIBROMOETHANE 183
Based on the results of these studies, the NCI (1978) concluded that 1,2-dibromoethane administered by gavage was carcinogenic to male and female
Osbome-Mendel rats and B6C3F 1 mice.
Van Duuren et al. (1985) exposed groups of30 male or female B6C3F 1 mice (n = 45 [males] or 50 [females] in controls) to 1,2-dibromoethane at a concentration of 4 mmol/litre in drinking-water (equivalent to a dose of 116 mg/kg of body weight per day for males and 103 mg/kg of body weight per day for females) for up to 456 (males) and 512 (females) days. Tissues from the lung, liver, kidneys, forestomach, and glandular stomach and any tissues appearing abnormal were examined histopathologically. Body weight was 10-20% depressed in exposed animals compared with controls (no other non-neoplastic effects were described). There were exposure-related increases in the incidences of squamous cell carcinomas or papillomas of the forestomach. Papillomas of the oesophagus were observed in 3/29 female mice, whereas none was observed in controls or exposed males. Groups of mice were also exposed to the metabolites bromoethanol or bromoacetaldehyde at a concentration of 4 mmolllitre (equivalent to doses of "16 [males] or 71 [females] mg/kg of body weight per day and 62 [males] and 85 [females] mg/kg of body weight per day, respectively). Exposure to bromoethanol induced a significant increase in the incidence of papillomas of the forestomach, whereas consumption of bromoacetaldehyde was not associated with a significant increase in tumour incidence at any site.
In an additional study in B6C3F 1 mice in which 1,2-dibromoethane was administered in drinking-water at a concentration of 2 mmolllitre (equivalent to a dose of 50 mg/kg of body weight per day) as a positive control for 18 months, there were significant increases in the incidences of squamous carcinomas of the forestomach in females and males, papillomas of the forestomach in females, papillomas of the oesophagus in males, and squamous carcinomas of the oesophagus in males (van Duuren et al., 1986). Repeated dermal exposure of male and female Ha:ICR Swiss mice to 25 or 50 mg of 1,2-dibromoethane in acetone for up to 594 days induced a significant increase in the incidence of skin and pulmonary tumours (van Duuren et al., 1979). Exposure to 1,2-dibromoethane via inhalation has also been carcinogenic in 3 experimental animals. B6C3F 1 mice exposed to concentrations of 77 mg/m and above for up to 2 years had increased incidences of several tumour types, including alveolar/bronchiolar carcinomas and adenomas, haemangiosarcomas, fibrosarcomas of the subcutaneous tissue, carcinomas of the nasal cavity, and adenocarcinomas of the mammary gland (NTP, 1982). Similarly, exposure to 1,2-dibromoethane at 77 mg/m3 and above for up to 2 years induced carcinomas, adenocarcinomas, adenomas, and adenomatous polyps of the nasal cavity, haemangiosarcomas, mesotheliomas of the tunica vaginalis, fibroadenomas of the mammary gland, and alveolar/bronchiolar adenomas and carcinomas (combined) in groups of F-344 rats (NTP, 1982). Adkins et al. (1986) also reported an increase in pulmonary adenomas in female A/J mice exposed to 1,2-dibromoethane at 154 or 384 mg/m3 for 6 months. 1,2-Dibromoethane did not demonstrate an ability to initiate tumour promotion in a dermal initiation/promotion assay in mice (van Duuren et al., 1979), although positive results were reported for initiation of liver foci in rats (Moslen, 1984). 184 PESTICIDES
5.4 Reproductive and developmental toxicity
Testicular atrophy was observed in Osbome-Mendel rats administered time weighted average doses of 38 mg/kg of body weight per day or more for up to
48 weeks and in B6C3F1 mice administered a time-weighted average dose of 107 mg/kg of body weight per day for up to 78 weeks in the long-term bioassays conducted by the NCI (1978) (discussed above in section 5.3). Effects on sperm density and motility and abnormal spermatozoa were observed in bulls orally administered 1,2-dibromoethane doses of 2 or 4 mg/kg of body weight per day for 12 months or 2 weeks, respectively (Amir and Volcani, 1965). In addition, Short et al. ( 1979) reported reduced testicular weights, reduced serum testosterone concentrations, inability to impregnate females, and atrophy of the testes, epididymis, prostate, and seminal vesicles in groups of 9 or 10 male Sprague Dawley CD rats exposed to 1,2-dibromoethane at 684 mg/m3 for 10 weeks, although increased mortality and decreased body-weight gain were also noted; no effects were 3 reported at 146 or 300 mg/m . However, there were no effects on reproductive parameters in groups of 10 male albino rats administered 1,2-dibromoethane at concentrations of 100 or 500 mg/kg of diet (equivalent to 10 and 50 mg/kg of body weight per day) for 90 days and mated with unexposed females. Histological examination of the testes revealed no exposure-related changes, and the mean number of litters per group, mean pup weight at birth, and sex ratio of pups were similar to controls (Shivanandappa et al., 1987). Short et al. ( 1979) also reported reversible abnormalities in the estrous cycle 3 3 in groups of 20 female rats exposed to 614 mg/m but not 154 or 300 mg/m ; decreased body-weight gain and increased mortality were also observed at 3 614 mg/m • No differences in the numbers of total implants, viable implants, or resorptions per dam were observed at any concentration. Morphological changes and abnormalities, including haematomas, exencephaly, and skeletal variations, were observed in fetuses of female Sprague-Dawley rats and CD-I mice exposed to 1,2-dibromoethane via inhalation at concentrations that were also toxic to the dams 3 (LOECs of292 and 154 mg/m , respectively) (Short et al., 1978).
5.5 Mutagenicity and related end-points
The genotoxicity of 1,2-dibromoethane has been extensively investigated in both in vitro and in vivo assays. In in vitro tests in prokaryotic organisms, 1,2-dibromoethane was generally genotoxic, both with and without exogenous metabolic activation (e.g. Principe et al., 1981; Zoetemelk et al., 1987), although negative results were obtained in some assays (e.g. Buselmaier et al., 1972; Shiau et al., 1980). In cultured mammalian cells, 1,2-dibromoethane caused forward mutations (e.g. Clive et al., 1979; Brimer et al., 1982), sister chromatid exchanges (Tezuka et al., 1980; Tucker et al., 1984), unscheduled DNA synthesis (Tennant etal., 1986; Working et al., 1986), and cell transformation (Perocco et al., 1991; Colacci et al., 1995). In in vivo assays, 1,2-dibromoethane induced recessive lethal mutations (e.g. Vogel & Chandler, 1974; Kale & Kale, 1995), gene mutations (Graf et al., 1984), and mitotic recombination (Graf et al., 1984) in Drosophila 1,2-D/BROMOETHANE 185 melanogaster. DNA damage was observed in the liver of mice and rats exposed orally or by intraperitoneal injection (e.g. Nachtomi & Sarma, 1977; Storer & Conolly, 1983 ). However, the substance did not induce dominant lethal mutations (e.g. Short et al., 1979; Barnett et al., 1992) or specific locus mutations (Russell, 1986; Barnett et al., 1992) in the germ cells of rodents. Similarly, negative results were obtained for sister chromatid exchange (Krishna et al., 1985), chromosomal aberrations (Krishna et al., 1985), unscheduled DNA synthesis (Bentley & Working, 1988), and micronucleus induction (Krishna et al., 1985; Asita et al., 1992) in mice or rats. Although the oxidative pathway of metabolism of 1,2-dibromoethane can generate metabolites capable of damaging DNA (i.e. bromoacetaldehyde and 2-bromoethanol), glutathione conjugation likely also contributes to the genotoxicity of this substance. DNA adduct formation has also been observed following administration of 1,2-dibromoethane, mostly (>95%) due to the glutathione conjugate (Sundheimer et al., 1982; Ozawa & Guengerich, 1983; Inskeep & Guengerich, 1984). Depletion of cellular glutathione inhibited unscheduled DNA synthesis in rat hepatocytes exposed to 1,2-dibromoethane, whereas inhibition of cytochrome P-450 mediated oxidation had no effect (Working et al., 1986).
5.6 Immunotoxicity
Administration of oral doses of up to 200 mg/kg of body weight per day for
14 days to groups of 8-20 female B6C3F 1 mice resulted in a number of changes in immune parameters, including numerical alterations and effects on in vitro function of various cell types, at doses that also induced changes in organ weights. However, no effects were observed in 111 vivo host resistance to infectious agents (Ratajczak et al., 1994).
6. EFFECTS ON HUMANS
Little information on the toxicity of ingested 1,2-dibromoethane in humans was identified. It was estimated that 200 mg/kg of body weight is lethal to humans, based on the death of a 60-kg woman who had ingested 12 g (Alexeeff et al., 1990). There were no significant increases in mortality due to neoplastic or non neoplastic causes in two studies of populations exposed to 1,2-dibromoethane in the workplace (Turner and Barry, 1979; Ott et al., 1980), although these studies were limited by the small size of the study populations, lack of accounting for possible confounding factors (such as smoking or exposure to other substances), and inadequate data on exposure. No differences in the frequency of sister chromatid exchanges or chromosomal aberrations were observed in 60 workers at six papaya packing plants exposed to a mean 1,2-dibromoethane concentration of 0.68 mg/m3 (with peak levels 3 of up to 2.0 mg/m ) for an average of 5 years compared with a group of 40 controls, even when only workers with long-term exposure (>5 years) or those with the highest peak exposure were considered (Steenland et al., 1986). 186 PESTICIDES
Adverse effects on reproductive ability, including reduced sperm counts, viability, and motility and an increase in sperm abnormalities, were also observed in a cross-sectional study of 46 papaya fumigators exposed to a time-weighted average 1,2-dibromoethane concentration of 0.68 mg/m3 for an average of 5 years, compared with a control group of 43 unexposed men, taking into account smoking and both caffeine and alcohol consumption (Ratcliffe et al., 1987; Schrader et al., 1987). Similar effects, along with reduced fertility, were also observed in other limited investigations of occupationally exposed men (Ter Haar, 1980; Takahashi et al., 1981; Wong et al., 1985; Schrader et al., 1988). However, limitations of these studies, including lack of appropriate controls, small group sizes, and lack of controlling for potentially confounding factors, preclude determination of an effect level.
7. PROVISIONAL GUIDELINE VALUE
1,2-Dibromoethane has induced an increased incidence of tumours at several sites in all carcinogenicity bioassays identified in which rats or mice were exposed to the compound by gavage (NCI, 1978), ingestion in drinking-water (van Duuren et al., 1985, 1986), dermal application (van Duuren et al., 1979), and inhalation (NTP, 1982; Adkins et al., 1986), although many of these studies were limited by high early mortality, limited histopathological examination, small group sizes, or use of only one exposure level. The substance acted as an initiator of liver foci in an initiation/promotion assay (Moslen, 1984) but did not initiate skin tumour development (van Duuren et al., 1979). 1,2-Dibromoethane was consistently genotoxic in in vitro assays, although results in in vivo assays were mixed. Biotransformation to active metabolites, which have been demonstrated to bind to DNA, is likely involved in the induction of tumours. Available data do not support the existence of a non-genotoxic mechanism of tumour induction. Therefore, based on the available data, it is concluded that there is sufficient evidence that 1,2-dibromoethane is a genotoxic carcinogen in rodents. As data on the potential carcinogenicity in humans are inadequate, and as it is likely that 1,2-dibromoethane is metabolized similarly in rodent species and in humans (although there may be varying potential for the production of active metabolites in humans, owing to genetic polymorphism), 1,2-dibromoethane is considered to be probably carcinogenic in humans, based on the results of studies in experimental species (IPCS, 1995; WHO, 1996). Similarly, IARC (1987) classified 1,2-dibromoethane in Group 2A (probably carcinogenic to humans). Although most of the available bioassays for 1,2-dibromoethane are limited, particularly those in which the compound was administered by ingestion, these studies can nevertheless be used to calculate approximate estimates of the carcinogenic potency of 1,2-dibromoethane. In view of the serious limitations of the studies, however, these estimates must be considered imprecise and, therefore, provisional. Lifetime low-dose cancer risks can be calculated by linearized multistage modelling of the incidences of haemangiosarcomas and tumours in the stomach, liver, lung, and adrenal cortex (adjusted for the observed high early mortality, where appropriate, and corrected for the expected rate of increase in tumour formation in rodents in a standard bioassay of 104 weeks) of rats and/or mice administered 1,2-DIBROMOETHANE 187
1,2-dibromoethane by gavage. The drinking-water concentrations that correspond 5 6 to excess lifetime cancer risks (for various tumour types) of 10-4, 10- , and 10- are 4-150 J.tg/litre, 0.4-15 J.tgllitre, and 0.04-1.5 J.tg/litre, respectively. 2 Because of the serious limitations of the critical studies, these concentrations are considered to be approximate estimates only, and hence guideline values derived from them should be considered provisional.
8. REFERENCES
Adkins B et al. (1986) Oncogenic response of strain A/J mice to inhaled chemicals. Journal of toxicology and environmental health, 17:311-322. Alexeeff GV, Kilgore WW, LiMY (1990) Ethylene dibromzde: toxicology and risk assessment. New York, NY, Springer-Verlag, pp. 49-122. Amir D, Volcani R (1965) Effect of dietary ethylene dibromide on bull semen. Nature (London), 206( 4979):99-1 00. Asita AO et al. (1992) Micronucleated reticulocyte induction by ethylating agents in mice. Mutatzon research, 271:29-37. ATSDR (1992) Toxicological profile for 1,2-dibromoethane. Atlanta, GA, US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry. Barnett LB et al. (1992) Ethylene dibromide: negative results with the mouse dominant lethal assay and the electrophoretic specific-locus test. Mutatzon research, 282:127-133. Bentley KS, Working PK (1988) Activity of germ-cell mutagens and nonmutagens in the rat spermatocyte UDS assay. Mutation research, 203:135-142. Berck B (1974) Fumigant residues of carbon tetrachloride, ethylene dichloride, and ethylene dibromide in wheat, flour, bran, middlings, and bread. Journal of agrzcultural and food chemistry, 22(6):977-984. Brim er PA, Tan EL, Hsie A W ( 1982) Effect of metabolic activation on the cytotoxicity and mutagenicity of 1,2-dibromoethane in the CHO/HGPRTsystem. Mutation research, 95:377-388. Buselmaier W, Rohrborn G, Propping P (1972) [Mutagenicity investigations with pesticides in host mediated assay and dominant lethal test in mice.] Biologisches Zentralblatt, 91:311-325 (in German). Clark AI et al. (1984a) Ambient air measurements of aromatic and halogenated hydrocarbons at urban, rural and motorway locations. Science of the total environment, 39:265-279. Cl ark AI et al. (1984b) Monitoring and assessment of ambient atmospheric concentrations of aromatic and halogenated hydrocarbons at urban, rural and motorway locations. Environmental pollution, 87:141-158. Clive D et al. (1979) Validation and characterization of the L5178Y/TK +/- mouse lymphoma mutagen assay system. Mutation research, 59:61-108. Colacci A et al. (1995) 1,2-Dibromoethane as an initiating agent for cell transformation. Japanese ;ournal of cancer research, 86:168-173. Daft JL (1989) Determination of fumigants and related chemicals in fatty and non fatty foods. Joumal ofagncultural and food chemistry, 37(2):560-564. Environment Agency Japan (1985) Chemzcals 111 the environment Annual report- Annex. Tokyo, pp. 16-28.
Although the study in which animals were exposed to I ,2-dibromoethane in drinking-water was considered to be too limited to use as the basis for derivation of a guideline value (owing to high mortality at the only exposure level), drinking-water concentrations corresponding to risks of 10--', w-', and 10-", derived based on the results of this study, would be similar to the range presented here. Likewise, values derived on the basis of the exposure-response relationship observed for non respiratory tract tumours in the study in which animals were exposed via inhalation are similar. 188 PESTICIDES
Environment Canada (1994) Summary of EDB concentrations at Canadian sites. Ottawa, Ontario, Environment Canada, Environmental Protection Service, Pollution Measurement Division (unpublished data). FAO/WHO (1972) 1971 evaluations of some pesticide residues m food Geneva, World Health Organization, p. 267 (WHO Pesticide Residues Series No. 1). Graf U et al. (1984) Somatic mutation and recombination test in Drosophila melanogaster. Envzronmental mutagenesis, 6:153-188. Harkov R et al. ( 1983) Measurement of selected volatile organic compounds at three locations in New Jersey during the summer season. Journal of the Air Pollution Control Association, 33(12): 1177-1183. Harkov R et al. (1984) Comparison of selected volatile organic compounds during the summer and winter at urban sites in New Jersey. Science of the total environment, 38:259-274. !ARC (1987) Ethylene dibromide. In: Overall evaluations of carcznogenzcity: An updatmg of !ARC Monographs volumes 1-42 Lyon, International Agency for Research on Cancer, pp. 204- 205 (JARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Suppl. 7). Inskeep PB, Guengerich FP (1984) Glutathione-mediated binding of dibromoalkanes to DNA: specificity of rat glutathione-S-transferases and d1bromoalkane structure. Carcznogeneszs, 5(6):805-808. IPCS (1995) Report of the 1994 meeting of the Core Assessment Group. Geneva, World Health Organization, International Programme on Chemical Safety, Joint Meeting on Pesticides (WHO/PCS/95. 7). Jakobson I et a! (1982) Uptake via the blood and elimination of 10 orgamc solvents following epicutaneous exposure of anesthetized guinea pigs. Toxzcology and applied pharmacology, 63:181-187. Kale P, Kale R (1995) Induction of delayed mutations by benzene and ethylene dibromide in Drosophila. Envtronmental and molecular mutagenesis, 25:211-215. Keough T et al. (1984) SeparatiOn and use of the perdeuterated analogue as an internal standard for the analysis of ethylene dibromide. Journal ofchromatography, 312:450-455. Koida K et al. (1986) [Determination of ethylene dibromide and dibromochloropropane in groundwater by gas chromatography-negative ion chemical ionization mass spectrometry.] Hiroshzma-shi Eisez Kenkyusho Nenpo, 5:34-36 (in Japanese). Krishna G et al. (1985) In vivo cytogenetic studies on mice exposed to ethylene dibromide Mutatton research, 158:81-87. Mackay D et al. ( 1982) Volatilization of organic pollutants from water. US Environmental Protection Agency (EPA-600/53-82-0 19) Mart! LR, de Kanel J, Dougherty RC ( 1984) Screening for organic contamination of groundwater: ethylene d1bromide in Georgia Irrigation wells. Envzronmental sczence and technology, 18:973-974 McCollister DD et al. (1956) Comparative inhalation toxicity of fumigant mixtures. American Medtcal Association archzves of industrzal health, 13 ·1-7. Moslen MT ( 1984) Increased incidence of hepatic foci and nodules in rats given one or two doses of 1,2-dibromoethane Toxicologic pathology, 12(4):307-314. Nachtomi E, Sarma DS (1977) Repair of rat liver DNA tn vzvo damaged by ethylene dibromide. Bwchemzcal pharmacology, 26: 1941-1945. NCI (1978) Bioassay of I ,2-dibromoethane for possible carcinogenicity. Bethesda, MD, National Cancer Institute, Division of Cancer Cause and Prevention, Carcinogenesis Testing Program, 64 pp. (Technical Report Senes No. 86; DHEW Publication No. (NIH) 78-1336). NTP (1982) Carcinogenests bwassay of /,2-dibromoethane m F344 rats and B6C3FJ nuce (mhalatwn study) Research Triangle Park, NC, US Department of Health and Human Services, National Toxicology Program, 163 pp. (NTP Technical Report Series No. 21 0). Ott MG, Scharnweber HC. Langner RR (1980) Mortality expenence of 161 employees exposed to ethylene d1bromide in two production units. British ;ournal of industrzal medzcine, 37·163-168. Ozawa N, Guengerich FP ( 1983) Evidence for formation of an S-[2-(N7-guanyl)ethyl]glutathione adduct in glutathione-mediated binding of the carcinogen 1,2-dibromoethane to DNA. Proceedmgs ofthe Natwnal Academy ofSctences ofthe Umted Stales ofAmerica, 80:5266-5270. 1,2-DIBROMOETHANE 189
Page GW (1981) Comparison of groundwater and surface water for patterns and levels of contamination by toxic substances. Envu-onmental sc1ence and technology, 15(12).14 75-1481. Perocco P et al. (1991) Transforming activity of ethylene dibromide in BALB/c 3T3 cells Research commumcattons tn chemical pathology and pharmacology, 73(2): 159-172 Pignatello JJ, Cohen SZ ( 1990) Environmental chemistry of ethylene dibromide m soil and ground water Revrews of envrronmental contammatton and toxrcology, 112.1-47. Pignatello JJ, Sawhney BL, Frink CR (1987) EDB. Persistence in soil [letter]. Scrence, 236:898. Ploemen JH et al. (1995) Polymorphism in the gluthathione conjugation activity of human erythrocytes towards ethylene dibromide and 1,2-epoxy-3-(p-nitrophenm,y)-propane. Bioclwmca Biophysrca Acta, 1243:469-476. Plotnick HB et al. (1979) The effect of dietary disulfiram upon the tissue distnbut10n and excretion of 14 C-l ,2-dibromoethane in the rat. Research communrcatrons rn chemrcal pathology and pharmacology, 26(3):535-545. Principe P et al (1981) Mutagenicity of chemicals of industrial and agncultural relevance in Salmonella, Streptomyces and Aspergillus. Journal of the scrence of food and agrrculture, 32:826-832. Ratajczak HV et al. (1994) Ethylene dibromide. evidence of systemic and immunologic toxicity without Impairment of m vivo host defenses. In V1vo, 8(5):879-884 Ratcliffe JM et al. (1987) Semen quality in papaya workers with long term exposure to ethylene dibromide. Brrt1sh;ournal of mdustrial med1cme, 44:317-326. Rowe VK et al. (1952) Toxicity of ethylene dibromide determined on experimental animals. Amerrcan Med1cal Assocwtton archives of mdustnal hyg1ene and occupatiOnal medicine, 6:158-173. Russell WL (1986) Positive genetic hazard predictiOns from short-term tests have proved false for results in mammalian spermatogonia with all environmental chemicals so far tested. In: Gene/re toxicology of envrronmental chenucals - Part 8: Genetic effects and applied mutagenesis. New York, NY, Alan R. Liss, pp. 67-74. Schrader SM et al. ( 1987) The use of new field methods of semen analysis m the study of occupational hazards to reproduction - the example of ethylene dibromide. Journal of occupational med1cme, 29·963-966. Schrader SM, Turner TW, Ratcliffe JM (1988) The effects of ethylene dibromide on semen quality: a comparison of short-term and chrome exposure. Reproductive toxicology, 2:191-198. Shiau SY et al. ( 1980) Mutagenicity and ON A-damaging activity for several pesticides tested with Bac1llus subtihs mutants. Mutatton research, 71.169-179. Shivanandappa T, Krishnakumari MK, Majumder SK (1987) Reproductive potential of male rats fed dietary ethylene dibromide. Journal offood safety, 8:147-155. Short RD et al. (1978) Inhalation of ethylene dibromide during gestation by rats and mice Toxicology and apphed pharmacology, 46:173-182. Short RD et al. (1979) Effects of ethylene dibromide on reproduction in male and female rats. Toxicology and applied pharmacology, 49:97-105. Simula TP, Glancey MJ, Wolf CR (1993) Human glutathione S-transferase-expressing Salmonella typhimurrum tester strains to study the activation/detoxification of mutagenic compounds· studies with halogenated compounds, aromatic amines and aflatoxin B I. Carcmogenesis, 14:1371-1376. Singh HB. Salas LJ, Stiles RE (1982) Distribution of selected gaseous organic mutagens and suspect carcinogens in ambient air. Envzronmental sc1ence and technology, 16(12):872-880 Steenland K et al. (1986) A cytogenetic study of papaya workers exposed to ethylene dibromide. Mutation research, 170:151-160. Steinberg SM, Pignatello JJ, Sawhney BL (1987) Persistence of 1,2-dibromoethane in soils: entrapment in intraparticle micropores. Environmental sc1ence and technology, 21(12): 1201-1208. Storer RD, Conolly RB (1983) Comparative in v1vo genotoxicity and acute hepatotoxicity of three I ,2-dihaloethanes. Carcinogenesis, 4(11): 1491-1494. 190 PESTICIDES
Stottmeister E, Hendel P, Engewald W (1986) [Gas-chromatographic trace analysis ofh1ghly volatile halogenated hydrocarbons in water with FID detectiOn.] Acta Hydroch1mica et Hydrobiologica, 14(6):573-580 (in German). Sundheimer DW et al. (1982) The bioactivation of I ,2-dibromoethane in rat hepatocytes: covalent binding to nucleic acids. Carcinogenesis, 3(10): 1129-1133. Suzuki T et al. (1989) Some problems with 1,2-dibromoethane residue analysis. Journal of agricultural and food chem1stry, 37(2):564-567. Takahashi W et al. (1981) DepressiOn of sperm counts among agricultural workers exposed to dibromochloropropane and ethylene dibromide. Bul!etm of environmental contammation and toxzcology, 27:551-558. Tennant RW, Stasiewicz S, Spalding JW (1986) Comparison of multiple parameters of rodent carcinogemcity and in vitro genetic toxicity. Environmental mutagenesis, 8:205-227. Terao H et al. (1985) [Influence of pesticide and fertilizer on the ground water in vegetable field zone.] Chikyu Kagaku, 19:31-3 8 (in Japanese). Ter Haar G (1980) An investigation of possible sterility and health effects from exposure to ethylene dibromide. In: Ames B, Infante 0, Peirtz R, eds. Ethylene dichlonde: a potential health risk? Cold Spring Harbor, NY, Cold Spring Harbor Laboratory, pp. 167-188 (Banbury Report No. 5). Tezuka H et a! ( 1980) Sister-chromatid exchanges and chromosomal aberrations in cultured Chinese hamster cells treated with pesticides positive in microbial reversion assays. Mutation research, 78:177-191. Townshend JL, Dirks VA, Marks CF ( 1980) Temperature, moisture and compaction and their effects on the diffusion of ethylene dibromide in three Ontario soils. Canadian journal of soil sc1ence, 60:177-184. Tucker JD et al. (1984) Detection of sister-chromatid exchanges in human peripheral lymphocytes induced by ethylene dibromide vapor. Mutation research, 138:93-98. Turner D, Barry PSI ( 1979) An epidemiological study of workers in plants manufacturing ethylene dibromide. Arh1v za Higl)enu Rada 1 Toksiko/ogl)u, 30:621--{526. van Bladeren PJ (1983) Metabolic activation of xenobiotics: ethylene dibromide and structural analogs. Journal of the Amencan College of Toxicologists, 2(3):73-83. van Duuren BL et al. (1979) Carcinogenicity of halogenated olefinic and aliphatic hydrocarbons in mice. Journal of the National Cancer Institute, 63(6): 1433-1439. van Duuren BL et al. (1985) Carcinogenicity bioassays of bromoacetaldehyde and bromoethanol - potential metabolites of dibromoethane. Teratogenesis, carcinogenesis, and mutagenesis, 5:393-403. van Duuren BL et al. (1986) Chronic bioassays of chlorinated humic acids in B6C3F 1 mice. Environmental health perspectives, 69: I 09-117. Vogel E, Chandler JL (1974) Mutagenicity testing of cyclamate and some pesticides in Drosophila melanogaster. Expenentw, 30:621--{523. WHO (1996) 1,2-Dzbromoethane. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria 177). Wong 0, Morgan RW, Whorton MD (1985) An epidemiologic surveillance program for evaluating occupational reproductive hazards. American journal of industrial medicine, 7:295-306. Woodrow JE, Majewski MS, Seiber JN (1986) Accumulative sampling of trace pesticides and other organics in surface water using XAD-4 resin. Journal of environmental science and health, 821(2):143-164. Working PK et al. (1986) Induction of DNA repair in rat spermatocytes and hepatocytes by 1,2- dibromoethane: the role of glutathione conjugation. Carcinogenesis, 7:467-472. Zoetemelk CE et al. (1987) 1,2-Dibromocompounds. Their mutagenicity in Salmonella strains differing in glutathione content and their alkylating potential. Bwchemica/ pharmacology, 36(11):1829-1835. 2,4-DICHLOROPHENOXYACETIC ACID (2,4-D)
First draft prepared by H. Galal-Gorchev International Programme on Chemical Safety World Health Organization, Geneva, Switzerland
2,4-Dichlorophenoxyacetic acid (2,4-D) was evaluated in the 1993 WHO Guidelines for drinking-water quality, and a guideline value of 30 11gllitre was recommended based on a TDI of 0.01 mg/kg of body weight (derived from a NOAEL of 1 mg/kg of body weight per day for effects on the kidney in 2-year studies in rats and mice). As JMPR re-evaluated this pesticide in 1996, the Coordinating Committee Meeting of December 1995 considered the updating of the guideline value for 2,4-D based on the JMPR evaluation to be of high priority.
1. GENERAL DESCRIPTION
The term 2,4-D is used here to refer to the free acid, 2,4-dichlorophenoxyacetic acid. 2,4-D is a strong acid and forms water-soluble salts with alkali metals and amines. Commercial 2,4-D products are marketed as the free acid, alkali and amine salts, and ester formulations. 2,4-D itself is chemically stable, but its esters are rapidly hydrolysed to the free acid (CCME, 1995).
1.1 Identity
CAS no.: 94-75-7
Molecular formula: C8H6Cl20 3
1.2 Physicochemical properties (Tom! in, 1994)
Property Value Water solubility 311 mg/litre (pH 1, 25°C) Vapour pressure 1.1 x 10-2 Pa (20°C) Log octanol-water partition coefficient 2.58-2.83 (pH I)
1.3 Organoleptic properties
The taste and odour threshold of 2,4-D is 3.13 mg/litre (US EP A, 1987). However, the taste threshold in water of 2,4-dichlorophenol, 2,4-D's main degradation product, is 0.3 11g/litre; as a result, public water supplies containing "traces" of 2,4-D are often shut down because of objectionable odours or tastes (WHO, 1984). 192 PESTICIDES
1.4 Major uses
2,4-D is a systemic herbicide used for post-emergence control of annual and perennial broad-leaved weeds in cereal cropland, on lawns, turf, and pastures, in forests, and in non-cropland (including areas adjacent to water). It is also used to control broad-leaved aquatic weeds. Impurities may be present in the technical product as a result of the manufacturing process (Health Canada, 1993).
1.5 Environmenta/fate
2,4-D can enter the environment through effluents and spills arising from its manufacture and transport and through direct application as a weed control agent. It is removed from the environment principally by biodegradation through several possible pathways. with the formation of 2,4-dichlorophenol as an intermediate. 2,4-D is removed from the atmosphere by photo-oxidation and rainfall, with a half life of less than 1 day. The half-life of 2,4-D in soil is reported to range from 4-7 days in most soil types to up to 6 weeks in acidic soils. 2,4-D is rapidly biodegraded in water, although some may be degraded by photolysis near the surface. Half-lives in water range from 1 to several weeks under aerobic conditions and can exceed 120 days under anaerobic conditions. 2,4-D is not expected to accumulate in bottom sediments and mud. Except for some algae, it does not bioaccumulate in aquatic or terrestrial organisms because of its rapid degradation (Health Canada, 1993).
2. ANALYTICAL METHODS
Residues of 2,4-D and its salts and esters in water are commonly measured by extraction, chemical derivatization, separation by gas-liquid chromatography, and electron capture detection. This method is suitable for the detection of picogram levels. Electrolytic conductivity detection is also used; its detection limit is 0.1 ~g/litre (Health Canada, 1993).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Atmospheric contamination by 2,4-D may occur as a result of volatilization and drift from application by spraying. Residues in the atmosphere are predominantly in the form of isopropyl and butyl esters. In large-scale studies in areas of intense 2,4-D use in Canada, about 40% of all air samples were found to 3 contain between 0.01 and 0.1 ~g of 2,4-D per m • In a general programme of air quality monitoring undertaken in citrus-growing regions in the USA, only I out of 880 samples analysed was found to contain 2,4-D, at a level of 4 ~g/m 3 (WHO, 1984). 2,4-DICHLOROPHENOXYACETIC ACID (2,4-D) 193
3.2 Water
2,4-D was detected, at a maximum concentration of 29 fig/litre, in 52 of 805 samples of raw and treated drinking-water from municipal and private water supplies in surveys conducted in six Canadian provinces from 1971 to 1986 (Health Canada, 1993). Surface water samples taken between 1983 and 1984 in Manitoba, Canada, indicated the presence of 2,4-D at concentrations ranging from 0.01 to 0.13 fig/litre (CCME, 1995). In the USA, reported levels of 2,4-D in drinking-water supplies have been below 0.5 fig/litre, with most levels below 0.1 fig/litre (US EP A, 1987). Generally, 2,4-D residues in surface waters are <0.1 fig/litre. This is not unexpected, in view of the relatively rapid biodegradation of 2,4-D in the environment (WHO, 1984).
3.3 Food
2,4-D and its degradation products do not tend to accumulate in food. Available evidence indicates that residues of 2,4-D rarely exceed a few tens of flg/kg in food. Exceptions may occur in liver and kidney from range animals; in berries and mushrooms grown in treated right-of-way areas; or when the herbicide is used in quantities far in excess of the rates applied in normal agricultural practice. High residues of 2,4-D can produce disagreeable odours or flavours in fruits and vegetables, and this may reduce the likelihood that highly contaminated foods are ingested (WHO, 1984).
3.4 Estimated total exposure and relative contribution of drinking-water
The available data are insufficient to determine whether food or water is the greater source of exposure to 2,4-D.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
2,4-D was rapidly absorbed, distributed, and excreted after oral administration to mice (Eiseman, 1984), rats (Smith et al., 1990), and goats (Guo & Stewart, 1993). At least 86-94% of an oral dose was absorbed from the gastrointestinal tract in rats. Once absorbed, 2,4-D was widely distributed in the body but did not accumulate because of its rapid clearance from the plasma and rapid urinary excretion. 2,4-D was excreted rapidly and almost exclusively (85-94%) in urine by 48 hours after treatment, primarily as unchanged 2,4-D. No metabolites have been reported other than conjugates. Pharmacokinetic studies with salts and esters of 2,4-D have shown that the salts rapidly dissociate and the esters are rapidly hydrolysed to 2,4-D, after which their fate was indistinguishable from that of the acid. The similarity in the fate of 2,4-D and its salts and esters explains their similar toxicity (FAO/WHO, 1997a). 194 PESTICIDES
In humans, following ingestion of a 5 mg/kg of body weight dose, 2,4-D was quickly absorbed and excreted rapidly in the urine; about 73% of the administered dose was found in the urine after 48 hours. No metabolites were detected (Sauerhoff et al., 1977).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
2,4-D, its amine salts, and its esters are slightly toxic when administered orally, the oral LD50 in rats being 400-2000 mg/kg of body weight (FAO/WHO, 1997a). WHO has classified 2,4-D as "moderately hazardous" (WHO, 1996).
5.2 Short-term exposure
Groups of 10 male and 10 female B6C3F1 mice were fed diets containing technical-grade 2,4-D (purity 96.1%) at 0, 1, 15, 100, or 300 mg/kg of body weight per day for 90 days. Treatment-related changes were observed in animals of both sexes at 100 and 300 mg/kg of body weight per day. These effects included decreases in glucose level in females, decreases in thyroxine activity in males, and increases in absolute and relative kidney weights in females. The NOAEL was 15 mg/kg of body weight per day on the basis of renal toxicity (Schultze, 199la). Groups of Fischer 344 rats (20 per sex per group) received 2,4-D (purity 97.5%) in the diet at doses of 0, 1, 5, 15, or 45 mg/kg of body weight per day for 90 days. Renal lesions were observed in animals of both sexes at 5 mg/kg of body weight per day and above. The NOAEL was 1 mg/kg of body weight per day on the basis of renal toxicity (Serota, 1983). In Fischer 344 rats (20 per sex per group) fed diets providing 2,4-D (purity 96.1 %) at doses of 0, 1, 15, 100, or 300 mg/kg of body weight per day for 13 weeks, treatment-related changes were observed in animals of both sexes at 100 and 300 mg/kg of body weight per day. These effects included decreases in body-weight gain, haematological and clinical chemical alterations, changes in organ weights, and histopathological lesions in the adrenals, liver, and kidneys. The NOAEL was 15 mg/kg of body weight per day (Schultze, 199lb). Groups of five male and five female beagle dogs were given gelatin capsules containing 2,4-D (purity 96.1%) at doses ofO, 0.3, 1, 3, or 10 mg/kg of body weight per day for 13 weeks. Renal lesions were observed at 3 and 10 mg/kg of body weight per day. The NOAEL was 1 mg/kg of body weight per day (Schultze, 1990). Groups of beagle dogs (five per sex per dose) were fed diets containing 2,4-D (purity 96.5%) at doses ofO, 1, 5, or 7.5 mg/kg of body weight per day for 52 weeks. At 5 and 7.5 mg/kg of body weight per day, body-weight gain was decreased, increases were observed in blood urea nitrogen, creatinine, alanine aminotransferase activity, and cholesterol, and histopathological lesions were observed in the kidneys and liver. The NOAEL was 1 mg/kg of body weight per day (Dalgard, 1993). 2,4-DICHLOROPHENOXYACETIC ACID (2,4-D) 195
5.3 Long-term exposure and carcinogenicity
In a 2-year study of toxicity and carcinogenicity, groups of 50 male and 50 female B6C3F 1 mice were fed diets providing 2,4-D doses (purity 97.5%) ofO, I, 15, or 45 mg/kg of body weight per day. Increases in absolute and/or relative kidney weights and renal lesions were observed at 15 and 45 mg/kg of body weight per day. The NOAEL was I mg/kg of body weight per day (Serota, 1987). In another 2-year study of toxicity and carcinogenicity, 2,4-D (purity 96.4%) was administered in the diet to groups of B6C3F1 mice (50 per sex per group) at doses of 0, 5, 62, or 120 mg/kg of body weight per day (males) or 0, 5, I 50, or 300 mg/kg of body weight per day (females). Dose-related increases in absolute and/or relative kidney weights and renal lesions were observed in animals at the mid- and high-dose levels in both sexes. The NOAEL was 5 mg/kg of body weight per day (Sott et al., I995). Groups of 50 male and 50 female Fischer 344 rats were fed diets containing technical-grade 2,4-D (purity 97.5%) at doses of 0, I, 5, I5, or 45 mg/kg of body weight per day for 2 years. Renal lesions were observed in animals of both sexes at 5 mg/kg of body weight per day and above. The NOAEL was I mg/kg of body weight per day (Serota, I986). In a further 2-year study, Fischer 344 rats (50 per sex per dose) were fed diets containing 2,4-D (purity 96.4%) at doses of 0, 5, 75, or I 50 mg/kg of body weight per day. Treatment-related effects were observed in males at I 50 mg/kg of body weight per day and in females at 75 mg/kg of body weight per day. The effects included decreases in body-weight gain and food consumption, increases in serum alanine and aspartate aminotransferase activities, decreased thyroxine concentrations, increases in absolute and relative thyroid weights, and histopathological lesions in the eyes, kidneys, liver, lungs, and mesenteric fat. The NOAEL was 75 mg/kg of body weight per day in males and 5 mg/kg of body weight per day in females (Jeffries et al., I995).
5.4 Reproductive and developmental toxicity
In a two-generation study, groups of 30 male and 30 female Fischer 344 rats were fed diets containing 2,4-D (purity 97.5%) at doses of 0, 5, 20, or 80 mg/kg of body weight per day for I05 days before mating (F 0 generation). The rats were dosed in an analogous manner during each mating, each gestation, and each lactation.
Reduced body weight in F1 dams and renal lesions in F0 and F1 adults were observed at 20 and 80 mg/kg of body weight per day. The NOAEL for parental and reproductive toxicity was 5 mg/kg of body weight per day (Rod well, I985). Groups of 15-I9 pregnant Sprague-Dawley rats were given 2,4-D (purity 98.7%) in corn oil by gavage at doses of I2, 25, 50, 75, or 88 mg/kg of body weight per day on days 6-I5 of gestation. Two control groups were used: one for the animals at 88 mg/kg of body weight per day and another for those at the lower doses. There was no maternal toxicity. Fetotoxicity was manifested as decreased fetal body weights at 50 mg/kg of body weight per day and above. The NOAELs were 196 PESTICIDES
88 mg/kg of body weight per day for maternal toxicity and 25 mg/kg of body weight per day for developmental toxicity (Schwetz et al., 1971). Groups of 35 pregnant Fischer 344 rats were given technical-grade 2,4-D (purity 97.5%) in corn oil by gavage at doses ofO, 8, 25, or 75 mg/kg of body weight per day during days 6-15 of gestation. Decreased body-weight gain of the dams during the dosing period and increased incidence of skeletal variations (7th cervical and 14th rudimentary ribs and missing sternebrae) were observed at 75 mg/kg of body weight per day. The NOAEL was 25 mg/kg of body weight per day for both maternal and developmental toxicity (Rodwell, 1983). Rabbits were given 2,4-D (purity 96.1 %) in 0.5% carboxymethylcellulose orally at 0, 10, 30, or 90 mglkg of body weight per day during days 6-18 of gestation. Maternal toxicity, which included clinical signs, abortions, and reduced body-weight gain during and after the treatment period, was observed only at the highest dose. No gross, visceral, or skeletal malformations or variations were observed in the fetuses at any dose. The NOAELs were 30 mg/kg of body weight per day for maternal toxicity and 90 mg!kg of body weight per day (the highest dose tested) for developmental toxicity (Hoberman, 1990).
5.5 Mutagenicity and related end-points
The genotoxic potential of2,4-D has been adequately evaluated in a range of assays in vivo and in vitro. Overall, the responses observed indicate that 2,4-D is not genotoxic, although conflicting results were obtained for mutation in Drosophila (FAO/WHO, 1997a).
5.6 Neurotoxicity
Groups of 10 male and 10 female Fischer 344 rats received 2,4-D in corn oil by gavage as single doses ofO, 15, 75, or 250 mg/kg of body weight. There were no treatment-related gross neuropathological changes at any dose. Animals of both sexes at the highest dose exhibited incoordination and gait abnormalities on day 1, but the signs disappeared by day 5. The NOAEL was 75 mg/kg of body weight (Mattsson et al., 1994a). Fischer 344 rats (15 per group per sex) were fed diets containing 2,4-D at 0, 5, 75, or 150 mg/kg of body weight per day for 12 months. Neurotoxicity, manifested as increased relative forelimb grip strength, was observed in animals of both sexes at 150 mg/kg of body weight per day. The NOAEL was 75 mg/kg of body weight per day (Mattsson et al., 1994b ).
5. 7 Toxicity of salts and esters of 2,4-D
JMPR also evaluated the acute, short-term, and developmental toxicity and mutagenicity of four amine salts of 2,4-D (diethanolamine [DEA], dimethylamine [DMA], isopropylamine [IPA], and triisopropanolamine [TIPA]) and two esters (butoxyethylhexyl [BEH] and 2-ethylhexyl [EH]). JMPR concluded that the toxicity of the salts and esters of2,4-D was comparable to that ofthe acid (FAO/WHO, 1997a). 2,4-DICHLOROPHENOXYACETIC ACID (2,4-D) 197
6. EFFECTS ON HUMANS
Epidemiological studies have suggested an association between exposure to chlorophenoxyacetic acid herbicides, including 2,4-D, and two forms of cancer in humans: soft-tissue sarcomas and non-Hodgkin lymphoma. The results of these studies are not, however, consistent; the associations found are weak, and conflicting conclusions have been reached by the investigators. Most of the studies did not provide information on exposure specifically to 2,4-D, and the risk was related to the general category of phenoxy herbicides, a group that includes 2,4,5-trichlorophenoxyacetic acid (2,4,5-T), which can be contaminated with dioxins. IARC has classified chlorophenoxy herbicides in Group 2B (the agent is possibly carcinogenic to humans) on the basis of limited evidence for carcinogenicity to humans and, for 2,4-D (and 2,4,5-T), inadequate evidence for carcinogenicity to animals (IARC, 1987). Case-control studies provide little evidence of an association between the use of2,4-D and soft-tissue sarcomas. Although some case-control studies have shown a relationship between the use of 2,4-D and non-Hodgkin lymphoma, others (even the positive studies) have produced inconsistent results, raising doubt about the causality of the relationship. Cohort studies of exposed workers have not confirmed the hypothesis that 2,4-D causes either neoplasm (FAO/WHO·, 1997a).
7. GUIDELINE VALUE
JMPR concluded, as did the 1993 Guidelines for drinking-water quality, that it was not possible to evaluate the carcinogenic potential of 2,4-D on the basis of the available epidemiological studies (FAO/WHO 1997a). JMPR established an ADI of 0--0.01 mg/kg of body weight for the sum of 2,4-D and its salts and esters, expressed as 2,4-D, on the basis of the NOAEL of l mg/kg of body weight per day in the 1-year study of toxicity in dogs and the 2-year study of toxicity and carcinogenicity in rats, using an uncertainty factor of 100 (FAO/WHO, 1997b). The resulting guideline value of 30 f!g/litre (using the ADI of 0.01 mg/kg of body weight and assuming a 60-kg body weight, drinking-water consumption of 2 litres/day, and a 10% allocation to drinking-water) is therefore the same as in the 1993 Guidelines for drinking-water quality, but is based on the more recent toxicological evaluation conducted by JMPR. This guideline value applies to 2,4-D, as salts and esters of 2,4-D are rapidly hydrolysed to the free acid in water.
8. REFERENCES
CCME (1995) 2,4-D. In: Canadwn water quu!tty gurdelines Ottawa, Ontario, Canadian Council of Ministers of the Environment. Dalgard DW ( 1993) 5:!-week dietary toxicity study wrth 2,4-D acid in dogs Unpublished report No. 2184-124 from Hazleton Laboratories America, Inc., Vienna, VA. Submitted to WHO by Industry Task Force 11 on 2,4-D Research Data, Indmnapo1is, IN Eiseman J (1984) The pharmacokinetrcs evaluatwn of 14c-lahelled 2.4-D m the mouse. Unpublished report No. 2184-104 from Hazleton Laboratories America. !ne , Vienna, VA. Submitted to WHO by Industry Task Force 11 on 2,4-D Research Data, Indtanapohs. IN. 198 PESTICIDES
FAO/WHO (1997a) Pesticide residues in food- 1996. Joint FAO/WHO Meeting on Pesticide Residues. Evaluations 1996. Part I!- Toxicological. Geneva, World Health Organization, International Programme on Chemical Safety (WHO/PCS/97 .I). FAO/WHO (1997b) Pesticide residues in food -1996. Report of the Joint Meeting of the FAO Panel of Experts on Pesticide Residues in Food and the Environment and the WHO Expert Committee on Pesticide Residues. Rome, Food and Agriculture Organization of the United Nations (FAO Plant Production and Protection Paper 140). Guo M, Stewart S (1993) Metabolism of 14C-rmg labelled 2,4-D in lactating goats. Unpublished report No. 40630 from ABC Laboratories, Inc., Columbia, MO. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, Indianapolis, IN. Health Canada (1993) 2,4-Dichlorophenoxyacetic acid. In: Guidelines for Canadian drmking water quality. Ottawa, Ontario. Ho berm an AM ( 1990) Developmental toxicity (embryo-fetal toxicity and teratogenic potential) study of2,4-dichlorophenoxyacetic acid (2,4-D acid) administered orally via stomach tube to New Zealand white rabbits. Unpublished report No. 320-003 from Argus Research Laboratories, Horsham, PA. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, Indianapolis, IN. IARC (1987) Overall evaluations of carcmogenicity: an updating of !ARC Monographs volumes 1- 42. Lyon, International Agency for Research on Cancer, pp. 156-160 (IARC Monographs
on the Evaluation of Carcinogenic Risks to Humans1 Suppl. 7). Jeffries TK et al. (1995) 2,4-D chronic/oncogenicity study in Fischer 344 rats. Unpublished report No. K-002372-064 from The Dow Chemical Company, Midland, MI. Submitted to WHO by Industry Task Force 11 on 2,4-D Research Data, Indianapolis, IN. Mattsson JL, McGuirk RJ, Yano BL (1994a) 2,4-D acute neurotoxicity study in Fischer 344 rats. Unpublished report No. K-002372-066 from The Dow Chemical Company, Midland, MI. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, Jndianapolis, IN. Mattsson JL, Jeffries TK, Yano BL (1994b) 2,4-Dichlorophenoxyacetlc acid chronic neurotoxicity study in Fisc her 344 rats. Unpublished report No. K-002372-064N from The Dow Chemical Company, Midland, MI. Submitted to WHO by Industry Task Force on 2,4-D Research Data, Indianapolis, IN. Rodwell DE (1983) A teratology study in F1scher 344 rats with 2,4-D acid. Unpublished report No. WIL-81135 from WIL Research Laboratories, Inc., OH Submitted to WHO by Industry Task Force II on 2,4-D Research Data, Indianapolis, IN. Rodwell DE (1985) A d1etary two-generation reproduction study m F1scher 344 rats with dichlorophenoxy acetic acid. Unpublished report No. WIL-81137 from WIL Research Laboratories, Inc., OH. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, IndJanapolis, IN. Sauerhoff MW et al. (1977) The fate of 2,4-dichlorophenoxyacetic acid (2,4-D) following oral administration to man. Toxicology, 8:3-11. Schultze GE (1990) Subchronic toxicity study m dogs with 2,4-dlchlorophenoxyacetic acid Unpublished report No. 2184-115 from Hazleton Laboratories America, Inc., Vienna, VA. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, Indianapolis, IN. Schultze GE (199la) Subchronic toxicity study m mice w1th 2,4-D acid. Unpublished report No. 2184-117 from Hazleton Laboratories America, Inc., Vienna, VA. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, Indianapolis, IN. Schultze GE (1991b) Subchronic toxicity study in rats with 2,4-D ac1d. Unpublished report No. 2184- 116 from Hazleton Laboratories America, Inc., Vienna, VA. Submitted to WHO by Industry Task Force 11 on 2,4-D Research Data, Jndianapolis, IN. Schwetz BA, Sparschu GL, Gehring PJ (1971) The effect of 2,4-dichloro-phenoxyacetic acid (2,4-D) and esters of 2,4-D on rat embryonal, foetal, and neonatal growth and development. Food and cosmet1cs toxicology, 9·801-817. Serota DG (1983) Subchronic tox1city study in rats with 2,4-D acid Unpublished report No 2184- 102 from Hazleton Laboratories America, Inc., Vienna, VA. Submitted to WHO by Industry Task Force 11 on 2,4-D Research Data, Indianapolis, IN. 2,4-DICHLOROPHENOXYACETIC ACID (2,4-D) 199
Serota DG (1986) Combined toxicity and oncogenicity study in rats: 2,4-dichlorophenoxyacetic acid (2,4-D). Unpublished report No. 2184-103 from Hazleton Laboratories America, Inc., Vienna, VA. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, Indianapolis, IN. Serota DG (1987) Oncogenicity study in mice with 2,4-dichlorophenoxyacetic acid (2,4-D). Unpublished report No. 2184-101 from Hazleton Laboratories America, Inc., Vienna, VA. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, lndianapolis, IN. Smith FA et al. (1990) Pharmacokinetics of 2,4-dichlorophenoxyacetic acid (2,4-D) in Fischer 344 rats. Unpublished report No. 0697 from The Dow Chemical Company, Midland, MI. Submitted to WHO by Industry Task Force 11 on 2,4-D Research Data, Indianapolis, IN. Sott WT et al. (1995) 2,4-Dichlorophenoxyacetic acid: dietary oncogenicity study in B6C3FJ mice two year final report. Unpublished report Nos. K-002372-063M and K-002372-063F from The Dow Chemical Company, Midland, Ml. Submitted to WHO by Industry Task Force II on 2,4-D Research Data, lndianapolis, IN. Tomlin C, ed. (1994} The pesticide manual: a world compendium, lOth ed. Farnham, British Crop Protection Council. US EPA (1987) Health advisory: 2,4-Dich/orophenoxyacetic ac1d. Washington, DC, US Environmental Protection Agency. WHO (1984) 2,4-Dichlorophenoxyacetic acid (2,4-D). Geneva, World Health Organization (Environmental Health Criteria 29). WHO (1996) The WHO recommended classification of pesticides by hazards and guidelines to classification. Geneva, World Health Organization, International Programme on Chemical Safety (WHO/PCS/96.3).
1,2-DICHLOROPROPANE (1,2-DCP)
First draft prepared by J.E.M. van Koten-Vermeulen National Institute ofPublic Health and Environmental Protection Bilthoven, The Netherlands
A provisional guideline value of 20 ~-tg/litre for 1,2-dichloropropane (1,2-DCP) was recommended in the second edition of the WHO Guidelines for drinking-water quality. The provisional guideline value was based on a LOAEL identified on the basis of a variety of systemic effects in a 13-week oral study in rats and an uncertainty factor of 10 000 (1 00 for inter- and intraspecies variation, 10 for the use of a LOAEL, and 10 to reflect limited evidence of carcinogenicity in animals and a limited toxicity database, particularly for reproductive studies). The guideline value was designated provisional because of the limited database and the use of the large uncertainty factor in its calculation. Because of the provisional guideline value and because an IPCS Environmental Health Criteria monograph for this pesticide was published in 1993, the Coordinating Committee for the updating of WHO Guidelines for drinking-water quality considered 1,2-DCP to be a high-priority chemical for inclusion in the 1998 Addendum.
1. GENERAL DESCRIPTION
1.1 Identity
CAS no.: Molecular formula:
1.2 Physicochemical properties (WHO, 1993/
Property Value Boiling point 96.8°C Melting point -100°C Density 1.1595 g/cm3 at 20°C Vapour pressure 27.9 kPa at 19.6°C Water solubility 2700 mg/litre at 20°C Log octanol-water partition coefficient 1.99
1.3 Organoleptic properties
1,2-DCP has a chloroform-like odour. The odour threshold in water is 10 jlg/litre.
3 Conversion factor in air: 1 mg/m = 0.214 ppm. 202 PESTICIDES
1.4 Major uses
1,2-DCP is used primarily as an intermediate in the production of perchloroethylene and other chlorinated products (ATSDR, 1989). It is also used as a solvent for fats, oils, resins, waxes, and rubber, as an insecticide fumigant on grain and soil, and to control peach tree borers. Other uses are as dry cleaning fluid, paint remover, metal degreasing agent, and lead scavenger for antiknock fluids. 1,2-DCP is also a component of"MIX DID," used as a pre-plant fumigant (WHO, 1993, 1996).
1.5 Environmenta/fate
The decomposition of 1,2-DCP in the atmosphere is rather slow; on the basis of reaction with hydroxyl radicals, the half-life of 1,2-DCP was >23 days. Phototransformation is likely to be the predominant process for the decomposition of 1,2-DCP, but vapour-phase photolysis was not detected after prolonged simulated sunlight irradiation in a reaction chamber (WHO, 1993, 1996). In water, 1,2-DCP is relatively resistant to hydrolysis and has a half-life of 175-1400 days (WHO, 1993). No biodegradation was observed in a semi-continuous activated sludge process, and there was also no biodegradation in standard 4-week tests that simulate biodegradation in environmental waters (ATSDR, 1989). Volatilization is likely to be the major route of loss from water. The relatively low soil adsorption coefficient as well as the high water solubility suggest that 1,2-DCP is not appreciably adsorbed onto soil but migrates from it to groundwater. 1,2-DCP is persistent in soil. More than 98% of the 1,2-DCP applied to loam sot! was recovered 12-20 weeks after treatment. Because of its high water solubility and low octanol-water partition coefficient (log Kow), bioaccumulation is unlikely to occur (WHO, 1993, 1996). A bioconcentration factor of 18 is estimated, indicating that there is very low potential for bioaccumulation in the food-chain (ATSDR, 1989).
2. ANALYTICAL METHODS
1,2-DCP is usually determined by a purge-and-trap gas chromatographic method used for the determination of volatile organohalides in drinking-water. Subsequent thermal desorption is used for its quantification. Confirmatory analysis is done by halide-specific detectors (e.g. Hall electrolytic conductivity detector) and mass spectrometry (detection limit 0.02--0.17 J.Lg/litre) (Lyman et al., 1982; WHO, 1996).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
1,2-DCP was monitored in 13 cities in Japan. Levels in ambient air were reported to range from 0.0065 to 1.4 J.Lg/m3 (WHO, 1993). Levels in 1982 were reported to be non-detectable in rural/remote areas, 0.26 J.Lglm 3 in urban/suburban areas, and 0.55 J.Lglm3 in source-dominated areas (WHO, 1996). An average 1,2-DICHLOROPROPANE (1,2-DCP) 203 concentration of 0.85-3.10 f.lg/m 3 in ambient air was detected in 1981-1982 and 1986--1987 in cities in Germany (BUA, 1996).
3.2 Water
In a 1983 study in the USA, 1,2-DCP was found in wells at concentrations up to 440 1-lg/litre. In another study, 1,2-DCP concentrations as high as 51 J.lg/litre were found in groundwater (WHO, 1993). In Japan in 1989, 1,2-DCP was detected in 20 out of 78 water samples (river, lake, port, and bay areas), with levels ranging from 0.000 01 to 0.14 J.lg/litre (WHO, 1993). Levels of 0.7-19.0 J.lg/litre were found in potable water samples collected in 1990 in eight homes in three communities in Connecticut, USA (WHO, 1993). In the Netherlands in 1995, 1,2-DCP was found in raw drinking-water in 13 out of 169 supplies (range 0.06-2.1 J.lg/litre) and in drinking-water in 11 out of 52 groundwater supplies (range 0.05-0.12 J.lg/litre) (Rijkinstitut voor Volksgezondheid en Milieuhygiene [RIVM], Bilthoven, personal communication, 1996). In a waterworks monitoring program in Germany in 1990, 1,2-DCP concentrations of 0.3 J.lgllitre were found in drinking-water from Schleswig-Holstein (BUA, 1996).
3.3 Estimated total exposure and relative contribution of drinking-water
3 At an air concentration of 0.26 J.lg/m , the exposure to I ,2-DCP will be 5.2 J.lg/day for an adult, assuming an air intake of 20 m3/day. Using the recently measured drinking-water concentrations ranging from 0.05 to 19 J.lg/litre, the maximal daily exposure for an adult consuming 2litres of water per day is 38 J.lg.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
After oral as well as inhalation exposure, 1,2-DCP is rapidly absorbed, metabolized, and excreted. The majority of the radioactivity is excreted within 24 hours. Elimination after oral or inhalation exposure occurs mainly via the urine (37-52%) and expired air (37--40%). It has been suggested that in rats, 1,2-DCP is dechlorinated and oxidized to epoxide intermediates, which are consequently hydrolysed and conjugated to form N-acety1-S-(2-hydroxypropyl) cysteine. Three N-acetylcysteine conjugates of dichloropropane were identified as the major urinary metabolites after oral or inhalation exposure (Timchalk et al., 1989; WHO, 1996).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
1,2-DCP has a low acute toxicity in experimental animals. In rats, oral LD50 values ranged from 1000 to 2000 mg/kg of body weight. The 8-hour LC50 for 3 inhalation is 9520 mg/m (ATSDR, 1989). The dermal LD 50 after a single application 204 PESTICIDES to the skin of rabbits was 8.75 mglkg of body weight (WHO, 1993, 1996). Single oral doses of 1,2-DCP caused salivation, lacrimation, dyspnoea, lethargy, reduced motility, haemorrhage in the gastrointestinal tract, haemolytic anaemia, and liver and kidney damage. 1,2-DCP does not cause any skin irritation in rabbits but is slightly irritating to the rabbit eye (BUA, 1996).
5.2 Short-term exposure
When rats were exposed to 1,2-DCP in corn oil at 0, 100, 250, 500, or 1000 mg/kg of body weight per day for 1, 5, or 10 days, the liver was the primary target organ. Central nervous system depression, decreased body-weight gain, and increased renal non-protein sulfhydryllevels were seen at 250 mglkg of body weight per day and higher. Liver damage, as reflected by enzyme changes and morphological alterations, was observed at 500 and I 000 mglkg of body weight per day. However, rats demonstrated an adaptive resistance to 1,2-DCP over 10 consecutive days of exposure, resulting in hepatic lesions being less severe at 10 days than at 5 days. However, nuclear enlargements in hepatocytes were observed at all dose levels at 5 and 10 days (Bruckner et al., 1989). In a 13-week study in which male rats were administered 1,2-DCP by gavage in corn oil at 0, 100, 250, 500, or 750 mg/kg of body weight per day for 5 days/week, increased mortality occurred at 500 and 750 mg/kg of body weight per day. Significant growth retardation was observed at all dose levels. Effects indicating haemolytic anaemia, such as decreased haematocrit, decreased haemoglobin concentration, increased serum bilirubin, haemosiderosis and hyperplasia of the erythropoietic elements of the spleen, and increased liver and spleen weights, were observed. At the lowest dose of 100 mg/kg of body weight per day, haemosiderosis and splenic hyperplasia were still present. Relative spleen and liver weights were increased at doses :2:250 mg/kg of body weight per day. Upon histopathology, changes in the liver were observed, such as mild hepatitis at 750 mg/kg of body weight per day and periportal vacuolization and active fibroplasia at 500 mg/kg of body weight per day. At 500 and 750 mg/kg of body weight per day, testicular degeneration, including reduced sperm production, increased numbers of degenerate sperm, and reduced numbers of sperm in the epididymis, occurred. A NOAEL could not be established in this study. The LOAEL was 100 mg/kg of body weight per day (71.4 mg/kg of body weight per day when corrected for 5 days/week dosing) (Bruckner et al., 1989). In 13-week studies in mice and rats, 1,2-DCP was administered by gavage, in corn oil, at dose levels of 0, 30 (mice only), 60, 125, 250, 500, or 1000 (rats only) mg/kg of body weight per day for 5 days/week. Increased mortality was observed in rats at 500 (50%) and 1000 (lOO%) mg/kg of body weight per day. At 500 mg/kg of body weight per day, mean body weights were decreased 16% in males and 8% in females at the end of the study. High-dose rats showed congestion of the liver in males and females and necrosis and fatty changes in the liver in females. No mortality occurred in mice, and only marginal body-weight depression was seen at the highest dose. No dose-related histological effects were seen (NTP, 1986). 1,2-DICHLOROPROPANE (1,2-DCP) 205
In a neurotoxicity study, F344 rats were administered 1,2-DCP by gavage at dose levels of 0, 20, 65, or 200 mg!kg of body weight per day, 5 days/week for 13 weeks. After 13 weeks, extensive neurohistopathology as well as histopathology of liver, kidneys, and spleen were performed on four rats per group. The remaining 11 rats per group were observed for another 9 weeks, after which 5 rats per group were subjected to gross pathological examination. Male body weight was reduced in rats at 200 mg/kg of body weight per day; the reduction was still evident at the end of the recovery period. No substance-related changes were found in the results of the neurological tests or of the neurohistopathological examination (Johnson & Gorzinski, 1988).
5.3 Long-term exposure
Long-term toxicological studies with 1,2-DCP are not available. The available carcinogenicity studies with mice and rats are described below (see section 5.6).
5.4 Reproductive and developmental toxicity
Testicular degeneration and an increased number of degenerate spermatogonia in the epididymis were seen in the 13-week oral gavage study in rats described above (see section 5.2) at doses of 500 and 750 mg/kg of body weight per day. The NOAEL for testicular toxicity in this study was 250 mg/kg of body weight per day ( 179 mg/kg of body weight per day when corrected for 5 days/week dosing) (Bruckner et al., 1989). In an oral two-generation study, rats were given 1,2-DCP in drinking-water at concentrations of 0, 240, 1000, or 2400 mg/litre (equivalent to 0, 33.6, 140, or 336 mg/kg of body weight per day). A decrease in water consumption (50% less than controls) as well as decreased body-weight gain in parental males and females
(both generations) were observed at the highest dose. In F0 females, red blood cell counts, haemoglobin concentration, and haematocrit were decreased at 2400 mg/litre. In parental males and females, increased hepatocellular granularity was observed at all dose levels. At 2400 mg/litre, neonatal body weight was decreased and neonatal mortality was slightly increased. No treatment-related changes in the reproductive organs or in other reproductive indices, including fertility, mating index, conception index, viable litters, gestation length, litter size, live pups, and sex ratio, were noted. The NOAEL in this study was 1000 mg/litre (equivalent to 140 mg!kg of body weight per day). According to the authors, the increased hepatocellular granularity is considered to be an adaptive change associated with the metabolism of 1,2-DCP and is not an indication of toxicity. Moreover, the US EPA (Kirk et al., 1990) as well as the IPCS (WHO, 1993) mentioned the observed effect but established the same NOAEL of 1000 mg/litre for this study. Teratogenicity studies are available in rats and rabbits. Rats were administered 1,2-DCP by gavage at doses of 0, 10, 30, or 125 mg/kg of body weight per day on days 6 through 15 of gestation. Toxic effects were seen in dams at the highest dose level (decreases in growth, food consumption, muscle tone, and 206 PESTICIDES
extensor flux reflex). In fetuses, the incidence of delayed ossification of skull bones was increased at 125 mg/kg of body weight per day. The NOAEL for maternal and fetal effects in the study was 30 mg/kg of body weight per day (Kirk et al., 1995). Rabbits were dosed by gavage at 0, 15, 50, or 150 mg/kg of body weight per day on days 7 through 19 of gestation. Effects similar to those in the rat study were found: maternal toxicity (anorexia, anaemia) at 150 mg/kg of body weight per day and increased incidence of delayed ossification of skull bones in fetuses at 150 mg/kg of body weight per day. The NOAEL for maternal and fetal effects in rabbits was 50 mg/kg of body weight per day (Kirk et al., 1995). Although maternal toxicity was apparent in these studies, no indication of teratogenicity was observed in rat or rabbit fetuses at any dose level. Significant increases in the incidence of delayed ossification of skull bones are considered secondary to decreased maternal body-weight gain.
5.5 Mutagenicity and related end-points
In in vitro mutagenicity studies, 1,2-DCP was positive m Salmonella typhimurium strains TA100 and TA1535, both with and without metabolic activation. In the S. typhimurium strains TA98, TA1537, and TA1538, no mutagenicity of I ,2-DCP could be detected (BUA, 1996; WHO, 1996). These results indicate that 1,2-DCP can cause base pair substitution. 1,2-DCP was mutagenic in mouse lymphoma cells in the thymidine kinase test and in an in vitro test in Aspergillus nidulans (forward mutation to 8-azaguanine resistance) (Myhr & Caspary, 1991; WHO, 1996). Negative results were obtained in a test studying the induction of somatic segregation in Aspergillus nidulans (Crebelli et al., 1984). Other in vitro studies for sister chromatid exchanges in Chinese hamster ovary cells and V79 cells were positive both without and with metabolic activation, as was a test for chromosomal aberrations in Chinese hamster ovary cells (Galloway et al., 1987; von der Hude et al., 1988). A test for DNA repair inS. typhimurium TA1535 (umu test) yielded negative results both with and without metabolic activation (BUA, 1996). In the SOS chromotest with Escherichia coli strain PQ37, 1,2-DCP was not genotoxic in concentrations up to the solubility limit (~2700 mg/litre), either with or without S9 mix (von der Hude et al., 1988). 1,2-DCP (0.1-10 mmol/litre) did not induce unscheduled DNA synthesis in human lymphocytes (Perocco et al., 1983). Negative results were obtained in two in vivo studies: a test for sex-linked recessive lethal mutations in Drosophila melanogaster with oral and inhalatory administration and a dominant lethal test in Sprague-Dawley rats, dosed for 14 weeks via drinking-water (WHO, 1993).
5. 6 Carcinogenicity
Two carcinogenicity studies are available (NTP, 1986). Groups of male and
female B6C3F1 mice and female F344/N rats were administered 1,2-DCP by gavage at 0, 125, or 250 mg!kg of body weight per day, 5 days/week for 103 weeks. Groups of 1,2-DICHLOROPROPANE (1,2-DCP) 207 male rats were administered 0, 62, or 125 mg/kg of body weight per day according to the same protocol. Mortality was increased in high-dose female mice and rats. An increased incidence of liver lesions (hepatomegaly, focal and centrilobular necrosis) was observed in male mice at both dose levels. The incidence of hepatocellular tumours (combined adenomas and carcinomas) in treated groups of mice was higher than that in the concurrent control groups, but was within the historical control range. Female rats in the highest dose group showed decreased survival, and body weight was decreased in both male (I4%) and female rats (24%) at the highest dose. In the same group, the incidence ofliver lesions (foci of clear cells and necrosis) was increased. The incidence of mammary gland adenocarcinomas was slightly increased in female rats (1/50, 2/50, and 5/50 in the control, low-dose, and high-dose groups, respectively). There were no effects on tumour incidences in male rats.
6. EFFECTS ON HUMANS
Several cases of acute poisoning resulting from accidental or intentional (suicide) overexposure to 1,2-DCP have been reported. Effects have been mainly on the central nervous system, liver, and kidneys. Haemolytic anaemia and disseminated intravascular coagulation as well as effects on the respiratory system, heart, and blood have also been described (WHO, I993, I996). Several cases of dermatitis and skin sensitization following occupational dermal exposure to solvent mixtures containing I ,2-DCP have been reported (WHO, I993, I996).
7. PROVISIONAL GUIDELINE VALUE
I,2-DCP was evaluated by IARC in I986 and I987 based on NTP studies. The substance was classified in Group 3 (not classifiable as to its carcinogenicity to humans) based on limited evidence for its carcinogenicity to experimental animals and an inability to evaluate its carcinogenicity to humans. Results from in vitro assays for mutagenicity were mixed. The in vivo studies, which were limited in number and design, were negative. In accordance with the IARC evaluation, the evidence from the long-term carcinogenicity studies in mice and rats was considered limited, and it was concluded that the use of a threshold approach for the toxicological evaluation of I ,2-DCP was appropriate. The 13-week toxicity study in which male rats were administered 1,2-DCP by gavage in corn oil for 5 days/week was considered to be the most appropriate study for derivation of a guideline value. A LOAEL of 100 mg/kg of body weight per day (71.4 mg/kg of body weight per day when corrected for 5 days/week dosing) was observed for changes in haematological parameters. Using an uncertainty factor of 5000 (I 00 for inter- and intraspecies variation, 10 for the use of a LOAEL instead of a NOAEL, and 5 for limitations of the database, including the limited data on in vivo genotoxicity and use of a subchronic study), a TDI of 14 !J.g/kg of body weight is derived. With an allocation of 10% of the TDI to drinking-water and assuming a 60-kg body weight and drinking-water consumption of 2 litres/day, the provisional guideline value is 40 !J.g/litre (rounded figure). The guideline value is considered to 208 PESTICIDES
be provisional owing to the magnitude of the uncertainty factor and the fact that the database has not changed since the previous guideline value was derived.
8. REFERENCES
A T~DR (1989) Toxicological profile for 1.2-dtchloropropane Atlanta, GA, US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry (TP-89/12). Bruckner JV et al. (1989) Oral toxicity of I ,2-dichloropropane. acute, short-term, and long-term studies in rats. Fundamental and applied toxicology, 12:713-730. BUA (1996) 1,2-Dichloropropane GDCh-Advisory Committee on Existing Chemicals of Environmental Relevance, Beratergremium fur Umweltrelevante Altstoffe. Stuttgart, S. Hirzel, Wissenschaftliche Verlagsgesellschaft (BUA Report 155, October 1994). Crebelli R et al. (1984) Induction of somatic segregation by halogenated aliphatic hydrocarbons in Aspergillus nidulans. Mutation research, 138.33-38. Galloway SM et al. (1987) Chromosome aberrations and sister chromatid exchanges in Chinese hamster ovary cells: Evaluation of I 08 chemicals. Environmental and molecular mutagenests, I 0 (Suppl. ): 1-175. !ARC (1986) Some halogenated hydrocarbons and pesticide exposures Lyon, International Agency for Research on Cancer (!ARC Monographs on the EvaluatiOn of Carcinogenic Risk of Chemicals to Humans, Vol. 41 ). !ARC (1987) Overall evaluations of carc111ogemclfy an updat111g of !ARC Monographs volumes 1- 42 Lyon, International Agency for Research on Cancer (IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Suppl. 7). Johnson KA, Gorzinski SJ (1988) Neurotoxicologic exam111atwn of Ftscher 34.J rats exposed to 1,2- ·dich/oropropane (DCP) vta gavage for 13 weeks Unpublished report (dated 16 November 1988) prepared by Dow Chemical Company, Mammalian and Environmental Toxicology Research Laboratory, Midland, Ml. Kirk HO et al. (1990) Propylene d1chloride two-generatiOn reproductwn study 111 Sprague-Dawley rats Unpublished propnetary report submitted to WHO by Dow Chemical Company, Mammalian and Environmental Toxicology Research Laboratory, Midland, Ml [cited in WHO, 1993]. Kirk HO et al. (1995) Developmental toxicity of I ,2-dichloropropane (PDC) in rats and rabbits following oral gavage. Fundamental and apphed toxicology, 28:18-26. Lyman WH, Reehle WF, Rosenblatt OH ( 1982) Handbook of chenucal property esllmalton methods. New York, NY, McGraw-Hill Book Co. Myhr BC, Caspary WJ (1991) Chemical mutagenesis at the thymidine kinase locus in L5178 Y mouse lymphoma cells: results for 31 coded compounds in the National Toxicology Program. Environmental and molecular mutagenesis, 18:51 [cited in BUA, 1996]. NTP (1986) Tox1cology and carcinogenesis studtes of 1,2-dichloropropane (propylene d1chlonde) (CAS No 78-87-5) in F34.J/N rats and B6C3F1 mice (gavage studtes) Research Tnangle Park, NC, US Department of Health and Human Services, NatiOnal Toxicology Program (Technical Report No. 263). Perocco P, Bolognesi S, Alberghim W (1983) Toxic activity of seventeen industrial solvents and halogenated compounds on human lymphocytes cultured 111 vttro. Toxicology letters, 16·69-75. Timchalk C et al. (1989) Propylene dichlonde pharmacokinetics and metabohsm 111 F1scher 344 rats fo/lowll1g oral and inhalation exposure Unpublished proprietary reported submitted to WHO by Dow Chemical Company, Mammalian and Environmental Toxicology Research Laboratory, Midland, Ml [cited in US Environmental Protection Agency Integrated Risk Information System (IRIS), 1996 (IRIS Accession No. 601)] von der Hude Wet al. (1988) Evaluation of the SOS chromotest. Mutatton research, 203:81. WHO (1993) 1.3-Dichloropropene, 1,2-dlchloropropane and miXtures Geneva, World Health Orgamzat10n, International Programme on Chemical Safety (Environmental Health Cntena 146). WHO (1996) Guidelmes for drmk111g-water quality, 2nd ed. Vol 2 Health criteria and other support111g mformatwn Geneva, World Health Organization. DIQUAT
First draft prepared by H. Ga/a/-Gorchev International Programme on Chemical Safety World Health Organization, Geneva, Switzerland
Diquat was not evaluated in the second edition of the WHO Guidelines for drinking-water quality. However, it was recommended for evaluation by the 1992 Final Task Group Meeting, and JMPR established an ADI for diquat in 1993.
1. GENERAL DESCRIPTION
1.1 Identity
Ion Dibromide CAS no.: 2764-72-9 85-007 Molecular formula: C12H12N2 C 12Ht2Br2N2
The IUPAC name for diquat is 9,10-dihydro-Sa,lOa-diazoniaphenanthrene ion. Diquat is sold as diquat dibromide and is usually formulated as an aqueous solution (270 g of diquat ion per litre) (FAO/WHO, 1995).
1.2 Physicochemical properties (FAO/WHO, 1995)
Physicochemical properties for the pure active ingredient (diquat dibromide) are as follows:
Property Value Vapour pressure
1.3 Major uses
Diquat is a non-selective contact herbicide and crop desiccant. On a global basis, pre-harvest desiccation to aid the harvesting of seed and fodder crops accounts for the use of two-thirds of the global volume of diquat, whereas one-third of the diquat sold is used as a weed killer. The regions of North America, Europe, Australia, and Japan consume 90% of the diquat used for herbicidal and crop desiccation purposes. Diquat may also be used (at or below 1 mg/litre) as an aquatic 210 PESTICIDES
herbicide for the control of free-floating and submerged aquatic weeds in ponds, lakes, and irrigation ditches (FAO/WHO, 1995).
1.4 Environmentalfate
Diquat is strongly adsorbed to soil, is not taken up by plant roots, and is not metabolically degraded by plants. The rate of degradation in soil, although slow, was found to be sufficient to ensure that diquat residues would not accumulate indefinitely in soil but would reach a plateau level when the amount degraded each year was equal to the amount of new addition. In the presence of sunlight, rapid and extensive photochemical degradation occurs. Diquat does not bioaccumulate in food. The potential for diquat to leach into potable water was tested using a model pond-soil-aquifer system (an extremely sandy soil with low adsorption capacity was used as the soil in the system). No diquat (<0.003 mg/litre) was found in any aquifer sample. Photodegradation of diquat is extensive in water. The major degradation product is 1,2,3,4-tetrahydro-1-oxopyrido[l ,2a]-5-pyrazinium ion (TO PPS). Diquat monopyridone is formed to only a limited extent. On further irradiation, TOPPS is degraded to picolinamide, picolinic acid, formic acid, oxalic acid, carbon dioxide, and other volatile fragments. Picolinamide in water is also known to undergo bacterial oxidation with ring opening to form maleic and fumaric acids. When diquat is added to natural waters to control aquatic weeds, residues in the water rapidly decline, owing mainly to the absorption of diquat into the aquatic plants, where it is firmly bound until the decaying weeds disintegrate into the bottom mud. The diquat is then irreversibly bound to the soil particles, leaving the water free of diquat residues. Half-lives of diquat in natural waters are generally less than 48 hours (FAO/WHO, 1995).
2. ANALYTICAL METHODS
Analytical methods are based on extraction of diquat by acid hydrolysis and clean-up and concentration by ion exchange chromatography followed by reduction and measurement of the diquat reduction products by gas-liquid chromatography with a nitrogen-phosphorus detector. The limits of determination are 4 11g/litre in water, 0.01 mg/kg in soil, and 0.02 mg/kg in animal tissues and food crops (FAO/WHO, 1995). In other methods, diquat residues have been determined spectrophotometrically with a limit of detection of 1 11g/litre in water (Earl & Boseley, 1988).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Because of its extremely low vapour pressure, diquat levels in air are expected to be low. After spraying, diquat levels decreased from 0.6 mg/m3 in the immediate spray area (tractor cabin) to 0.06 mg/m3 in a treated field 10 minutes after DIQUAT 211 spraying; diquat was not detectable 20 minutes after spraying. No diquat was detected at a distance of 400 m from the treated field (WHO, 1984).
3.2 Water
Groundwater was analysed for diquat at two sites in Japan where the product had been used commercially for 5 and 15 years. No diquat was detected in the water, the limit of detection being 0.1 mg/litre (FAO/WHO, 1995). Following its use as an aquatic herbicide at normal application rates, diquat residues in water have been found to decrease rapidly to essentially undetectable levels within 7-14 days (WHO, 1984).
3.3 Food
When diquat is used as a herbicide to control weeds, no residues (<0.05 mg/kg) are found in the harvested crop. When diquat is used as a desiccant, the product is sprayed directly onto the crop, and significant residues are present in the crop at harvest. Diquat concentrations of <0.02-0.7 mg/kg have been reported in beans, lentils, peas, potatoes, and other food crops harvested 3-21 days after spraying (FAO/WHO, 1995).
3.4 Estimated total exposure and relative contribution of drinking-water
Because of diquat's rapid degradation in water and strong adsorption onto sediments, the contribution of drinking-water to total exposure is expected to be low. Major sources of exposure are expected to be air in the occupational setting and food for the general population.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
When administered orally, 14 C-diquat was poorly absorbed from the gastrointestinal tract of rats, cows, and goats and was eliminated mainly via the faeces during the first 24 hours (>90% of the administered doses); the small fraction that was absorbed was eliminated principally via the urine (Hemingway et al., 1974; Griggs & Davis, 1975; Johnston et al., 1990). After oral administration of diquat to rats, goats, and cows, the major excreted product was diquat in both faeces and urine. Minor metabolites included diquat monopyridone and diquat dipyridone, as well as TOPPS, picolinamide, and picolinic acid, which undergo further degradation and natural incorporation (Black et al., 1966; Griggs & Davis, 1975; Williams et al., 1991). Diquat is poorly absorbed through human skin in vitro and in vivo (Feldmann & Maibach, 1974; Scott et al., 1991). No other studies on the kinetics of diquat in volunteers have been published, but observations are available on accidental and suicidal ingestion. Toxicological analysis of the serum of a patient who had ingested 20 m! of "Reglone" (an aqueous formulation containing about 200 g of diquat ion 212 PESTICIDES per litre) showed a diquat level of 0.4 mg/litre at the time of admission to the hospital. At postmortem examination on the fifth day after ingestion, a diquat concentration of approximately 0.2 mg/kg was measured in liver, kidneys, muscle, and eye liquid (WHO, 1984).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
Median oral lethal doses (LD50s) of diquat range from 100 to 300 mg of ion per kg of body weight for mice, rats, rabbits, guinea-pigs, dogs, hens, and monkeys; a median oral LD50 of 30 mg of ion per kg of body weight has been estimated for cows (FAO/WHO, 1994). Diquat has been classified by WHO as moderately hazardous (IPCS, 1996).
5.2 Short-term exposure
In a 90-day feeding study, groups of 12 male and 12 female AlpK:APfsD rats received dietary concentrations of 0, 20, 100, or 500 mg of diquat ion per kg of feed. The NOAEL was 100 mg/kg of feed (equal to 8.5 mg of diquat ion per kg of body weight per day for males and 9.2 mg of diquat ion per kg of body weight per day for females), based upon cataract formation and reduction in body-weight gain, food consumption, and plasma protein at the next higher dose (Hodge, 1989a). In a 1-year study, groups of four male and four female beagle dogs received doses of 0, 0.5, 2.5, or 12.5 mg of diquat ion per kg of body weight per day. The NOAEL was 0.5 mg of diquat ion per kg of body weight per day, based upon lens opacity in females at the next higher dose (Hopkins, 1990).
5.3 Long-term exposure and carcinogenicity
In an 80-week feeding study, groups of CD-1 mice (60 per sex) were fed diquat dibromide (98.5% pure) at dietary concentrations of 30, 150, or 500 mg/kg of feed. Because of lethal toxicity, the highest dietary concentration was reduced to 400 mg/kg of feed after 3 weeks and to 300 mg/kg of feed after 5 weeks, and the dead mice were replaced. Reduced growth rates (both sexes) and increased frequency of hepatic vacuolation in males were observed at 150 mg/kg of feed. The NOAEL was 30 mg/kg of feed, equivalent to 4.5 mg/kg of body weight per day (Ashby, 1987). In a 2-year study, groups of 60 male and 60 female mice (C57BL/10fCD- 1Alpk) were fed a diet containing 0, 30, I 00, or 300 mg of diquat per kg of feed. The test material was technical-grade diquat dibromide, containing 26.7% (w/v) diquat ion. No treatment-related carcinogenic effects were seen in the study (there were reductions in the incidence of certain tumours at 300 mg/kg of feed), and the incidence of cataract formation was not related to the test material. The NOAEL was 30 mg/kg of feed, equal to 3.6 mg of diquat ion per kg of body weight per day, based DIQUAT 213 on reduction in body-weight gain and increased relative kidney weights at higher doses (Hodge, 1992). In a 2-year feeding study, groups ofWistar-derived rats (35 per sex) received diquat dibromide (lOO% pure= 53.6% ion) at dietary concentrations of 0, 15, 25, or 75 mg/kg of feed. The NOAEL was 25 mg/kg of feed, equivalent to 1.3 mg/kg of body weight per day, based on cataract formation at the highest dose (Rogerson & Broad, 1978). In another 2-year study, diquat dibromide (technical grade) was administered in the diet to groups of Sprague-Dawley rats (60 per sex) at concentrations of 0, 5, 15, 75, or 375 mg/kg of feed. The LOAEL was 15 mg!kg of feed (equal to 0.58 and 0.72 mg of diquat ion per kg of body weight per day for males and females, respectively) based upon cataract formation. The NOAEL was 5 mg!kg of feed, equal to 0.19 mg of diquat ion per kg of body weight per day (males) and 0.24 mg of diquat ion per kg of body weight per day (females). There was no evidence of carcinogenicity in rats (Colley et al., 1985).
5.4 Reproductive and developmental toxicity
A three-generation reproductive toxicity study was conducted in Wistar derived rats using diquat dibromide monohydrate (100% pure). Three groups (12 males and 24 females per group) received diquat dibromide in aqueous solution at dietary concentrations of 0, 125. or 500 mg!kg of feed when they were approximately 35 days old. Thereafter, the three groups and their progeny remained on the diet throughout the study. A NOAEL could not be determined, as there was a
(trivial) decreased weight gain in F0 and F 1 animals at the lowest dose tested (125 mg/kg of feed, equivalent to 6.3 mg of diquat ion per kg of body weight per day) (Fletcher et al., 1972). In a multigeneration reproductive toxicity study, groups of Alpk:APfSD rats (30 per sex) were fed diets containing 0, 16, 80, or 400 mg of diquat per kg of feed. The top dose received by the F, parents was reduced after 4 weeks to 240 mg/kg of feed. Although cataract formation was mostly confined to the 240 mg/kg of feed group. a low incidence was seen at 80 mg/kg of feed in the F 1 female parents. The NOAEL was 16 mg/kg of feed, equivalent to 0.8 mg/kg of body weight per day (Hodge, 1990). Numerous teratogenicity studies have been conducted. Diquat in deionized water was administered by gavage to groups of eight female Wistar-derived rats at doses of 0, 4, 12, 24, or 40 mg of diquat ion per kg of body weight per day from day 7 to day 16 of gestation. On day 22, the females were killed and their uteri examined for live fetuses and intrauterine deaths. No fetal abnormality was observed. Fetal toxicity (reduced fetal weight gain) was seen at 40 mg of ion per kg of body weight per day. Maternal toxicity, as indicated by dose-related reduced food consumption and body-weight gain. was observed at all doses. The NOAEL for fetal toxicity was 24 mg of diquat ion per kg of body weight per day (Milburn. 1989). In another teratogenicity study, groups of 24 female rats were given diquat by gavage in deionized water at doses of 4. 12, or 40 mg of ion per kg of body weight per day from day 7 to day 16 of gestation. Maternal toxicity (reduced weight 214 PESTICIDES gain and food consumption), significant reductions in fetal weight, litter weight, and gravid uterine weight, and fetal defects in ossification were seen at 40 mg of ion per kg of body weight per day. The NOAEL for both maternal and fetal toxicity was 12 mg of ion per kg of body weight per day (Wickramatne, 1989). In a study in rabbits, diquat was given orally at doses of 1.3, 2.5, or 5.0 mg of ion per kg of body weight per day from day 1 to day 28 of gestation. There was no evidence of any effect on embryonic or fetal development. The NOAEL was 2.5 mg of diquat ion per kg of body weight per day based on mild reduction in maternal weight gain at the highest dose (Hodge, 1987). In a second study in rabbits, groups of7-8 female New Zealand white rabbits were administered diquat by gavage in deionized water at doses of 0, 1, 3, 7, or 10 mg of ion per kg of body weight per day from day 7 to day 19 of gestation. No evidence of fetotoxicity was observed. The NOAEL was 1 mg of ion per kg of body weight per day based on maternal weight loss and reduced weight gain and food intake at 3 mg of ion per kg of body weight per day (Hodge, 1989b ). In a third study, groups of 20 New Zealand white rabbits were administered diquat by gavage in deionized water at doses of 0,,1, 3, or 10 mg of ion per kg of body weight per day from day 7 to day 19 of gestation. The NOAEL was 1 mg of ion per kg of body weight per day based on maternal toxicity (reduced weight gain and food consumption) and skeletal effects in the fetuses (partially ossified sternebrae) at 3 and 10 mg of ion per kg of body weight per day (Hodge, 1989c ).
5.5 Mutagenicity and related end-points
Diquat has been adequately tested in a series of genotoxicity assays in vitro and in vivo. Chromosomal aberrations were induced in vitro, but there was no other evidence of genotoxicity. JMPR concluded that diquat was not genotoxic (FAO/WHO, 1994).
5.6 Toxicity ofmetabolites
In Wistar-derived rats, diquat dipyridone and monopyridine were less toxic than diquat when given subcutaneously (Parkinson, 1974; Crabtree, 1976).
The LD 50 ofTOPPS in male and female Wistar rats was about 3000 mg/kg of body weight. TOPPS gave negative results in several in vitro genotoxicity tests (FAO/WHO, 1994).
6. EFFECTS ON HUMANS
The acute lethal dose of diquat dibromide is considered to be 6-12 g for humans (WHO, 1984). Symptoms of diquat poisoning include oliguria, coma, shock and cardiac arrest, renal failure leading to anuria, and ventricular fibrillation (Vanholder et al., 1981). Cataracts have not been observed in those engaged in diquat manufacture or formulation (WHO, 1984; Bonsall, 1990). DIQUAT 215
7. PROVISIONAL GUIDELINE VALUE
In 1993, JMPR established an ADI of 0-0.002 mg of diquat ion per kg of body weight based on a NOAEL of 0.19 mg of diquat ion per kg of body weight per day (based on cataract formation at the next higher dose) identified in a 2-year study in rats and using an uncertainty factor of 100 (FAO/WHO, 1993). Issues relevant to the establishment of a guideline value for diquat in drinking-water were addressed by JMP (FAO/WHO, 1996). JMP concluded that the ADI established by JMPR was relevant for the establishment of a drinking-water guideline value, and that a more accurate determination of potential dietary exposure would be useful in setting a drinking-water guideline. Assuming a 60-kg person consumes 2 litres of drinking water per day and allocating 10% of the JMPR ADI (1.9 ~g/kg of body weight, if not rounded) to drinking-water, a health-based value of 6 ~g/litre (rounded figure) can be calculated for diquat ion. However, the limit of detection of diquat in water is 1 ~g/litre, and its practical quantification limit is about 10 ~g/litre. As a result, a provisional guideline value of 10 ~g/litre is established for diquat ion, based on the practical quantification limit.
8. REFERENCES
Ash by R ( 1987) Diquat dibromide monohydrate: evaluation of potential carcinogeniclly zn dietary administratwn to mtce for 80 weeks. Vols. I & 2. Unpublished report prepared by Life Sciences Research Ltd., Suffolk. Supplied to WHO by Zeneca Agrochemicals, Femhurst, Surrey (Final Report CTL/C/3761; LSR 87/ILYOOl/375). Black WJ et al. (1966) Residues in herbage and silage and feeding experiments following the use of diquat as a desiccant. Journal of the science offood and agriculture, 17:506-509. Bonsall JL (1990) Letter to F. Tripet, Stauffer Chemicals, Switzerland, dated 9 October 1990. Colley Jet al. (1985) Diquat dibromide evaluation ofpotential carcinogemcity and chronic toxicity by prolonged dietary administration to rats. Unpublished report prepared by Huntingdon Research Centre, Cambridgeshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report ICI 406/83763). Crabtree HC (1976) Comparison of the subcutaneous toxtctly of diquat, diquat monopyridone and diquat dtpyndone. Unpublished report prepared by Imperial Chemical Industries PLC, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Femhurst, Surrey (Report CTLIP/351 ). Earl M, Boseley AD (1988) The determination of restdues of diquat in water and other hquid samples- a spectrophotometric method JCI Agrochemicals Residue Analytical Method No. ARAM 7B Unpublished report dated December 1988. FAO/WHO (1993) Pestlctde residues in food- 1993. Report of the Joint Meetmg of the FAO Panel of Experts on Pestictde Residues in Food and the Environment and the WHO Expert Group on Pesticide Residues. Rome, Food and Agriculture Organization of the United Nations (FAO Plant Production and Protection Paper 122). FAO/WHO (1994) Pesticide residues in food- 1993. Joint FAO/WHO Meeting on Pesticide Residues. Evaluations- 1993. Part II- Toxicology. Geneva, World Health Organization (WHO/PCS/94.4). FAO/WHO (1995) Pesticide residues in food -1994. Evaluations- 1994. Part I- Residues, Vol I. Rome, Food and Agriculture Organization of the United Nations (FAO Plant Production and Protection Paper 131/1). FAO/WHO ( 1996) Pesticide residues in food- 1995. Report of the Joint Meeting of the FAO Panel of Experts on Pesticide Residues in Food and the Environment and the WHO Expert Group on Pesticide Residues (Annex IV). Rome, Food and Agriculture Organization of the United Nations (FAO Plant Production and Protection Paper 133). 216 PESTICIDES
Feldmann RJ, Maibach HI (1974) Percutaneous penetration of some pesticides and herbicides m man. Toxzcology and applzed pharmacolog~>, 28:126-132 Fletcher K, Gnffiths D, Kinch DA (1972) Diquat dzbromzde three-generatzon reproductwn study m rats Unpublished report prepared by Imperial Chemical Industries Ltd., Industrial Hygiene Research Laboratories. Supplied to WHO by Zeneca Agrochemicals, Femhurst, Surrey (Report HO/IH/R/331 A). Griggs RE, Davis JA (1975) Drquat excretwn and metaholzsm m a goat Unpublished report prepared by !Cl Plant Protection Ltd Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report AR 2585 A). Hemingway RJ, Leahey JP, Davis JA ( 1974) Dzquat. metabolzsm of dzquat and rts photoproducts m a cow Unpublished report prepared by ICI Plant Protection Ltd. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report AR 2448 B). Hodge MC (1987) Drquat drbromide teratogemc studres m the rabha. Unpublished report prepared by Imperial Chemical Industries PLC, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report OH/CTL!Pll4B). Hodge MC (1989a) Diquat 90 day feedrng study in the rat Unpublished report prepared by Impenal Chemical Industries PLC, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIP/1 832, revised). Hodge MC (1989b) Diquat· embryotoxrcrty study m the rahhzt Unpublished report prepared by ICI Central Toxicology Laboratory, Cheshire Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIP/2196). Hodge MC (1989c) Dzquat embryotoxzcrty study rn the rabbi/ Unpublished report prepared by ICI Central Toxicology Laboratory, Cheshire Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTL/P/2379). Hodge MC (\990) Dzquat multzgeneratwn study 111 the rat. Unpublished report prepared by ICI Central Toxicology Laboratory, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIP/2462). Hodge MC (1992) Dzquat. two year feedmg study m mrce. Unpublished report prepared by ICI Central Toxicology Laboratory, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTL/P/3409). Hopkins MN (I 990) Drquat I year feedmg study 111 dogs. Unpublished report prepared by ICI Central Toxicology Laboratory, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIP/2596) IPCS (I 996) The WHO recommended classification of pestzczdes by hazards and guidelrnes to classificatwn. Geneva, World Health Organization, International Programme on Chemical Safety (WHO/PCS/96.3). Johnston AM et al. (1990) The elimination of 14C-dzquat m the rat followmg smgle dose oral administration (low level). Unpublished report prepared by Inveresk Research International, Tranent. Supplied to WHO by Zeneca Agrochemicals, Femhurst, Surrey (Report CTLIC/2554). Milburn GM (1989) Dzquat embryotoxzcl/y study in the rat Unpublished report prepared by ICI Central Toxicology Laboratory, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIP/2275) Parkinson GR (1974) Dzquat monopyndone· acute and subacute oral toxzcity Unpublished report prepared by ICI Central Toxicology Laboratory, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIP/122 13) Rogerson A, Broad RD (1978) Diquat dibromzde 2 year feeding study m rats- amended report. Unpublished report prepared by Imperial Chemical Industnes PLC, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIP/253A). Scott RC, Walker M, Mawdesley SJ (1991) Diquat dzbromrde. rn vztro absorptwn from techmcal concentrate ("Reglone 40 ") and spray strength solutwn through human skm Unpublished report prepared by ICI Central Toxicology Laboratory, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey. Vanholder R et al. (1981) Diquat intoxication: report of two cases and review of the literature. American ;ournal ofmedicine, 70: 1267-I 271. DIQUAT 217
WHO (1984) Paraquat and d1quat. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria 39). Wickramatne GA (1989) D1quat· teratogenicity study in the rat Unpublished report prepared by Imperial Chemical Industnes PLC, Cheshire. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIP/2331 ). Williams SG, Cameron BD, McGuire GM (1991) Identification of the ma;or radioactive components in urme and faeces from rats following single oral administratiOn of [I4cj-drqual Unpublished report prepared by Inveresk Research International, Tranent. Supplied to WHO by Zeneca Agrochemicals, Fernhurst, Surrey (Report CTLIC/2523).
GLYPHOSATE
First draft prepared by H. Galal-Gorchev International Programme on Chemical Safety World Health Organization, Geneva, Switzerland
Glyphosate was not evaluated in the second edition of the WHO Guidelines for drinking-water quality. JMPR evaluated glyphosate in 1986 and established an ADI of 0-0.3 mg/kg of body weight. A 1994 IPCS Environmental Health Criteria monograph is available from which a higher ADI could be surmised. As this pesticide has been found in drinking-water in the Netherlands, the Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality considered it a high-priority chemical for evaluation.
1. GENERAL DESCRIPTION
1.1 Identity
CAS no.: 1071-83-6
Molecular formula: C3H 8N05P
The IUP AC name for glyphosate is N-(phosphonomethyl)glycine. Glyphosate is a weak organic acid; it consists of a glycine moiety and a phosphonomethyl moiety. It is usually formulated as the isopropylamine salt in aqueous solution at a concentration of about 360 g of glyphosate per litre. The CAS number of glyphosate-isopropylammonium is 38641-94-0, and its molecular formula is C6H 17N20 5P (WHO, 1994).
1.2 Physicochemical properties (WHO, 1994)
Property Value Vapour pressure <10-5 Pa at 25°C (negligible) Melting point 185°C (decomposes at 199°C) Log octanol-water partition coefficient -2.8 Water solubility 10.1 g/litre at 20°C Specific gravity 1.70 g/cm3
1.3 Major uses
Glyphosate is a broad-spectrum post-emergence herbicide. It has a high activity when applied to foliage, and it is used worldwide in both agriculture and forestry. Glyphosate is also used for aquatic weed control (WHO, 1994). 220 PESTICIDES
1.4 Environmental/ate
Glyphosate is strongly bound to soil particles and is not taken up by the roots of plants. It is metabolized very little by plants, the major metabolite being aminomethylphosphonic acid (AMP A). Glyphosate readily translocates from treated foliage to other parts of the plant. Residues from treated weeds passing into the soil are not taken up by other plants (FAO/WHO, 1986). Microbial biodegradation of glyphosate occurs in soil, aquatic sediment, and water. The main route of biodegradation of glyphosate appears to be by splitting the C-N bond to produce AMP A, the principal microbial metabolite; AMP A is also biologically degradable, with liberation of carbon dioxide. Degradation occurs more rapidly in aerobic than in anaerobic conditions. Half-lives for biodegradation in soil vary widely and range between a few days and several months; in water, half-lives between 12 hours and 7 weeks have been measured (CCME, 1989). Glyphosate is chemically stable in water and is not subject to photochemical degradation (FAO/WHO, 1986). The low mobility of glyphosate in soil indicates a minimal potential for the contamination of groundwater. Glyphosate can, however, enter surface and subsurface waters by direct use near aquatic environments or by runoff or leaching from terrestrial applications. This has been substantiated by reports that indicate the presence of glyphosate residues in water from direct overspray in forestry operations, from runoff, and from irrigation canal discharges. Furthermore, the possibility of aquatic contamination from drift during agricultural or silvicultural applications also exists. Depending upon the suspended solids loading and the microbial activity of flowing water, glyphosate may be transported several kilometres downstream from the site of aquatic application (CCME, 1989). Glyphosate is not expected to bioaccumulate in food in view of its high water solubility and its ionic character. Although residues of glyphosate were found in fish, crustaceans, and molluscs after exposure to water containing glyphosate, residues declined to about 50--90% of the accumulated levels when these aquatic organisms were subsequently exposed to water free from glyphosate for 14--28 days (FAO/WHO, 1986).
2. ANALYTICAL METHODS
Various analytical methods for the determination of glyphosate have been described, including thin-layer chromatography, high-performance liquid chromatography, and gas chromatography-mass spectrometry. The limits of determination were 0.02-50 J.tg/litre in water, 0.05-1 mg/kg in soil, 0.01-0.05 mg/kg in plants, and about 0. 3 J.tg/m3 in air. The limit of determination of AMP A in water is reported to be 1.2 J.tg/litre (WHO, 1994).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Concentrations in air are available only from studies on exposures of workers involved in application of the herbicide. Air concentrations during silvicultural GLYPHOSATE 221
1 1 spraying were mostly below 1.3 1-1g/m ; the highest value observed was 15.7 1-1g/m . The highest estimated exposure (dermal and inhalation) of about 8000 1-lg/hour, as reported in a study with spray applicators, corrected for incomplete absorption. equals about 40 1-1g/kg of body weight per day (8-hour working day for a 60-kg adult) (WHO, 1994).
3.2 Water
In a survey conducted in 1988-1989 in the Netherlands, surface water contained 0.5-1 1-1g of glyphosate per litre and 6 1-1g of the metabolite AMP A per litre (WHO, 1994). In Canada, glyphosate residues as high as 5153 1-1g/litre were measured after direct aerial application over lakes, ponds, or streams. Glyphosate concentrations in water declined to a few 1-lg/litre or to non-detectable levels hours or days post-treatment, depending on the extent of vegetation present. The concentration of AMP A in water without substantial vegetation was about 3 1-1g/litre (CCME, 1989). In the USA, pond water contained 90-1700 1-1g of glyphosate per litre and 2-35 1-1g of AMPA per litre, whereas stream water contained 35-1237 and <1.0-10 1-1g ofglyphosate and AMPA, respectively, per litre (WHO. 1994).
3.3 Food
No information was available on direct measurements of glyphosate in foodstuffs (as part of food surveillance) or total diets. The only information available comes from residue levels resulting from supervised trials. In pre-planting use of glyphosate, residues of glyphosate and its metabolite were not detected (<0.05 mg/kg) in cereal grains at harvest. Pre-harvest application of glyphosate to cereals and pulses resulted in mean residue levels ranging from 0.2 to 4.8 mg/kg, when the glyphosate was used according to good agricultural practice. Industrial processing of wheat to flour resulted in a decrease in glyphosate level from 1.6 to 0.16 mg/kg (FAO/WHO, 1986). Fish exposed to water containing 10 mg of glyphosate per litre for 14 days contained 0.2-0.7 mg of glyphosate per kg. Residues were reduced when fish were exposed to glyphosate-free water. In controlled feeding studies, mean residues of glyphosate found in muscle tissues of pigs, poultry, and cattle were <0.05 mg/kg. Livers of these animals contained up to 0.12 mg/kg, whereas residues in cattle milk were not detectable (FAO/WHO, 1986).
3.4 Estimated total exposure and relative contribution of drinking-water
Use of glyphosate as a herbicide may result in the presence of residues in air, drinking-water, crops, and animal tissues destined for human consumption. Main routes of exposure to glyphosate are expected to be inhalation and dermal exposure in the occupational setting and consumption of water and food for the general population. Because of its sorption to particulate matter and its microbial degradation in the aquatic environment (CCME, 1989), the major source of exposure to glyphosate is expected to be food. 222 PESTICIDES
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
The results of oral studies with 14C-glyphosate in rats, rabbits, and goats indicate that absorption from the gastrointestinal tract is incomplete and amounts to approximately 30% of the dose or less. On day 7 after administration of a single oral dose of 14C-glyphosate to rats, the isotope was widely distributed throughout the body, with the highest concentration found in the bones. Biotransformation of glyphosate occurs to a very low degree only. In rats, it was shown that almost all of the 14C in urine and faeces, after a single oral administration of 14 C-glyphosate, was present . as unchanged parent compound. Elimination through exhaled air is very low. AMP A was the only metabolite, accounting for only 0.2-0.3% of the applied dose of 14C-glyphosate (WHO, 1994). In a study of the metabolic fate of AMP A in rats, AMP A was only moderately absorbed (approximately 20%); excretion was almost exclusively via the urine, with less than 0.1% of the dose expired as carbon dioxide (FAO/WHO, 1987).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
Glyphosate and its formulations have very low acute toxicity by the oral and dermal administration routes. Median oral lethal doses (LD50s) of glyphosate range from 1950 to >5000 mg/kg of body weight for mice, rats, and goats (WHO, 1994). Glyphosate has been classified by WHO as unlikely to present an acute hazard in normal use (IPCS, 1996).
5.2 Short-term exposure
In a 13-week feeding study, groups of 15 male and 15 female Charles River CD-1 mice were fed technical glyphosate (purity 98.7%) in their diet at dose levels of 0, 0.5, 1.0, or 5.0%. No effect on appearance or survival was observed. Growth retardation and increased weights of brain, heart, and kidneys were observed at 5.0%. Liver weights were increased at 1.0% and 5.0%. Limited histopathology showed no adverse effects. The authors of the study concluded that the NOAEL was 1.0% glyphosate in the diet, equal to 1890 mg/kg of body weight per day (Bio/Dynamics Inc., 1979; FAO/WHO, 1987; WHO, 1994). In a 13-week feeding study, Sprague-Dawley rats received 0.1, 0.5, or 2% technical glyphosate in their diet. No effects on appearance, survival, or growth were observed. Haematology, blood biochemistry, and urinalysis, carried out at test end only, were also unaffected. Organ weights determined for liver, kidneys, and testes were not affected. Limited histopathology showed no adverse effect in any tissue. The NOAEL in this study was 2% g1yphosate in the diet (the highest dose tested), equal to 1267 mg/kg of body weight per day (Monsanto, 1987). GLYPHOSATE 223
Two further 13-week studies in rodents were conducted. Both mice (B6C3F 1) and rats (F-344/N) were administered glyphosate (purity approximately 99%) in feed at levels of 0, 3125, 6250, 12 500, 25 000, or 50 000 mglkg of feed (NTP, 1992). In mice, reduced weight gains were observed at 50 000 mg!kg of diet in both sexes. Dose dependent lesions in the parotid gland were observed at 6250 mg/kg of feed and higher but were not seen at the lowest dose level tested. The NOAEL in this study was 3125 mglkg of feed, equal to 507 mglkg of body weight per day (NTP, 1992). In rats, reduced weight gains were observed in males at 25 000 mg/kg of feed and in both sexes at 50 000 mg/kg of feed. Clinical chemistry showed increased alkaline phosphatase and alanine aminotransferase at 6250 mg/kg of feed in males and at 12 500 mg/kg of feed in females. Decreases in sperm count were observed in males at 25 000 and 50 000 mglkg of feed. Cytoplasmic alterations of the parotid and submandibular salivary glands, consisting of basophilic changes and hypertrophy of acinar cells, were observed. Effects on the salivary glands were observed at the lowest dose tested (3125 mglkg of feed, equal to 205 mg/kg of body weight per day for males and 213 mg/kg of body weight per day for females). Thus, a NOAEL could not be identified in this study (NTP, 1992). Groups of six male and six female beagle dogs were administered technical glyphosate (96.1% pure) in gelatin capsules at dose levels of 0, 20, 100, or 500 mg/kg of body weight per day for 52 weeks. No effects were observed with respect to clinical signs, body weight, feed consumption, ophthalmoscopy, haematology, urinalysis, gross pathology, and histopathology. The NOAEL in this study was 500 mg/kg of body weight per day, the highest dose tested (FAO/WHO, 1987; WHO, 1994).
5.3 Long-term exposure and carcinogenicity
In a combined chronic toxicity and carcinogenicity study, groups of Charles River CD-1 mice (50 per sex per group) were fed technical glyphosate in the diet for 24 months at levels of 0, 0.1, 0.5, or 3.0%. No effect on survival or appearance was noted. Body weights were decreased in the males of the high-dose group. Haematology and organ weights showed no effects. Histopathology in liver revealed an increased incidence of central lobular hepatocyte hypertrophy and hepatocyte necrosis among high-dose males. Hyperplasia of the urinary bladder was increased in frequency in mid- and high-dose males (incidences: 3/49, 3/50, 10/50, and 8/50), but not in treated females. There were no statistically significant increases in the frequency of neoplastic lesions. The NOAEL in this study was 0.5% glyphosate, equal to 814 mg/kg of body weight per day (Bio/Dynamics Inc., 1983). Groups of Charles River Sprague-Dawley rats (50 per sex per dose) were fed technical glyphosate in their diets at dose levels of about 0, 3, 10, or 32 mg/kg of body weight per day for 26 months. Survival, appearance, haematology, blood biochemistry, urinalysis, and organ weights were not changed. Slight growth retardation during part of the study was noted in the high-dose males. The incidence of interstitial cell tumours in testes showed a statistically significant increase (incidences: 0/50, 3/50, 1/50, and 6/50; historical control range: 3-7%) (Bio/Dynamics Inc., 1981a). This finding, in itself constituting evidence of a 224 PESTICIDES carcinogenic effect in rats, should be judged in light of the absence of an effect at much higher dose levels in the more recent 2-year study in rats (see below). This is also valid for the slight growth retardation. The NOAEL was 32 mg/kg of body weight per day, the highest dose tested (Bio/Dynamics Inc., 198la). In the recent 2-year study, groups of Charles River Sprague-Dawley rats (60 per sex per dose) were fed technical glyphosate in their diets at dose levels of about 0, 100, 410, or 1060 mg/kg of body weight per day for 24 months. There was no effect on survival or appearance. Growth was retarded in the high-dose females. Haematology and blood biochemistry showed no effects. In the high-dose males, the urine specific gravity and urine pH were increased. A statistically significant increased incidence of degenerative lens changes was found among the high-dose males; however, this finding was within the historical control range. Liver weights were increased in the high-dose males only. Increased incidence of inflammation ofthe gastric squamous mucosa was observed in the mid- and high-dose groups (incidences in males: 2/58, 3/58, 5/59, and 7/59; females: 0/59, 3/60, 9/60, and 6/59; historical range: 0-13.3%). The incidence of pancreatic islet cell adenomas was increased (statistically significant) among low- and high-dose animals. However, these effects were within the historical control range. No pancreatic carcinomas were found. The NOAEL in this study was 410 mg/kg of body weight per day (Monsanto, 1990a).
5.4 Reproductive and developmental toxicity
Groups of female Charles River CD-I rats were administered technical glyphosate by gavage at dose levels of 0, 300, 1000, or 3500 mg/kg of body weight per day on days 6-19 of gestation. At 3500 mglkg of body weight per day, the following effects were observed: increased incidence of soft stools, diarrhoea, breathing rattles, red nasal discharge, reduced activity, increased mortality (6/25 dams dying before the end of the treatment period), growth retardation, increased incidence of early resorptions, decreases in total number of implantations and the number of viable fetuses, and increased number of fetuses with reduced ossification of sternebrae. At the lower dose levels, these effects were absent. The NOAEL in this study was 1000 mg/kg of body weight per day (IRDC, 1980a). Groups of 16 female Dutch belted rabbits received technical glyphosate by gavage in 0.5% Methocel at dose levels of 0, 75, 175, or 350 mg/kg of body weight per day on days 6-27 of gestation. The control group received the vehicle only. The incidence of diarrhoea and soft stools was increased in the high-dose group and also, to a slight degree, in the mid-dose group. The incidence of nasal discharge was increased in the high-dose group only. In the mid- and high-dose groups, 2 and 10 dams, respectively, died during the study from unknown causes. The IPCS Task Group concluded that the NOAEL was 175 mg/kg of body weight per day (IRDC, 1980a; WHO, 1994). In a three-generation study, groups of Sprague-Dawley rats were given glyphosate (98.7% pure) in the diet at doses ofO, 3, 10, or 30 mg/kg of body weight per day for 60 days. The only effect noted was an increased incidence of unilateral renal tubular dilation in the F3b male pups of the high-dose group (incidence not determined in mid-dose group; earlier litters not examined). The NOAEL in this GLYPHOSATE 225 study was 30 mg/kg of body weight per day, the highest dose tested (Bio/Dynamics Inc., 1981b; WHO, 1994). In a more recent two-generation feeding study, Sprague-Dawley rats received glyphosate at doses of 0, 100, 500, or 1500 mg/kg of body weight per day. Soft stools and decreased body weights in parent animals and slightly decreased litter size and pup weights were seen in the high-dose group. Decreased body weights of parents and pups were seen to a slight degree in the mid-dose group. No histological effect on kidneys was present in the F2b male pups (15 and 23 pups examined in control and high-dose groups, respectively; first generation and F2• pups not examined). The NOAEL in this study was 500 mg/kg of body weight per day (Monsanto, 1990b; WHO, 1994). In its evaluation of these latter two reproductive toxicity studies, the IPCS Task Group noted that the number of pups submitted to histopathological examination in both studies was limited. These limitations made it difficult to evaluate the renal effect seen in pups at 30 mg/kg of body weight per day in the Bio/Dynamics Inc. (1981b) study (WHO, 1994).
5.5 Mutagenicity and related end-points
Glyphosate was consistently without mutagenic effect in a range of genotoxicity assays in vitro and in vivo (WHO, 1994).
5.6 Toxicity ofmetabolites
No toxicological data were reported on AMP A in the IPCS monograph on glyphosate (WHO, 1994).
6. EFFECTS ON HUMANS
Several cases of (mostly intentional) intoxications with technical glyphosate herbicide formulation have been reported. A typical symptom is erosion of the gastrointestinal tract. No compound-related effects were observed in a test group of five applicators prior to and after exposure for l week. No controlled studies have been conducted in humans.
7. CONCLUSIONS
JMPR established an ADI for glyphosate of 0-0.3 mg/kg of body weight, based upon a NOAEL of 32 mg/kg of body weight per day, the highest dose tested, identified in a 2-year study in rats and using an uncertainty factor of 100 (Bio/Dynamics Inc., 1981 a; FAO/WHO, 1987). This study has been superseded by the Monsanto (1990a) study, in which higher levels of glyphosate were administered, resulting in a higher NOAEL (410 mg/kg of body weight per day) for long-term toxicity/carcinogenicity. The IPCS Task Group on the Environmental Health Criteria monograph for glyphosate identified a NOAEL of 175 mg/kg of body weight per day in the 226 PESTICIDES teratogenicity study in rabbits (IRDC, 1980b) and considered that an uncertainty factor of 100 would be appropriate for the derivation of an ADI for glyphosate, given the elaborate data sets available. This gives an ADI of0-1.75 mg/kg of body weight (WHO, 1994). Using this ADI and assuming a 60-kg person consuming 2 litres of drinking-water per day, a health-based value of 5 mg/litre (rounded figure) is obtained for an allocation of 10% of the ADI to drinking-water. Because of its low toxicity, the health-based value derived for glyphosate is orders of magnitude higher than glyphosate concentrations normally found in drinking-water. Under usual conditions, therefore, the presence of glyphosate in drinking-water does not represent a hazard to human health. For this reason, the establishment of a numerical guideline value for glyphosate is not deemed necessary. AMP A is the major metabolite of glyphosate. It was noted that most AMP A found in water comes from sources other than glyphosate degradation, and that AMP A is scheduled to be evaluated by JMPR.
8. REFERENCES
Bio/Dynamics Inc. (I 979) A three month feeding study of glyphosate (Roundup techmca/) in mice (Pro;ect No. 77-211) -final report. Unpublished report prepared by Bw/Dynamics Inc., Division of Biology and Safety Evaluation, East Millstone, NJ. Submitted to WHO by Monsanto Ltd. Bio/Dynamics Inc. ( 1981 a) A lifetime feeding study of g(vphosate (Roundup techmca/) in rats Unpublished report prepared by Bio/Dynamics Inc., Division of Biology and Safety Evaluation, East Millstone, NJ. Submitted to WHO by Monsanto Ltd. (Project No. 410n7; BDN-77-416). Bio/Dynamics Inc. (198lb) A three-generatwn reproduction study in rats with glyphosate Final report. Unpublished report prepared by Bio/Dynamics Inc., Division of Biology and Safety Evaluation, East Millstone, NJ. Submitted to WHO by Monsanto Ltd. (Project No. 77-2063, BDN-77-147). Bio/Dynamics Inc. (1983) A chronic feeding study of glyphosate (Roundup technical) m m1ce. Unpublished report prepared by Bio/Dynamics Inc., Division of Biology and Safety Evaluation, East Millstone, NJ. Submitted to WHO by Monsanto Ltd. (Project No. 77-2061; BDN-77-420). CCME (I 989) Canadwn water quality guidelines. Ottawa, Ontario, Environment Canada, Canadian Council of Ministers of the Environment. FAO/WHO (1986) Pesticide residues in food- 1986. Evaluations- 1986. Part I- Res1dues Rome, Food and Agriculture Organization of the United Nations (FAO Plant Production and Protection Paper 78). FAO/WHO (1987) PestiCide res1dues m food- 1986 Joint FAO/WHO Meeting on Pestic1de Residues. Evaluatwns- 1986. Part li- Toxicology. Geneva, World Health OrganizatiOn (FAO Plant Production and Protection Paper 78/2). IPCS (1996) The WHO recommended classificatiOn of pesticides by ha=ards and gwdelmes to classification Geneva, World Health Organization, International Programme on Chemical Safety (WHO/PCS/96.3). IRDC (1980a) Test article - Techmcal glyphosate: Teratology study m rats. Unpublished report prepared by International Research and Development Corporation, Mattawan, Ml. Submitted to WHO by Monsanto Ltd. (Study No. 401-054, Reference No. IR-79-016). IRDC ( 1980b) Test article - Technical glyphosate: Teratology study 111 rabbits Unpublished report prepared by International Research and Development Corporation, Mattawan, Ml. Submitted to WHO by Monsanto Ltd. (Study No. 401-056, Reference No. IR-79-018). GLYPHOSATE 227
Monsanto (1987) 90 day study of glyphosate administered in feed to Sprague-Dawley rats Unpublished report No. MSL 7575, prepared and submitted to WHO by Monsanto Ltd., Monsanto Environmental Health Laboratory, St. Louis, MO (Project No. ML-86-3511EHL 86128). Monsanto (1990a) Chronic study of glyphosate administered in feed to albino rats Unpublished report prepared and submitted to WHO by Monsanto Ltd., Monsanto Environmental Health Laboratory, St. Louis, MO (Project No. MSL-10495). Monsanto (1990b) Two generation reproduction feeding study with glyphosate in Sprague-Dawley rats Unpublished report prepared and submitted to WHO by Monsanto Ltd., Monsanto Environmental Health Laboratory, St. Louis, MO (Project No. MSL-10387). NTP (1992) NTP technical report on toxicity studzes ofglyphosate (CAS No. 1071-83-6). Research Triangle Park, NC, National Toxicology Program (Toxicity Report Series No. 16). WHO (1994) Glyphosate. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria 159).
PENTACHLOROPHENOL
First draft prepared by J. Tuomisto Division of Environmental Health, National Public Health Institute Kuopio, Finland
A provisional guideline value for pentachlorophenol (PCP) of 9 Jlg/litre was established in the 1993 WHO Guidelines for drinking-water quality. The guideline value was based on a NOAEL of 3 mg/kg of body weight per day, observed in rat reproduction studies, an uncertainty factor of 1000 (100 for inter- and intraspecies variation and 10 for potential carcinogenicity of technical PCP), and an allocation of 10% of the resulting TDI to drinking-water. The guideline value was designated as provisional. as it was evaluated only at the Final Task Group Meeting of September 1992, on the basis of a 1987 IPCS Environmental Health Criteria monograph. The Coordinating Committee for the Updating of WHO Guidelines for drinking-water quality considered it desirable to evaluate PCP in the 1998 Addendum.
1. GENERAL DESCRIPTION
1.1 Identity
CAS no.: Molecular formula:
Unpurified technical pentachlorophenol (PCP) contains other chlorophenols and chlorophenoxyphenols, as well as several microcontaminants, particularly polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs). Because of attempts to reduce toxic impurities, various technical preparations may be quite different both temporally and geographically. Chlorophenol preparations in various countries have varied from almost pure PCP to those in which 2,3,4,6-tetrachlorophenol was the main component, 2,4,6-trichlorophenol concentrations were significant, and PCP accounted for as little as 5-10% (IARC, 1986).
1.2 Physicochemical properties (WHO, 1987)
Property Value Physical state light tan to white needle-like crystals Water solubility (at 20°C) 14 mg/litre (pH 5) 2 g/litre (pH 7) 8 g/litre (pH 8) Log octanol-water partition coefficient 3.56 (pH 6.5) 3.32 (pH 7.2) Vapour pressure 2 x 10-6 kPa at 20°C 230 PESTICIDES
Property Value Melting point 19!°C Boiling point 31 0°C (decomposition) Density 1.987 g/cm3 pK. 4.7 at 25°C
1.3 Organoleptic properties
The odour threshold of PCP in water has been given as 0.86 mg/litre at 30°C (ATSDR, 1994) and 1.6 mg/litre (temperature not specified) (WHO, 1987). The taste threshold is 0.03 mg/litre (temperature not specified) (WHO, 1987). In cold water, the taste threshold is increased (Lamp et al., 1990).
1.4 Major uses
PCP and other chlorophenols are used primarily for protecting wood from fungal growth. Their use is in decline, and they have been abandoned from most other applications, such as indoor disinfectant, leather and textile application, and herbicide uses. In several countries, their use has been totally discontinued (e.g. Sweden, Germany, Finland) or practically abandoned as a result of severe restrictions (e.g. Denmark). However, PCP is still an important pesticide in some developing countries because of its low cost and broad spectrum. In some developed countries (e.g. France, USA), several thousand tonnes are produced annually (IARC, 1991; McConnell et al., 1991). Even in those countries where PCP use has been abandoned, PCP continues to be an important environmental contaminant, because it is imported via various materials treated with it.
1.5 Environmental/ate
PCP and other chlorophenols can be metabolized by numerous aquatic and soil microorganisms, but environmental conditions are usually unfavourable for biodegradation (Salkinoja-Salonen et al., 1989; Bajpai & Banerji, 1992; Orser & Lange, 1994; Laine & Jorgensen, 1996). Slow elimination in surface waters, high persistence in sediments, formation of stable metabolites, and the limited adaptation of microorganisms to chlorophenols owing to their high microbial toxicity imply that chlorophenols are practically non-biodegradable in the aquatic environment (UBA, 1996). In addition, the trace contaminants, especially PCDDs and PCDFs, are not metabolized. Chlorophenols are relatively water soluble in the anionic form. At pH 6.7, 99% of PCP is ionized and in easily leachable form. Therefore, soil contamination may lead to the contamination of groundwater as well. The solubility of the dioxin!furan contaminants is very low (in the order of 1o-s g/litre ), and these contaminants are readily sorbed onto soil particles and other surfaces; hence, the risk of contamination of finished drinking-water is not great. In surface waters, organic or clay particles easily transport dioxins/furans to distant sites from their origin (Koistinen et al., 1995). PENTACHLOROPHENOL 231
PCP is accumulated by aquatic organisms through uptake from the surrounding water or along the food-chain (WHO, 1987).
2. ANALYTICAL METHODS
A capillary gas chromatographic method with electron capture detection is the most widely applied method of assaying PCP (usually methylated or acetylated) after acid extraction to diethyl ether or another organic solvent (WHO, 1987; IARC, 1991; Jorens & Schepens, 1993). The detection limit is 0.005-0.01 J.lg/litre. Recently, other methods have been advocated, such as gas chromatography with atomic emission detection (Tumes et al., 1994), gas chromatography-mass spectrometry using selected ion monitoring (Kontsas et al., 1995), and high performance liquid chromatography (Frebortova & Tatarkovicova, 1994).
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
Because of its high vapour pressure, PCP easily evaporates from treated wood surfaces, and the loss may be as high as 30-80% a year (WHO, 1987). Its indoor use in buildings is highly inadvisable (Jorens & Schepens, 1993; Liebl et al., 1995) and has caused a number of poisonings. Volatilization from water is pH dependent, and only the un-ionized form seems to be volatile (WHO, 1987). Although PCP is ubiquitous, there is little information on its concentrations in ambient air (WHO, 1987; Thompson & Treble, 1995). Fallout as calculated from Finnish snow samples ranged from 1.49 to 136.0 f.lg/m 2 (Paasivirta et al., 1985).
3.2 Water
In the past, PCP concentrations as high as 25 000-150 000 f.lg/litre in industrial effluent were reported; at one time, 30-40 t were calculated to be transported by the Rhine per year (WHO, 1987). Concentrations up to 10 500 f.lg!litre have been reported locally in a river (Fountaine et al., 1976), but concentrations in water samples are usually below 10 J.lg/litre (WHO, 1987). Monitoring data showed that PCP concentrations generally decreased (from 0.07-0.14 f.lg/litre in 1988 to 0.01-0.02 f.lg/litre in 1993) in the River Elbe after PCP production was stopped in Germany in 1986 and its use was banned in 1989. Such a trend was not seen in the Rhine and its tributaries, where concentrations were even higher in 1990-1991 than in 1980-1989 (maximum levels up to 0.23 f.lg/litre ); the cause is not known (UBA, 1996), but it indicates continuing environmental contamination. Under special conditions, PCP may accumulate to very high concentrations in ground water. In the spring water of a Finnish village, concentrations of 70-140 f.lg/litre were found, and investigation of the groundwater reservoir revealed extremely high concentrations of 56 000-190 000 f.lg/litre in the deep parts of the reservoir (Lampi et al., I 990). The obvious source was a local sawmill using chlorophenols since 1940s. 232 PESTICIDES
Bottom sediments usually contain higher concentrations of PCP than the overlying waters. In Canadian water samples, PCP concentrations ranged from non detectable to 7.3 J..Lg/litre; in the respective sediments, concentrations ranged from non-detectable to 590 J..Lg/kg (WHO, 1987). In a lake downstream of the above mentioned Finnish village, total chlorophenol concentrations of 2.6-11.0 J..Lg/litre were detected (Lampi et al., 1990), whereas the concentration in sediment was 100 J..Lg/kg dry weight or more (Lampi et al., 1992a). In New Zealand, the highest PCP concentrations (3.6 J..Lg/litre in surface water and 400 J..Lg!kg dry weight in sediments) were found downstream from a sawmill (Gifford et al., 1996). PCP concentrations in drinking-water are usually in the range ofO.Ol-0.1 J..Lg/litre (WHO, 1987). It has been thought that PCP's odour threshold is low and should render water unpalatable, but a recent episode in Finland demonstrated that people may drink water containing up to I 00 J..Lg/litre of chlorophenols with few complaints (Lampi et al., 1990). As stated above, evaporation depends on pH and temperature, and detection in cold water is unreliable.
3.3 Food
In Canada, analysis of 881 pork liver tissue samples revealed a gradual decline in PCP levels in 1988-1989 from those in earlier years. Some 6.6% of the samples contained levels in excess ofO.l mg/kg, the highest level being 0.72 mg/kg. Of 51 beef liver samples, 2.0% had levels in excess of 0.1 mg/kg, the maximum level being 0.35 mg/kg. Examination of 214 chicken and 68 turkey liver samples showed only one with a level above 0.1 mg/kg; this incident was traced to the use of wood shavings as bedding (Agriculture Canada, 1989). Another study in Slovakia found detectable PCP concentrations in 79% of food samples from school kitchens; the average PCP concentration in positive samples was 6.3 J..Lg/kg (Veningerova et al., 1994).
3.4 Estimated total exposure and relative contribution of drinking-water
Daily net intake of PCP in the general population has been estimated to be 5-35 J..Lg/day (Reigner et al., 1992a). The long-term average daily intake ofPCP by the general population in the USA was estimated to be 16 J..Lg/day using six-compartment environmental partitioning models; food, especially fruits, vegetables, and grains, accounted for 99.9% of the total exposure (Hattemer-Frey & Travis, 1989). In Canada, the estimated daily intake is 0.05 J..Lg!kg of body weight per day, mostly via food or indoor air (Coad & Newhook, 1992). It seems likely that food accounts for the majority of intake unless there is specific local chlorophenol contamination causing increased concentrations in drinking-water or exposure from log homes treated with PCP.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
The kinetics of PCP have been reviewed on a number of occasions (Ahlborg & Thunberg, 1980; WHO, 1987; IARC, 1991), and only a general overview and a PENTACHLOROPHENOL 233 few recently emphasized points are presented here. PCP is probably well absorbed orally (Reigner et al., 1992a), although bioavailability may be influenced by food (Yuan et al., 1994). During occupational exposure, PCP is well absorbed via the dermal and pulmonary routes (IARC, 1991). Good oral (WHO, 1987; Reigner et al., 1992b; Yuan et al., 1994) and dermal (Wester et al., 1993) absorption is supported by animal studies. It is likely that most PCP is conjugated to glucuronic acid and excreted into urine, although the rate of glucuronidation has been controversial (Reigner et al., 1992a). It is somewhat debatable whether an important rodent metabolite, tetrachloro-1 ,4-hydroquinone, is formed in humans, but at most it is a minor metabolite (Reigner et al., 1992a). PCP is avidly bound (99.5%) to plasma proteins in humans (Reigner et al., 1993). There is some uncertainty as to the elimination rate ofPCP, and half-lives from 33 hours to 16 days have been reported (Reigner et al., 1992a). Both protein binding and urinary pH could cause variation, but there may also be methodological reasons for the differences found. One may safely assume that the half-life is a few days. The bioconcentration factor for PCP is rather low, and steady-state levels in human adipose tissue amount to only 2-4 times the average daily intake (Geyer et al., 1986).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
Oral LD50 values range from 36 to 177 mg/kg of body weight in mice and from 27 to 175 mg/kg of body weight in rats (IARC, 1991).
5.2 Short-term exposure
A number of toxic effects in short-term tests have been attributed to impurities present in technical-grade PCP preparations. Rats receiving technical grade PCP at 500 mg/kg of feed (equivalent to 25 mg/kg of body weight per day) for 8 months had slow growth rates, liver enlargement, porphyria, and increased activities of some liver microsomal enzymes (Goldstein et al., 1977; Kimbrough & Linder, 1978). These were not seen when purified PCP was used, and in fact they are typical effects of tetrachlorodibenzo-p-dioxin (TCDD)-like compounds (Pohjanvirta & Tuomisto, 1994). Some effects on enzyme activity, such as strong inhibition of sulfotransferase activity, are due to PCP itself(Boberg et al., 1983). A number of effects on immune systems have been noted: reduced humoral immunity and impairment of T-cell cytolytic activity in mice (Kerkvliet et al., 1982a,b) and decreased cell-mediated and humoral immunity in rats (Exon & Koller, 1983). The effects are attributed to dioxin/furan impurities (Kerkvliet et al., 1985). Thyroid hormones were decreased in female Wistar rats after 4-week dosing by gavage with pure PCP at 30 mg/kg of body weight per day without a consequent increase in thyroid-stimulating hormone (Jekat et al., 1994), suggesting a hypophyseal or hypothalamic action. It is of some interest that PCP decreased the uptake of thyroxin into cerebrospinal fluid and the brain (van Raaij et al., 1994). 234 PESTICIDES
Male Wistar rats (number per group unspecified) administered PCP (purity unspecified) at 1.0 or 3.0 mmol!litre in drinking-water ad libitum for 3-4 months showed degenerative changes in peripheral nerves, including degeneration of myelin sheath and loss of neurotubules and neurofilaments and other axoplasmic components in A- and B- but not C-type fibres. Hepatocyte swelling and vacuolar degeneration were also seen, as well as some proximal tubular damage in the kidneys (Villena et al., 1992).
5.3 Reproductive and developmental toxicity
PCP is not teratogenic in rats, but it is embryo/fetotoxic at high doses (Schwetz et al., 1978; Exon & Koller, 1982; Welsh et al., 1987). Rats were administered 3 or 30 mg/kg of body weight of purified PCP 62 days before mating, during mating, and to females during gestation and lactation. The high dose caused reductions in the number of offspring, neonatal body weight, neonatal survival, and growth of weanlings. The NOAEL was 3 mg/kg of body weight per day (Schwetz et al., 1978). Developmental effects of purified (98% purity) and technical-grade (88.4% purity) PCP in rats were investigated by administering PCP orally at doses ranging from 5 to 50 mg/kg of body weight per day from day 6 to day 15 of gestation. Purified PCP caused delayed fetal development without simultaneous maternal toxicity at 5 mg/kg of body weight per day. Technical-grade PCP did not have any effects on the mother or fetus at 5 mg/kg of body weight per day (Schwetz et al., 1974). Female Sprague-Dawley rats were exposed to PCP (95% pure) at 0, 5, 50, or 500 mg/kg of feed (equivalent to 0, 0.25, 2.5, or 25 mg/kg of body weight per day). This study was designed to produce progeny that were exposed to PCP both pre- and postnatally. Decreased litter sizes and increased number of stillborn were observed at the highest dose (Exon & Koller, 1982). Male and female Sprague-Dawley rats were fed purified PCP (>99% pure) at doses of 0, 4, 13, or 43 mg/kg of body weight per day for 181 days, through mating and pregnancy. PCP was embryolethal and maternally toxic at the highest dose tested (Welsh et al., 1987).
5.4 Mutagenicity and related end-points
The genetic toxicology of PCP was thoroughly reviewed by Seiler (1991 ). PCP does not seem to produce DNA damage, and the few scattered observations of mutagenic activity were attributed to oxygen radical formation by its metabolite. This seems to be supported by later studies that implicate the rodent metabolite of PCP, tetrachloro-1 ,4-hydroquinone (Jansson & Jansson, 1992a; Dahlhaus et al., 1994, 1995; Naito et al., 1994; Waidyanatha et al., 1994; Sai-Kato et al., 1995; W ang & Lin, 1995). There. is some evidence of weak clastogenic effects in chromosomal aberration assays in vitro and in lymphocytes of exposed persons in vivo (Bauchinger et al., 1982; Seiler, 1991 ). Also, 2,4,6-trichlorophenol caused chromosome malsegregation rather than mutations in Chinese hamster ovary cells (Jansson & Jansson, 1992b; Armstrong et al., 1993). PENTACHLOROPHENOL 235
5.5 Carcinogenicity
Early carcinogenicity studies on PCP were negative or equivocal (Innes et al., 1969; Schwetz et al., 1978), but there were deficiencies in these studies (IARC, 1991). In a carefully executed study of the US NIP (McConnell et al., 1991), both a generic technical-grade PCP and a preparation with lower levels of impurities (about
2 orders lower with respect to PCDDs and PCDFs) were fed to B6C3F1 mice (50 per sex per group) for 2 years at average doses of 0, 18, or 3 5 mg/kg of body weight per day (technical grade) and 0, 18, 35, or 116 mg/kg of body weight per day (purer grade). Dose-dependent and significantly elevated levels of hepatocellular adenomas and carcinomas, phaeochromocytomas, and haemangiosarcomas were observed. Overall, the purer preparation was not less tumorigenic than the technical-grade PCP. Hepatic tumours and phaeochromocytomas were more common in males, whereas haemangiosarcomas were seen only in females. McConnell et al. (1991) came to a plausible conclusion that tumours are caused primarily by PCP itself and not by impurities, although hexachlorodibenzo-p-dioxin may play a small part in accentuating the effect. IARC (1991) concluded that there is sufficient evidence for the carcinogenicity of PCP in experimental animals. The US NIP is conducting a 2-year feeding study in F344 rats (as cited by Yuan et al., 1994). There is a paucity of carcinogenicity data on 2,3,4,6-tetrachlorophenol. 2,4,6-Trichlorophenol was found to cause hepatocellular carcinomas or adenomas in mice of both sexes and lymphomas and leukaemias in male rats (National Cancer Institute, 1979). This was considered to provide sufficient evidence for its carcinogenicity in experimental animals (IARC, 1982).
6. EFFECTS ON HUMANS
The paramount difficulty in interpreting human studies, especially long-term studies, has been simultaneous exposure to other chemicals (Johnson, 1990; Saracci et al., 1991 ). Those working in chemical industries are often simultaneously exposed to chlorophenols, solvents, dibenzodioxins, dibenzofurans, and the chemical being synthesized. Those working in forestry or agriculture are often simultaneously exposed to chlorophenols, chlorophenoxy acids, dibenzodioxins and dibenzofurans, as well as other pesticides. Owing to the extreme toxicity and high carcinogenic potency of TCDD in laboratory animals, there has been a tendency to attribute most of the toxicity to dioxins and furans. However, as dioxins and furans exist only as impurities, any exposure to them means perhaps close to a million-fold higher exposure to the main chemical (McConnell et al., 1991; Vartiainen et al., 1995a). An exception is early years of 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) production, which caused remarkably high TCDD exposure (Flesch-Janys et al., 1995). In addition, the vapour pressure and water solubility of chlorophenols are substantially higher than those of dioxins and furans, favouring higher exposure via air and water. Also, absorption of chlorophenols through skin is probably better than that of dioxins (Wester et al., 1993; Pohjanvirta & Tuomisto, 1994). Therefore, the primary relevance of the measured concentrations of kinetically persistent 236 PESTICIDES dioxins/furans may be that they indicate past exposure to the main chemical, especially if the exposure is long-lasting. The slow pattern of the kinetics of dioxins and furans dictates that, for their own toxicity, a steady gradual exposure is more important than an occasional short exposure, unless this is very large. In fact, the long half-life of dioxins (5-10 years or more) (Pirkle et al., 1989; Wolfe et al., 1992; Flesch-Janys et al., 1994; Needham et al., 1994) means that a steady-state level will be achieved only in 30-40 years at a constant intake. This renders it unlikely that an occasional exposure, such as in forestry and agriculture, would significantly change the body balance, even if the concentrations of the main chemical increase dramatically during spraying or handling. The situation is quite different in chemical industries. If the workers are exposed continuously for years or decades (Fingerhut et al., 1991; Flesch-Janys et al., 1995), then extremely high dioxin levels (thousands ofpg/g fat) may be accumulated.
6.1 PCP poisoning
PCP and other higher chlorinated phenols act cellularly to uncouple oxidative phosphorylation and to inhibit A TPase and several other enzymes (Jorens & Schepens, 1993). This leads to excessive heat production and fever. Symptoms of acute poisoning include central nervous system disorders, dyspnoea, and hyperpyrexia, leading to cardiac arrest. Marked rigor mortis is typical (IARC, 1991 ). In general, acute human poisoning is seen only after large accidental or suicidal doses. The estimated minimal lethal dose in humans has been calculated to be 29 mg/kg of body weight (IARC, 1991 ). The human toxicology of PCP was recently reviewed by Jorens & Schepens ( 1993 ). In addition to metabolic, respiratory, and central nervous system effects (see above) seen only after an acute large oral dose, they classified the clinical features as effects on skin and haematopoietic tissue, renal effects, and gastrointestinal effects. Some of these effects. notably chloracne, are probably due to dioxin/furan impurities. These effects have occasionally been seen after heavy occupational exposure. In a Chinese PCP production plant, high prevalence of chloracne, increased urinary porphyrin excretion, and decreased motor nerve conduction velocities were observed in the high-exposure area (Cheng et al., 1993; Coenraads et al., 1994). These are likely due to dioxin-like impurities. Drinking-water exposure was associated with gastrointestina1 symptoms (nausea, pains, diarrhoea) and mild skin disorders (itching, eczema) (Lampi et al., 1993). The latter findings can be assumed to be due to chlorophenols, as no dioxin exposure was noted (Vartiainen et al., 1995b).
6.2 Immunological effects
Immunological abnormalities suggested in one study were T-cell activation and autoimmunity, functional immunodepression, and B-cell dysregulation (McConnachie & Zahalsky, 1991). These were found among 38 individuals living in log houses treated with PCP; hence, if true, the effects are more plausibly due to volatile PCP than to non-volatile impurities. T -lymphocyte dysfunction was also PENTACHLOROPHENOL 237 implied in 188 patients who were exposed to PCP (Daniel et al., 1995). In another study, no major clinical or laboratory signs of immune deficiency were found (Colosio et al., 1993).
6.3 Cancer
A number of studies since the 1970s implicated the group of chemicals including chlorophenols, chlorophenoxy acids, chlorinated dibenzo-p-dioxins, and chlorinated dibenzofurans in the causation of cancer, especially soft-tissue sarcoma and non-Hodgkin lymphoma (Lilienfeld & Gallo, 1989; Johnson, 1990; IARC. 1991 ). Owing to several inconsistencies, the case remained debatable. Several important contributions were published during the 1990s that give further information on such an association (Eriksson et al., 1990; Wigle et al., 1990; Zahm et al., 1990; Zober et al., 1990; Coggon et al., 1991; Fingerhut et al., 1991; Green, 1991; Manz et al., 1991; Saracci et al., 1991; Lampi et al., 1992b; Scherr et al., 1992; Smith & Christophers, 1992; Bertazzi et al., 1993; Bueno de Mesquita et al., 1993; Kogevinas et al., 1993, 1995; Lynge, 1993; Hardell et al., 1994; Mikoczy et al., 1994; Flesch-Janys et al., 1995). There is a tendency for the risk to be non-significant or borderline in many well-executed studies. This results from small numbers of very rare cancers, but it may also be an indication of the difficulties of accurate exposure assessment. Although individual studies are not always convincing, there is certain coherence in the results incriminating this group of chemicals in the risk of soft-tissue sarcoma, less so with non-Hodgkin lymphoma. The relative contribution of each particular chemical is even less clear. as several chemicals among the group are animal carcinogens and as there is a paucity of information on others. There was no indication in a large international multi centre cohort of a difference in soft-tissue sarcoma risk between those workers who were probably exposed to TCDD and those who were exposed only to the main chemical (Saracci et al., 1991 ). A similar observation has been made in several other studies (Pearce & Bethwaite, 1992; Zahm & Blair, 1992; Lynge, 1993). In nested case control studies (Kogevinas et al., 1995) within the previous cohort (Saracci et al., 1991 ), relatively high odds ratios were noted for soft-tissue sarcoma by exposure to both phenoxy acids and dioxins/furans. Again, risks were not specifically associated with those herbicides contaminated with TCDD. Individuals with PCP exposure were too few for conclusions to be drawn (Kogevinas et al., 1995). Generally, the role of chlorophenols remains more elusive than that of other chemicals of the group, because the exposed groups are smaller and therefore the number of cases of these relatively rare cancers is often only from zero to five. This decreases the statistical power in cohorts and increases the possibility of chance findings. In a relatively large case-control study (105 cases and 335 controls) on non Hodgkin lymphoma, Hardell et al. (1994) observed an odds ratio of 4.8 (Cl 2.7-8.8) for chlorophenols, with PCP being the most common type. There seemed to be a higher risk in those exposed for a longer time. In a similar case-control study on soft tissue sarcoma (237 cases and 237 controls), a risk ratio of 5.25 (95% Cl 1.69-16.34) was observed in a high-exposure chlorophenol group (Eriksson et al., 1990). 238 PESTICIDES
In a large US multicentre occupational cohort, there was a significant correlation between long-term exposure to dioxins/furans and soft-tissue sarcoma (based on three cases), but again there is no certainty on whether the causative agent (provided the correlation indicates a true causal relationship) was dioxins/furans or the bulk chemicals (Fingerhut et al., 1991). A more detailed picture was obtained in a German cohort in a herbicide-producing plant where the exposure of each worker during a 33-year period was modelled (Flesch-Janys et al., 1995), but still 2,4,5-trichlorophenol or 2,4,5-T exposure occurred simultaneously with dioxins. In this study, a dose-related increase was observed in total mortality, cancer mortality, and ischaemic heart disease mortality. The effect was seen especially in the highest quintile, and the estimated highest dioxin equivalent concentrations were extremely high, over 4000 pg/g blood fat (normal young population in Europe, 10-30 pg/g fat; older people, up to 100 pg/g fat) (WHO, 1989, 1996; Vartiainen et al., 1995b). Viewed against these concentrations, the dioxin concentrations in the studies of Eriksson et al. (1990) and Hardell et al. (1994) cannot be even close to those found in chemical industries; if their observations are true, the cancers are more likely due to chlorophenols/phenoxy acids than to contaminant dioxins. Some light may be thrown on the discussion of the roles of chlorophenols and their impurities by an episode in Finland, where 3500 inhabitants were potentially exposed to chlorophenols but not dioxins/furans for at least 15 years through contaminated groundwater (Lampi et al., 1990, 1992a; Vartiainen et al., 1995a,b). In this cohort, increased incidences of soft-tissue sarcoma (six cases) and non-Hodgkin lymphoma (12 cases) were observed during three successive 5-year periods (Lampi et al., 1992b). In a case--control study, a significantly elevated risk ratio was observed for non-Hodgkin lymphomas among persons who consumed fish from the downstream lake or drinking-water from the village waterworks (Lampi et al., 1992b ). This implies that chlorophenols alone without the contribution of the minor dioxin!furan impurities may be associated with cancer at relevant exposure levels. In conclusion, there is some, although not irrefutable, evidence that chlorophenol preparations, including PCP, may cause cancer in humans. There seems to be no reason to believe that the problem could be overcome by advocating more purified preparations. In fact, there is little evidence that minor carcinogenic impurities, such as TCDD and other chlorinated dioxins and furans, would be more important than the main chemical, and there is limited direct evidence on the basis of both animal experiments and epidemiological studies that pure chlorophenols may, at relevant concentrations, pose a risk of cancer in humans. As far as water is concerned, only the risk of chlorophenols is plausible, as the PCDD/PCDF impurities are practically non-soluble in drinking-water.
7. PROVISIONAL GUIDELINE VALUE
IARC (1990) classified PCP in Group 2B - the agent is possibly carcinogenic to humans, because of inadequate evidence of carcinogenicity in humans and sufficient evidence in experimental animals. There is suggestive, although inconclusive, evidence of the carcinogenicity of PCP from epidemiological studies of populations exposed to mixtures including PCP. There is conclusive PENTACHLOROPHENOL 239 evidence of carcinogenicity in one animal species, although there are notable variations in metabolism between experimental animals and humans. It was therefore considered prudent to treat PCP as a potential carcinogen. Adequate dose-response data for carcinogenicity are available only from toxicological studies in animals. Based on multistage modelling of tumour incidence in the US NTP bioassay without incorporation of a body surface area correction, although recognizing that there are interspecies differences in metabolism, the concentration of PCP associated with a 1o-s excess lifetime cancer risk is similar to the current guideline value. The current provisional guideline value. of 9 Jlg/litre is therefore retained.
8. REFERENCES
Agriculture Canada (1989) Annual report on residue and formula compliance momtoring for fiscal year 1988/89 Ottawa, Ontario, Agriculture Canada, Agri-Food Safety Division, Food Inspection Directorate. Ahlborg UG, Thunberg T (1980) Effects of2,3,7,8-tetrachlorodibenzo-p-dioxin on the in vivo and in vitro dechlorination of pentachlorophenol. Archives of toxicology, 40:55-61. Armstrong MJ, Galloway SM, Ashby J (1993) 2,4,6-Trichlorophenol (TCP) induces chromosome breakage and aneuploidy in vttro. Mutation research, 303:101-108. A TSDR ( 1994) Toxicological profile for pentachlorophenol Atlanta, GA, US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry. Bajpai R, Banerji S (1992) Bioremediation of soils contaminated with pentachlorophenol. Annals of the New York Academy ofSciences, 665:423--434. Bauchinger M et al. (1982) Chromosome changes in lymphocytes after occupational exposure to pentachlorophenol (PCP). Mutatlon research, 102:83-88. Bertazzi PA et al. (1993) Cancer incidence in a population accidentally exposed to 2,3,7,8- tetrachlorodibenzo-para-dioxin. Epidemiology, 4:398--406. Boberg EW et al. (1983) Strong evidence from studies with brachymorphic mice and pentachlorophenol that 1'-sulfooxysafrole is the major ultimate electrophilic and carcinogenic metabolite of 1' hydroxysafrole in mouse liver. Cancer research, 43:5163-5173. Bueno de Mesquita HB et al. (1993) Occupational exposure to phenoxy herbicides and chlorophenols and cancer mortality in The Netherlands. American journal ofindustnal medtcine, 23:289-300. Cheng WN et al. (1993) A health survey of workers in the pentachlorophenol section of a chemical manufacturing plant. Amencan journal of industrial medicine, 24·81-92. Coad S, Newhook RC (1992) PCP exposure for the Canadian general population: a multimedia analysis. Journal ofexposure analysis and environmental epidemiology, 2:391--413. Coenraads PJ et al. (1994) Chloracne. Some recent issues. Dermatologic clinics, 12:569-576. Coggon D, Pannett B, Winter P (1991) Mortality and incidence of cancer at four factories making phenoxy herbicides. British journal of industnal medicine, 48:173-178. Colosio C et al. (1993) Toxicological and immune findings in workers exposed to pentachlorophenol (PCP). Archives of environmental health, 48:81-88. Dahlhaus M, Almstadt E, Appel, KE (1994) The pentachlorophenol metabolite tetrachloro-p hydroquinone induces the formation of 8-hydroxy-2-deoxyguanosine in liver DNA of male
B6C3F1 mice. Toxicology letters, 74:265-274. Dahlhaus M et al. (1995) Induction of 8-hydroxy-2-deoxyguanosine and single-strand breaks in DNA ofV79 cells by tetrachloro-p-hydroquinone. Mutatwn research, 329:29-36. Daniel V et al. (1995) Impaired in-vitro lymphocyte responses in patients with elevated pentachlorophenol (PCP) blood levels. Archives of environmental health, 50:287-292. Eriksson M, Hardell L, Adami HO (1990) Exposure to dioxins as a risk factor for soft tissue sarcoma: a population based case--(:ontrol study. Journal ofthe National Cancer Institute, 82:486--490. 240 PESTICIDES
Exon JH, Koller LD (1982) Effects of transplacental exposure to chlorinated phenols Environmental health perspectives, 46:137-140. Exon JH, Koller LD (1983) Effects of chlorinated phenols on immunity in rats. Internattonal journal of immunopharmacology, 5: 131-136. Fingerhut MA et al. (1991) Cancer mortality in workers exposed to 2,3,7,8-tetrachlorodibenzo-p dioxin. New England journal of medic me, 324:212-218. Flesch-Janys D et al. (1994) First results of an investigation of the elimination of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/F) in occupationally exposed persons. Organohalogen compounds, 21:93-99. Flesch-Janys D et al. (1995) Exposure to polychlorinated dioxins and furans (PCDD/F) and mortality in a cohort of workers from a herbicide-producing plant m Hamburg, Federal Republic of Germany. Amencanjournal ofepidemiology, 142:1165-1175. Fountaine JE, Joshipura PB, Keliher PN (1976) Some observations regarding pentachlorophenol levels in Haverford Township, Pennsylvania. Water research, I 0:185-188. Frebortova J, Tatarkovicova V (1994) Trace enrichment of chlorinated phenols from drinking water on chemically bonded sorbents for high-performance liquid chromatography. Analyst, 119:1519-1523. Geyer H, Scheunert I, Korte F (1986) Bioconcentration potential of organic environmental chemicals in humans. Regulatory tox1cology and pharmacology, 6:313-347. Gifford JS et al. (1996) Pentachlorophenol (PCP), PCDD, PCDF and pesticide concentrations in a freshwater lake catchment. Chemosphere, 32:2097-2113. Goldstein JA et al. (1977) Effects of pentachlorophenol on hepatic drug-metabolizmg enzymes and porphyria related to contamination With chlorinated dibenzo-p-dioxms and dibenzofurans. Biochemical pharmacology, 26:1549-1557. Green LM (1991) A cohort mortality study of forestry workers exposed to phenoxy acid herbicides. Bntishjournal ofmdustnal med1cme, 48:234-238 Hardell L, Eriksson M, Degerman A ( 1994) Exposure to phenoxyacetic acids, chlorophenols, or organic solvents in relation to histopathology, stage, and anatomical localization of non Hodgkin's lymphoma. Cancer research, 54:2386-2389. Hattemer-Frey HA, Travis CC (1989) Pentachlorophenol: environmental partitioning and human exposure. Archives of environmental contammation and toxtcology, 18:482-489. IARC ( 1982) Chemicals, industrial processes and mdustries associated with cancer 1n humans, !ARC Monographs, Volumes I to 29 Lyon, InternatiOnal Agency for Research on Cancer, pp. 249-250 (IARC Monographs on the Evaluation of the Carcinogemc Risk of Chemicals to Humans, Suppl. 4). IARC (1986) Some halogenated hydrocarbons and pestlctde exposures Lyon, International Agency for Research on Cancer, pp. 319-356 (IARC Monographs on the EvaluatiOn of the Carcinogenic Risk of Chemicals to Humans, Vol. 41). IARC (1991) Occupational exposures in msect1c1de applicatwn, and some pesllcides Lyon, International Agency for Research on Cancer, pp. 3 71-402 (!ARC Monographs on the Evaluation ofCarcmogenic Risks to Humans, Vol. 53). Innes JR et al. ( 1969) Bioassay of pesticides and industrial chemicals for tumorigenicity in mice: a preliminary note. Journal ofthe National Cancer Inslilute, 42:1101-1114. Jansson K, Jansson V (1992a) Induction of micronuclei in V79 Chmese hamster cells by tetrachlorohydroquinone, a metabolite of pentachlorophenol. Mutatwn research, 279:205-208. Jansson K, Jansson V (1992b) Genotoxicity of 2,4,6-tnchlorophenol in V79 Chinese hamster cells. Mutation research, 280:175-179. Jekat FW et al. (1994) Effects of pentachlorophenol (PCP) on the pituitary and thyroidal hormone regulation in the rat. Toxicology letters, 71:9-25. Johnson ES (1990) Association between soft tissue sarcomas, malignant lymphomas, and phenoxy herbicides/chlorophenols: evidence from occupational cohort studies. Fundamental and apphed toxrcology, 14:219-234. Jorens PG, Schepens PJ (1993) Human pentachlorophenol poisoning. Human and expenmental toxicology, 12:4 79-495. PENTACHLOROPHENOL 241
Kerkvliet NI et al. (1982a) lmmunotoxicity of technical pentachlorophenol (PCP-T)· depressed humoral immune responses toT-dependent and T-independent antigen stimulation m PCP-T exposed mice. Fundamental and applied toxicology, 2:90-99 Kerkvliet NI, Baecher-Steppan L, Schmitz JA (1982b) lmmunotoxicity of pentachlorophenol (PCP): increased susceptibility to tumor growth in adult mice fed technical PCP-contaminated diets. Tox1cologv and applied pharmacology, 62:55-{)4 Kerkvliet NI, Brauner JA, Matlock JP (1985) Humoral immunotoxicity of polychlorinated diphenyl ethers, phenoxyphenols, dwxins and furans present as contaminants of technical grade pentachlorophenol. Toxicology, 36·307-324. K1mbrough RD, Linder RE (1978) The effect of technical and purified pentachlorophenol on the rat liver. Toxicology and applied pharmacologv, 46:151-162. Kogevinas M et al. ( 1993) Cancer incidence and mortality in women occupationally exposed to chlorophenoxy herbicides, chlorophenols, and dwxms. Cancer causes and control, 4.547-553. Kogevinas M et al (1995) Soft tissue sarcoma and non-Hodgkin's lymphoma in workers exposed to phenoxy herbicides, chlorophenols, and dioxins: two nested case-control studies. Ep1denuologv, 6:396--402. Koistmen J et al. (1995) Contamination of p1ke and sediment from the KymiJOki river by PCDEs, PCDDs, and PCDFs· contents and patterns compared to pike and sediment from the Bothnian Bay and seals from Lake Saimaa. Environmental ~c1ence and technology, 29:2541-2547. Kontsas H et al. (1995) Gas chromatographic-mass spectrometric determmation of chlorophenols m the urine of sawmill workers with past use of chlorophenol-contammg anti-stain agents Ana~vst, 120:1745-1749. Laine MM, Jorgensen KS (1996) Straw compost and bwremediated soil as inocula for the bioremediation of chlorophenol-contaminated soil. Applied enVIronmental microbwlogy, 62.1507-1513. Lampi P et al. (1990) Population exposure to chloropheno1s, dibenzo-p-dioxins and dibenzofurans after a prolonged ground water pollution by chlorophenols. Chemospherc, 20:625-634. Lam pi P et al. ( 1992a) Chlorophenols in lake bottom sediments· a retrospective study of drinking water contamination. Chemosphere, 24·1805-1824. Lamp1 P et al. (1992b) Cancer incidence following chlorophenol exposure in a community m southern Finland. Archrves of enVIronmental health, 47: 167-175. Lam pi Pet al. ( 1993) Long-term chlorophenol exposure causes gastrointestinal and skm symptoms in humans. In: Proceedmgs of the first Fmmsh conference of environmental scrences Kuop10, Kuopio University, pp. 412--415 (Kuopio University PublicatiOn 14). Liebl B et al. (1995) [Evaluation of wood preservative pollutants of indoor air.] Gesundhelfswesen, 57:476--488 (in German). Lilienfeld DE, Gallo MA (1989) 2,4-D, 2,4,5-T, and 2,3,7,8-TCDD: an overview Ep1demzolog1cal rev1ew, 11:28-58. Lynge E (1993) Cancer in phenoxy herbicide manufacturing workers in Denmark 1947-87 - an update. Cancer causes and control, 4·261-272. Manz A et al. (1991) Cancer mortality among workers in chemical plant contaminated with dioxin. Lancet, 338:959-964. McConnach1e PR, Zahalsky AC (1991) Immunological consequences of exposure to pentachlorophenol. Archrves of enVIronmental health, 46·249-253 McConnell EE et al. (1991) Toxicology and carcinogenesis studies of two grades of
pentachlorophenol in B6C3F 1 mice. Fundamental and applied tmxology, 17:519-532 Mikoczy Z, Schutz A, Hagmar L (1994) Cancer mc1dence and mortality among Swedish leather tanners. Occupatwnal and environmental medrcme, 51 :530-535. Naito S et al. (1994) Role of active oxygen species in DNA damage by pentachlorophenol metabolites. Mutatwn research, 310.79-88. National Cancer Institute (1979) Bwassay of 2,4,6-tnch/oropheno/ for possible carcmogemcity Washington, DC, US Department of Health, Education and Welfare (Technical Report Series No 155, DHEW Publication No (NIH)79-1711 ). 242 PESTICIDES
Needham LL et al. (1994) Half-life of2,3,7,8-tetrachlorodibenzo'p-dioxin in serum of Seveso adults: intenm report. Organohalogen compounds, 21:81-85. Orser CS, Lange CC (1994) Molecular analysis of pentachlorophenol degradation. BiodegradatiOn, 5:277-288. Paasivirta J et al. (1985) Polych lorinated phenols, guaiacols and catechols in the environment. Chemosphere, 14:469-491. Pearce N, Bethwaite P (1992) Increasing incidence of non-Hodgkin's lymphoma: occupational and environmental factors. Cancer research, 52:5496s-5500s. Pirkle JL et al. (1989) Estimates of the half-life of 2,3,7,8-tetrachlorodibenzo-p-dioxin in Vietnam veterans of Operation Ranch Hand. Journal of toxicology and environmental health, 27.165-171. Pohjanvirta R, Tuomisto J (1994) Short-term toxicity of 2,3,7,8-tetrachlodibenzo-p-dioxin (TCDD) in laboratory animals: effects, mechanisms and animal models. Pharmacological rev1ews, 46:485-549. Reigner BG, Bois FY, Tozer TN (1992a) Assessment of pentachlorophenol exposure in humans using the clearance concept Human expenmental toxicology, 11: 17-26. Reigner BG, Rigod JF, Tozer TN (1992b) Disposition, bioavailability, and serum protein binding of pentachlorophenol in the B6C3F, mouse. Pharmaceutical research, 9: I 053-1057. Reigner BG, Bois FY, Tozer TN (1993) Pentachlorophenol carcmogemcity- extrapolation of risk from mice to humans. Human expenmental tox1cology, 12.215-225. Sai-Kato K et al. (1995) Pentachlorophenol-induced oxidative DNA damage in mouse liver and protective effect of antioxidants. Food chemistry and toxicology, 33:877-882. Salkinoja-Salonen MS et al. (1989) Cleanup of old industrial sites. Advances m applied biotechnology, 4:347-367. Saracci R et al. (1991) Cancer mortality in workers exposed to chlorophenoxy herbicides and chlorophenols. Lancet, 338: I 027-1032. Scherr PA, Hutchison GB, Neiman RS (1992) Non-Hodgkin's lymphoma and occupational exposure. Cancer research, 52 (Sup pi. 19):5503s-5509s. Schwetz BA, Keeler PA, Gehring PJ (1974) The effect of purified and commercial grade pentachlorophenol on rat embryonal and fetal development. Toxicology and applied pharmacology, 28:151-161. Schwetz BA et al. (1978) Results of two-year toxicity and reproduction studies on pentachlorophenol in rats. In: Ranga Rao K, ed. Pentachlorophenol. New York, NY, Plenum Press, pp. 301-309. Seiler JP (1991) Pentachlorophenol. Mutation research, 257:27-47. Smith JG, Christophers AJ (1992) Phenoxy herbicides and chlorophenols: a case control study on soft tissue sarcoma and malignant lymphoma Bnllshjournal ofcancer, 65.442-448. Thompson TS, Treble RG ( 1995) Use of pine needles as an mdicator of atmospheric contamination by pentachlorophenol. Chemosphere, 31:4387-4392. Turnes Ml et al. (1994) Determination of chlorophenols in drinking water samples at the subnanogram per millilitre level by gas chromatography with atomic emission detection. Journal ofchromatography, 683:21-29. UBA (1996) Bewertung der UmweltgefahrhckeLt ausgewahlter Altsto.!Je dw·ch das Umweltbundesamt, Tell JJ (Texte 38/96). Berlin, Umweltsbundesamt, pp. 59-69 van Raaij JA et al. (1994) Reduction of thyroxine uptake into cerebrospinal fluid and rat bram by hexachlorobenzene and pentachlorophenol. Toxicology, 94:197-208 Vartiainen T et al. (1995a) Polychlorodibenzo-p-dioxin and polychlorodibenzofuran concentrations in lake sediments and fish after a ground water pollution with chlorophenols Chemosphere, 30:1439-1451 Vartiainen T et al. (1995b) Polychlorodibenzo-p-dwxin and polychlorodibenzofuran concentrations m human fat samples in a village after pollution of drinking water by chlorophenols Chemosphere, 30:1429-1438 Veningerova M et al. (1994) [Polychlorinated phenols in total diet.] Zeitschnft fur LebensmLttel Untersuchung und -Forschzmg, 199:317-321 (in German). Villena F et al. (1992) Morphological changes on nerves and histopathological effects on liver and kidney on rats by pentachlorophenol (PCP). Comparative bwchem1stry and physwlogy C, I 0 1·353-363. PENTACHLOROPHENOL 243
Waidyanatha S et al. (1994) Measurement of hemoglobin and albumin adducts of tetrachlorobenzoquinone. Chemtcal research m toxtcology, 7:463-468. Wang YJ, Lin JK (1995) Estimation of selected phenols in drinking water with in situ acetylation and study on the DNA damaging properties of polychlorinated phenols. Archtves of envtronmental contammation and toxicology, 28:537-542. Welsh JJ et al (1987) Teratogenic potential of purified pentachlorophenol and pentachloroanisole in subchronically exposed Sprague-Dawley rats. Food chemistry and toxtcology, 25:163-172. Wester RC et al. (1993) Percutaneous absorption of pentachlorophenol from soil. Fundamental and applied toxicology, 20:68-71. WHO (1987) Pentachlorophenol. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Cnteria 71). WHO (1989) Levels of PCBs, PCDDs, and PCDFs in breast milk Results of WHO-coordmated mterlaboratory quality control studies and analytical field studies. Copenhagen, World Health Organization Regional Office for Europe, p. 92 (WHO Environmental Health Series 34). WHO (1996) Levels of PCBs, PCDDs and PCDFs in human milk Second round of WHO coordinated exposure study Copenhagen, World Health Organization Regional Office for Europe (Environmental Health in Europe, No. 3). Wigle DT et al. (1990) Mortality study of Canadian male farm operators: non-Hodgkin' s lymphoma mortality and agricultural practices in Saskatchewan. Journal of the Nat10nal Cancer Institute, 82:575-582 .. Wolfe W et al. (1992) Dioxin half-life in veterans of Operation Ranch Hand. Organohalogen compounds, 10:239-242. Yuan JH et al (1994) Toxicokinetics of pentachlorophenol m the F344 rat. Gavage and dosed feed studies. Xenobiotica, 24:553-560. Zahm SH, Blair A (1992) PestiCides and non-Hodgkin's lymphoma. Cancer research, 52:5485s-5488s. Zahm SH et al. (1990) A case--{:ontrol study of non-Hodgkin's lymphoma and the herbicide 2,4- dichlorophenoxyacetic acid (2,4-D) in Eastern Nebraska. EpidemiOlogy, I :349-356. Zober A, Messerer P, Huber P (1990) Thirty-four-year mortality follow-up of BASF employees exposed to 2,3,7,8-TCDD after the 1953 accident. InternatiOnal arcluves of occupatiOnal and environmental health, 62:139-157.
TERBUTHYLAZINE (TBA)
First draft prepared by P. Chambon Laboratoire Santi Environnement Hygiene de Lyon, Lyon, France
Terbuthylazine was not evaluated in the 1993 WHO Guidelines for drinking water quality, and no IPCS Environmental Health Criteria monograph or evaluation by JMPR has been prepared. Terbuthylazine has been found in water and was considered by the Coordinating Committee for the updating of the Guidelines to be of high priority for evaluation.
I. GENERAL DESCRIPTION
1.1 Identity (Tom/in, 1994)
CAS no: 5915-41-3 Molecular formula: C9HI6CINs
The IUPAC name for terbuthylazine (TBA) IS 6-chloro-N-(1,1- dimethylethyl)-N-ethyl-1 ,3,5-triazine-2,4-diamine.
1.2 Physicochemical properties (Green, 1991)
Property Value Melting point 177-179°C Vapour pressure 0.15 mPa at 25°C Volatility 0.014 mg/m1 at 20°C Density 1.188 at 20°C Octanol-water partition coefficient 1096 Water solubility 8.5 mg/litre at 20°C Hydrolytic stability at pH I: 8 days (half-life in water at 20°C) at pH 5: 86 days at pH 7: >200 days at pH 9: >200 days at pH 13: 12 days
1.3 Major uses (Anonymous, 1989, 1995; Green, 1991; Schneider, 1995; Werner, 1996)
TBA is a herbicide that belongs to the chloro-triazine family. In plants, it acts as a powerful inhibitor of photosynthesis. The substance is taken up through both roots and leaves and is distributed throughout the plant after being taken up through the roots. This enables it to be used in both pre- and post-emergence treatment. TBA is a selective herbicide for maize, sorghum, potatoes, peas, sugar cane. vines, fruit 246 PESTICIDES trees, citrus, coffee, oil palm, cocoa, olives, rubber, and forestry in tree nurseries and new planting. It is particularly effective against annual dicotyledons.
1.4 Environmentalfate
Metabolism of IBA in maize was investigated using 14C-labelled IBA. When IBA was applied to the leaves, most of the radioactivity remained in situ; when applied to the roots, it quickly spread throughout the entire plant. IBA degrades rapidly through the following process: hydrolysis at the chlorine substituent (to form the hydroxy-derivative) or N-dealkylation of a side chain, preferably the ethyl chain; N-dealkylation of the second side chain with formation of 2-chloro-4,6-diamino-s-triazine (or hydroxy-diamino-s-triazine = ammeline ); and hydroxylation of the ammeline to 2-amino-4,6-dihydroxy-s-triazine (ammelide) (Green, 1991). In soil, metabolism studies using 14C-labelled IBA show cleavage of the side chains and hydrolysis of the chlorine substituent, followed by mineralization by cleavage of the heterocycle with the formation of nitrogen-containing derivatives and carbon dioxide. Bacteria and fungi are capable of breaking down s-triazines. Decomposition by means of photolysis was also observed on the surface of the soil. Degradation of TBA occurs under a variety of envirorimental conditions. The speed of decomposition is strongly influenced by temperature, moisture levels, microbial activity, pH, and aeration. Soil mobility studies showed an adsorption of IBA onto soil particles within 2 hours; adsorption increased with humus content of the soil. The mobility of IBA is lower than that of atrazine but highly dependent upon soil type (Green, 1991; Anonymous, 1995; Schneider, 1995; Werner, 1996). Degradation of IBA in natural water depends on the presence of sediments and biological activity. Although studies with Rhine River and pond water gave half-lives of> 1 year, newer findings indicate a half-life of approximately 50 days. In studies investigating the photolysis of IBA in aqueous solution, a half-life of > 1 month was estimated using a sun-like light source, in agreement with a reported half-life of about 3 months in natural sunlight. However, for very high sunlight intensities (full midday sunlight), the half-life was 39 hours (Green, 1991; Anonymous, 1995; Schneider, 1995; Palm & Zetzsch, 1996; Werner, 1996). IBA has a low vapour pressure and can therefore be expected to have little volatility. Traces of IBA in air are diluted by diffusion and degraded by reactions with hydroxyl radicals. Under these conditions, the half-life of IBA in air is approximately 1.5 days (Zetzsch, 1993).
2. ANALYTICAL METHODS
IBA is extracted from water by solid-liquid extraction on RP-C18 material (RP = reversed phase), eluted with a solvent and then separated, identified and quantified by high-performance liquid chromatography using ultraviolet detection at 220 nm. The limit of detection is about 0.1 ).lg/litre (ISO, 1997). TERBUTHYLAZINE (TBA) 247
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Air
No data are available, but results ofrain analysis (Palm, 1995) indicate that TBA concentrations in air would be very low.
3.2 Water
TBA levels were monitored in the River Po (Italy) over a period of 3 years (1988-1991); concentrations between 0.0 and 0.3 !J.g/litre were observed (Brambilla, 1993). In the Friuli-Venezia area in Italy, well-water samples from seven wells were analysed up to 29 times between October 1991 and April1992. TBA was detected in 7 of a total of 200 samples: 3 samples contained TBA concentrations between 0.1 and 0.2 IJ.g/litre, and 4 contained TBA levels above 0.2 IJ.g/litre (Barbina, 1993). In Germany, a large survey of TBA in water was carried out over a 3-year period (1990-1992). Out of 9565 samples analysed, "94.2% were negative, 5.3% were positive (below 0.1 !J.g/litre ), and 0.4% contained TBA at concentrations above 0.1 IJ.g/litre. Samples were taken from drinking-water, groundwater, and surface water (Wolter, 1993). Among the drinking-water samples, 0.7% were positive. The detection limit was not given but was probably around 0.01 IJ.g/litre. In the Rhine River in 1994, the mean TBA concentration found was 0.013 IJ.g/litre at Kembs and Laufenburg (Giiggi, 1995). In 1995, mean TBA concentrations had reached 0.020 !J.gllitre at Kembs and 0.022 IJ.g/litre at Laufenburg (Giiggi, 1996). In 1993, a survey was conducted at one of the main tributaries of the Nidda River (one of the tributaries of the Main River in Germany). Out of 105 samples analysed, 5 contained TBA concentrations above 0.1 IJ.g/litre, with a maximum concentration of 0.28 IJ.g/litre (Seel, 1994). In Switzerland between March 1988 and March 1989, TBA was found in Lake Zurich at a mean concentration of 0.013 IJ.g/litre. During the same period, rainwater was collected from roofs and analysed. Concentrations of TBA were between 0.01 and 0.198 IJ.g/litre. The presence of TBA could have resulted from direct volatilization during herbicide application or wind erosion of soil particles from treated areas (Buser, 1990). The groundwater situation in different countries was surveyed by the French Ministry of Agriculture and Fisheries. In Germany and Sweden, 22 out of 3204 samples and 6 out of 230 samples were positive for TBA (above 0.1 !J.gllitre), respectively (Dabene, 1993).
3.3 Food
In the USA, a large survey of pesticide residues on food products was carried out from 1989 to 1991: 6970 samples (80% domestic and 20% foreign) were taken from 81 varieties of vegetables and fruit. No TBA residues were found at a detection limit of 0.5 mg/kg (Schattenberg, 1992). 248 PESTICIDES
TBA application rates used on most annual crops are in the range of 0.4-2.0 kg/ha; at this concentration, no measurable residues ofTBA (0.01-0.1 mg/kg) were observed among harvest products. Residues were found only in young maize grown for silage (maximum 0.1 mg/kg); this maize was given experimentally to cattle and poultry, and no measurable residues were detected in meat, eggs, or milk (Green, 1991; Anonymous, 1995).
3.4 Estimated total exposure and relative contribution of drinking-water
At a concentration of 0.2 flg/litre in drinking-water, the daily intake of TBA would be 0.4 flg. Data are insufficient to allow the relative contribution of drinking water to total intake to be determined.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
14 C-labelled TBA administered orally to rats is quickly absorbed, completely metabolized, and excreted via the urine and faeces with a half-life of 16-17 hours (Green, 1991 ). After a single oral dose, at least 60% was absorbed from the gastrointestinal tract and 75% was excreted within the first 24 hours. After 168 hours, 99% of TBA was excreted. About 46%, 14%, and 2% of the administered dose was eliminated in bile, urine, and faeces, respectively (Werner, 1996). Seven days after a single oral administration of TBA at 0.5 mg/kg of body weight, tissue residues were below 0.007 mg of TBA equivalents per kg, except in kidneys (0.02-0.05 mg/kg), liver (0.01 mg/kg), and blood (0.01 mg/kg) (Green, 1991). Following dermal application to rats, 30% of the applied dose was found in urine and faeces (Werner, 1996). A metabolic study in the cow using 14C-labelled TBA at a dose of 1 mg/kg of body weight per day for 10 days showed that almost I 00% of the radioactivity was excreted in urine and faeces. About 0.1% was detected in milk (Werner, 1996). The main metabolic degradation of TBA occurs through oxidative de-ethylation, oxidation of one methyl group of the tert-butyl group, and subsequent conjugation of the alcohol with glucuronic acid. The alcohol may be further oxidized into carboxylic acid. Minor pathways have been described with the formation of sulfate esters of the alcohol derivative, dechlorination via glutathione with subsequent formation of mercapturic acid derivatives, and the formation of 2-hydroxytriazine metabolites (Werner, 1996).
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure (Werner, 1996)
The oral LD 50 for TBA in mice was 7700 mg/kg of body weight. Five male and five female Tif/RAI f rats received TBA at 2000 mg/kg of body weight. One female died on day 2 after administration. All surviving animals TERBUTHYLAZINE (TBA) 249 recovered within 6-7 days. In a second rat study, TBA was administered to 20 male and 20 female OF A.SD rats at dose levels of I 000, 1590, or 2510 mg/kg of body weight. Fifteen out of 40 animals (3, 6, and 6 at the low, middle, and high doses, respectively) died within 3 days. The LD 50 was found to be 1590 mg/kg of body weight. Clinical observation revealed piloerection, prostration, and diarrhoea. The main necropsy findings were marked congestion of the lung and autolysis of the alimentary canal.
In rats, the dermal LD 50 was above 2000 mg/kg of body weight, and the acute 3 inhalation LC 50 was above 5300 mg/m .
5.2 Short-term exposure (Werner, 1996)
TBA was fed to groups of 10 male and 10 female rats in the diet at dose levels of 0, 6, 30, 100, or 300 mg/kg of feed for 90 days. A treatment-related decrease in body-weight gain was observed in both sexes at dose levels of 100 and 300 mg/kg of feed. Changes in haematology and clinical chemistry parameters were observed without macroscopic or histopathological modifications; most were reversible during a 4-week recovery period given to the 300 mg/kg of feed group. The NOAEL was 30 mg/kg of feed, equal to 2.1 mg/kg of body weight per day. In a 1984 oral toxicity study, groups of five male and five female New Zealand white rabbits were administered TBA at dose levels of 0, 5, 50, or 500 mg/kg of body weight per day for 28 days on a 5 days/week basis. Because of high mortality, dose levels were reduced to 0, 5, 20, and 100 mg/kg of body weight per day 3 days after the start of treatment. Marked toxicity was still observed, with high mortality and body-weight loss at 5001100 mg/kg of body weight per day. Three males and two females at the top dose level were kept for a 14-day recovery period. Signs of systemic toxicity were noted at all dose levels, including sedation, dyspnoea, diarrhoea, and tremors. A decrease in food consumption was observed at all dose levels. Haematology revealed a decrease in red blood cell parameters at 500/100 mg/kg of body weight per day. A number of organ weights decreased, including liver (at all dose levels) and heart, thymus, and testes at 20 and 500/100 mg/kg of body weight per day. Organ-weight changes were mostly reversible during the 2-week recovery period. Congestion and haemorrhage of the respiratory and digestive tracts, thymus, bone marrow, spleen, mediastinum, and lymph nodes were observed in rabbits found dead. Additionally, atrophy of thymus, lymph nodes, and spleen as well as immature testes were observed in rabbits at the highest dose. It was not possible to determine the NOAEL in this study. In a 1987 oral toxicity study, groups of five male and five female New Zealand white rabbits were fed TBA at dose levels of 0, 0.05, 0.5, or 5 mg/kg of body weight per day for 28 days. No deaths occurred at the highest dose level, and no significant indications of any target organ toxicity were observed. The NOAEL in this study was 5 mg/kg of body weight per day, the highest dose tested. The difference in results between the 1984 and 1987 studies may be due to the different origins of the rabbits used. Administration of TBA to dogs in the diet for 1 year at dose levels of 0, 10, 50, 250, or 500 mg/kg of feed resulted in decreased body-weight gain and food 250 PESTICIDES consumption at dose levels ;:::so mg/kg of feed and slight reduction in some red blood cell parameters in females at dose levels ;:::250 mg/kg of feed, but no specific signs of toxicity were noted. The NOAEL in this study was 10 mg/kg of feed, equal to 0.4 mg/kg of body weight per day.
5.3 Long-term exposure and carcinogenicity (Werner, 1996)
Administration of TBA to mice (50 per sex per dose) in the diet at dose levels of 0, 30, 150, or 750 mg/kg of feed for 2 years resulted in an increased survival of treated males, which was statistically significant only at 30 and 750 mg/kg of feed. A moderate decrease in body-weight gain and food consumption was observed in both sexes at 750 mg/kg of feed. No treatment-related neoplastic findings were observed. The NOAEL in this study was 150 mg/kg of feed, equal to 15 mg/kg of body weight per day. In two lifetime studies, Tif/RAI f rats (1120 rats in total; number per sex per dose not specified) were fed TBA at dietary levels of 0, 30, 150, or 750 mg/kg of feed and 0, 6, or 30 mg/kg of feed. The second study was started 6 months after the beginning of the first one, after it became clear that, because of the effects on body weight, 30 mg/kg offeed was not the NOAEL. In the first study, an increased survival of high-dose males (60% survival compared with 22% in the control group) and a dose-dependent decrease in body weight gain in both sexes were noted. Haematology revealed effects on red blood cell parameters in mid- and high-dose females. Pathology revealed an increased incidence of non-neoplastic lesions in the liver, lung, thyroid, and testis. An increased incidence of mammary gland carcinomas was observed in high-dose females, but the number of fibroadenomas was simultaneously decreased at this dose. The incidence of mammary gland tumours was within the historical control range, and the number of females bearing mammary gland tumours was identical in control and high-dose females. The increased incidence of Leydig cell tumours in high-dose males seems to be related to the increased "time of risk," as most of the tumours were observed in old rats after 2 years or at terminal sacrifice. The NOAEL was not determined in this study. In the second study, the only effect noted was a slight decrease in body weight gain at 30 mg/kg of feed (equal to 1 mg/kg of body weight per day) in both sexes. The NOAEL in this study was 6 mg/kg of feed, equal to 0.22 mg/kg of body weight per day.
5.4 Reproductive and developmental toxicity (Werner, 1996)
Groups of 24 pregnant Tif/RAI f SPF rats were administered TBA by gastric intubation at dose levels of 0, 1, 5, or 30 mg/kg of body weight per day from days 6 to 15 of gestation. Decreased body-weight gain and food consumption were noted at the high dose level. No effects on reproductive parameters were noted. As indicated by delayed or absent ossification of phalanges, fetal development was slightly delayed at 30 mg/kg of body weight per day. No treatment-related external, TERBUTHYLAZINE (TBA) 251 visceral, or skeletal malformations or abnormalities were recorded. The maternal and fetal NOAELs were both determined to be 5 mg/kg of body weight per day. Groups of 16--22 New Zealand white rabbits were treated with TBA at dose levels of 0, 0.5, 1.5, or 4.5 mg/kg of body weight per day by oral intubation during days 7-19 of gestation. No significant effects on reproductive parameters were observed. Groups of 21 pregnant Russian rabbits (Chbb:HM) were administered TBA at dose levels of 0, 0.5, 1.5, or 5 mg/kg of body weight per day by oral intubation during days 7-19 of gestation. Increased body-weight loss was observed at the highest dose during the treatment period and was later compensated by a body weight gain during the post-treatment period. A dose-related decrease in food consumption was observed in the mid- and high-dose groups during the treatment period. No effects on reproductive or fetal parameters were observed. Based on the effect on food consumption at 1.5 mg/kg of body weight per day, the maternal NOAEL was 0.5 mg/kg of body weight per day. In the absence of any treatment-related fetal findings, the fetal NOAEL was 5 mg/kg of body weight per day, the highest dose tested. In a two-generation study, the effects on the reproductive performance of Sprague-Dawley rats were assessed by feeding them TBA in the diet at dose levels of 0, 6, 60, or 300 mg/kg of feed. The F0 generation consisted of groups of 32 male and 32 female rats. In the F 1 generation, groups of 28 male and 28 female rats were used. Animals of the F 0 generation were approximately 6 weeks old at the beginning of treatment; treatment was maintained for 70 days prior to mating (age: 16 weeks).
On day 21 post-partum, 28 male and 28 female pups (F 1) were retained. F0 males and females as well as surplus F1 pups were sacrificed. The selected F1 animals were maintained on their respective diets and mated at the age of 16 weeks. All F2• pups were sacrificed at day 21 post-partum. Apparently non-pregnant F, females were mated for the second time with males that had been successful at the first mating.
The F2b pups as well as the F1 males and females were sacrificed on day 4 post partum. At the high dose level, decreased body-weight gain, decreased food and water consumption, slightly higher number of infertile pairings, slightly higher pup mortality, and retarded pup growth were observed. At the mid-dose level, parental (decreased body-weight gain and food consumption) but no reproductive effects were observed. The NOAEL for parental effects was 6 mg/kg of feed, equal to 0.4 mg/kg of body weight per day; the NOAEL for reproductive effects was 60 mg/kg of feed, equal to 21.3 mg/kg of body weight per day.
5.5 Mutagenicity (Werner, 1996)
No increased incidence of back mutations was observed in Salmonella typhimurium strains TA98, TA100, TA1535, or TA1537 with or without metabolic activation at TBA concentrations up to 1 mg per plate. In a second study with the same strains plus Escherichia coli WP2 uvrA at TBA concentrations up to 5 mg per plate, no mutagenic activity could be detected with or without metabolic activation. With mouse lymphoma cells at TBA concentrations up to 1 mg/ml, no increase in mutant frequency was observed (with or without metabolic activation). In 252 PESTICIDES a test with Chinese hamster V79 cells, no induction of gene mutations resulted with TBA concentrations up to 380 mg/ml, with or without metabolic activation. In another test with Chinese hamster cells, no increased incidence of specific aberrations was observed at TBA concentrations up to 380 mg/ml, with or without metabolic activation. With human lymphocytes, no increased number of specific chromosomal aberrations was observed up to a TBA concentration of 500 mg/ml in experiments with or without metabolic activation. With human fibroblasts or rat hepatocytes, no induction of unscheduled DNA synthesis was observed at concentrations of TBA up to 125 and 1000 mg/ml, respectively. In in vivo tests, doses of TBA up to 3000 mglkg of body weight were administered to Chinese hamsters twice within 24 hours. No increased incidence of chromosomal aberrations was observed. A micronucleus test in mice with dose levels up to 5000 mg/kg of body weight gave the same results.
6. EFFECTS ON HUMANS
There are no data available on the effects of TBA on humans.
7. GUIDELINE VALUE
There is no evidence that TBA is carcinogenic or mutagenic. A TDI approach was therefore used in the derivation of a guideline value for TBA in drinking-water. A NOAEL of 0.22 mg/kg of body weight per day was identified in a 2-year toxicity and carcinogenicity study in rats for decreased body-weight gain at the next higher dose of 1 mg/kg of body weight per day. Using an uncertainty factor of 100 (for inter- and intraspecies variation), the TDI is 2.2 ~J.g/kg of body weight. Assuming a 60-kg person consumes 2 litres of drinking-water per day and allocating 10% of the TDI to drinking-water, a guideline value of 7 ~J.g/litre (rounded figure) can be calculated for TBA in drinking-water.
8. REFERENCES
Anonymous ( 1989) Brochure Gardoprimlherb1c1de-product profile March (C1ga-Geigy Document). Anonymous (1995) Assessment GS 13529 Terbuthyla=me- consumer safety 18 December (Ciba Geigy Document). Barbina MT (1993) Pesticide residues in groundwater in Friuh-Venezia Gtulia. In· Proceedmgs of the IX Symposium on Pest1c1de Chemistry. Pwcenza. I 1-13 October 1993, pp 729-738 (Ciba Geigy Document). Brambilla A (1993) The fate of atrazme pesticides in River Po water. Sc1ence of the total environment, 132:339-348 Dabene E (1993) Recherche de prodz11ts phytosamtmres dans /es eaux souterrames Resultats pour quelques pays· Allemagne. Etats-Ums. Grande-Bretagne. ltaile. Pays-Bas, Suede Ministere de !'Agriculture et des Peches maritimes, Bureau de I' Agriculture et des Ressources naturelles, April. Green OH (\991) Terbuthylazme. lnformatwn on the acllve substance September (Ctba-Geigy Document). TERBUTHYLAZINE (TBA) 253
Gilggi M ( 1995) Rheinwasseruntersuchung-Pjlanzenschut::mlftel. Jahres-bericht 1994 CIBA Forschungsdienste Zentrale Analytik, I February (Ciba-Geigy Document). Gilggi M (1996) Rheinwasseruntersuchung-Pflanzenschutzmtttel. Jahres-bericht 1995 CIBA Forschungsdienste Zentrale Analytik, 22 February (Ciba-Geigy Document). ISO (1997) Water quality- Determinatwn of selected plant treatment agents- Methods using htgh performance !tqutd chromatography with UV detectwn after so!td-liqutd extractwn. Geneva, InternatiOnal Organization for Standardtzation (ISOII369: 1997 (E)). Palm WU ( 1995) Pjlan::enschut::mlllel m der Atmosphare. Eine Literaturstudie. Hannover, Fraunhofer Institute ofToxtcology and Aerosol Research, 2 April (Ciba-Getgy Document). Palm WU, Zetzsch C (1996) Investigation of the photochemistry and quantum yields of triazines using polychromatic irradiation and UV -spectroscopy as analytical tools. International journal of environmental and analytical chemistry, 65:313-329. Schattenberg HJ (1992) Pesticide restdue survey of produce from 1989-1991. Journal of the Assocwtwn of Official Ana~vttcal Chemtsts, 75:925-932 Schneider M (1995) Terbuthylazine Fate and behavwur in sot!s and surface water 12 October (Ciba-Geigy Document). See! P (1994) Eintrage von Pflanzenschutzmittel in ein Fliessgewasser.-Versuch einer Bilanzterung. Vom Wasser, 83:357-372. Tomlin C, ed. (1994) The pesticide manual· a world compendium, lOth ed. Farnham, British Crop Protection Council. Werner C (1996) Toxicological evaluation of terbuthylazme (GS 13529) 25 October (Ciba-Geigy Document). Wolter R (1993) Pflanzenschutzmtttel-Funde im Wasser Seminar des WABOLU, 12 October (Ciba Geigy Document). Zetzsch C ( 1993) Determination of the OH-rate constant of TBA ads or bed on aerosols Fmal report Hannover, Fraunhofer Institute of Toxicology and Aerosol Research (Ciba-Geigy Document).
CHLOROFORM
First draft prepared by M.E.Meek Bureau of Chemical Hazards, Environmental Health Directorate Health Protection Branch, Health Canada Ottawa, Ontario, Canada
A guideline value for chloroform of 200 J.!g/litre was established in the 1993 WHO Guidelines for drinking-water quality. The guideline value was based on extrapolation of an observed increase in kidney tumours in male rats exposed to chloroform in drinking-water for 2 years, although it was recognized that chloroform may induce tumours through a non-genotoxic mechanism. The guideline value was calculated using the linearized multistage model to correspond with a 1o-s excess lifetime cancer risk. Because of the increasing database on mechanisms of induction of tumours by chloroform and because an IPCS Environmental Health Criteria monograph on chloroform was published in 1994, the Coordinating Committee for the updating of the WHO Guidelines recommended that chloroform be re-evaluated for the 1998 Addendum.
1. GENERAL DESCRIPTION
1.1 Identity
CAS no: 67-66-3
Molecular formula: CHCI3
Chloroform is the most commonly occurring trihalomethane (THM); THMs are halogen-substituted single-carbon compounds with the general formula CHX3•
1.2 Physicochemical properties1
Chloroform is degraded photochemically, is not flammable, and is soluble in most organic solvents. However, its solubility in water is limited. Phosgene and hydrochloric acid may be formed by chemical degradation. The most important physical properties of chloroform are presented below (IARC, 1979; Budvari et al., 1989).
Property Value Physical state Clear, colourless liquid Boiling point at 101.3 kPa 61.3°C Melting point -63.2°C
1 Conversion factor in air: 1 mg/m3 = 0.20 ppm. 256 DISINFECTANT BY-PRODUCT
Property Value Relative density (20°C) 1.484 Auto-ignition temperature >1000°C Water solubility (25°C) 7.5-9.3 g/litre Vapour density (I 01.3 kPa, 0°C) 4.36 kg/m 3 Vapour pressure 8.13 kPa at 0°C; 21.28 kPa at 20°C Stability air- and light-sensitive; breaks down to phosgene, hydrogen chloride, and chlorine Log octanol-water partition coefficient 1.97
1.3 Organoleptic properties
Chloroform has a characteristic odour and a burning, sweet taste. Its odour threshold values are 2.4 mg/litre in water and 420 mg/m3 in air (Budvari et al., 1989; ATSDR, 1993).
1.4 Major uses
Commercial production of chloroform was 440 000 t in 1987. Chloroform has been used primarily in the production of chlorodifluoromethane, although smaller amounts have also been used as solvents, as cleaning agents, and m fumigants. Although chloroform was used in the past as an anaesthetic and m proprietary medicines, these applications have been prohibited in many countries.
1.5 Environmenta/fate
It is assumed that most chloroform present in water is ultimately transferred to air as a result of its volatility. Chloroform has a residence time in the atmosphere of several months and is removed from the atmosphere through chemical transformation. It is resistant to biodegradation by aerobic microbial populations of soils and aquifers subsisting on endogenous substrates or supplemented with acetate. Biodegradation may occur under anaerobic conditions. Bioconcentration in freshwater fish is low. Depuration is rapid (WHO, 1994a).
2. ANALYTICAL METHODS
There are several methods for the analysis of chloroform in air, water, and biological materials. The majority of these methods are based on direct column injection, adsorption on activated adsorbent or condensation in a cool trap, then desorption by solvent extraction or evaporation by heating and subsequent gas chromatographic analysis. In water, detection limits range from 0.02 to 1 f..lg/litre (WHO, 1994a; ISO, 1997). CHLOROFORM 257
3. ENVIRONMENTAL LEVELS AND HUMAN EXPOSURE
3.1 Outdoor air
Chloroform levels in ambient air in remote areas of the USA range from 0.1 to 3 3 0.25 J.lg/m • In urban and source-dominated areas, concentrations are 0.3-9.9 J.lg/m 3 and 4.1-110 J.lg/m , respectively (ATSDR, 1993). The population-weighted mean concentration of chloroform at 17 urban sites sampled across Canada in 1989 was 0.2 J.lglm3 (Environment Canada, 1992). Hourly average concentrations of chloroform in the Netherlands, determined during 1979-1981, were generally 0.15 J.lg/m3 (estimated detection limit) or less, the maximum value being 10 J.lg/m3 (den Hartog, 1980, 1981 ). Average concentrations of chloroform during 1990 in four German cities (Berlin, Tiibingen, Freudenstadt, 3 3 and Leipzig) ranged from 0.26 to 0.9 J.lg/m ; the maximum value was 30 J.lg/m , detected in Tiibingen (Toxicology and Environmental Health Institute of Munich Technical University, 1992).
3.2 Indoor air
Concentrations of chloroform in indoor air are generally higher than those in ambient outdoor air owing primarily to volatilization during water use. In a population survey in the USA, the observed increase was approximately 85%, which was consistent with assumptions concerning daily water use and likely release of chloroform from water into air (Wallace, 1987). In a nationwide survey of 757 randomly selected one-family houses in Canada sampled over a 10-month period in 1991, the mean level of chloroform in indoor air 3 3 was 4.1 J.lg/m ; the maximum value was 69 J.lglm (Otson et al., 1992). Ullrich (1982) 3 reported comparable concentrations in indoor air (1-3 J.lg/m ) in Germany. Chloroform levels in the air of enclosed swimming pools are also increased as a result of transfer from water. They vary as a function of several factors, such as the degree of ventilation, the level of chlorination, water temperature, the degree of mixing at the water surface, and the quantity of organic precursors present (Lahl et al., 1981a). Over a period of 11 months, the levels of chloroform directly above the water surface in indoor public swimming pools in Bremen, Germany, ranged from 3 3 10 to 380 J.lg/m , with an average of about 100 J.lg/m (Batjer et al., 1980; Lahl et al., 1981a). Ullrich (1982) reported a similar mean value in four public swimming pools in Germany. In an experimental study in which the mean concentration of chloroform in water in an indoor swimming pool was increased from 159 to 553 J.lg/litre, corresponding mean air chloroform levels ranged from 2.9 to 8.0 mg/m3 (Levesque et al., 1994).
3.3 Water
Sources of chloroform in the aquatic environment include paper bleaching with chlorine, chlorination of recreational (pool) water, cooling water, and 258 DISINFECTANT BY-PRODUCT wastewater. Chloroform is present in drinking-water through direct contamination of the source and through formation from naturally occurring organic compounds during chlorination. The rate and degree of formation of chloroform during chlorination are a function primarily of the concentrations of chlorine and humic acid, temperature, and pH. Levels vary seasonally, with concentrations generally being greater in summer than in winter (WHO, 1994a). Concentrations of chloroform in groundwater vary widely, depending principally on proximity to hazardous waste sites (ATSDR, 1993). For example, chloroform was detected at levels ranging from 11 to 866 J..lg/litre in samples from five out of six monitoring wells drilled 64 m apart in a direction perpendicular to the northward flow of groundwater at a contaminated site in Pittman, Nevada, USA (the depth ofunconfined groundwater was 2-4 m at this selected site) (Kerfoot, 1987). Finished drinking-water collected in 1988-1989 from 35 US sources, 10 of which were located in California, in all four seasons contained median concentrations of chloroform ranging from 9.6 to 15 J..lg/litre. The overall median for all four seasons was 14 J..lg/litre (Krasner et al., 1989). In a 1987-1989 survey conducted in the USA, the mean chloroform concentration in finished water for surface water systems serving more than 10 000 people was 38.9 J..lg/litre (90th percentile, 74.4 J..lg/litre). The comparable mean value in the distribution system was 58.7 J..lg/litre (US EPA, 1992). In a national survey of the water supplies of 70 communities in Canada conducted during the winter of 1976-1977, concentrations of chloroform in treated water of the distribution system 0.8 km from the treatment plant averaged 22.7 J..lg/litre (Williams et al., 1980). Concentrations at 10 different locations in southern Ontario sampled in the early 1980s were 4.5-60 J..lg/litre in water leaving the treatment plant and 7.1-63 J..lg/litre 1.6 km from the plant (Oliver, 1983). In a more recent survey of samples collected during winter and summer of 1993 at 53 sites in nine provinces of Canada, chloroform was the major THM detected at all but three sites, these being groundwater sources that were treated with minimal chlorination. The contribution of chloroform to total THMs (ranging from 75.5 to 91.4%) was higher in summer than in winter and was slightly higher for chlorine chlorine versus chlorine-chloramine or ozone-chloramine treatment. Although the majority of treatment facilities had relatively low total THM levels (<50 J..lg/litre), a small number had relatively high levels (>I 00 J..lg/litre ), particularly in the summer (except for chlorine-chloramine disinfection) (Health Canada, 1995). Chloroform levels in drinking-water in 100 German cities sampled in 1977 ranged from <0.1 to 14.2 J..lg/litre and averaged 1.3 flg/litre. Concentrations were similar in other surveys conducted in Germany in the late 1970s and early 1980s (Lahl et al., 198la). Mean levels of chloroform in drinking-water in the Netherlands in 1994 ranged up to 8.9 flg/litre (Versteegh et al., 1996).
3.4 Food
In an early study in Germany, chloroform was detected in several foodstuffs, in particular decaffeinated coffee (20 J..lg/kg), olive oil (28 J..lg/kg), pork (1 0 J..lg/kg), CHLOROFORM 259 and sausages (17 flg/kg). Occasionally, concentrations were higher: up to 80 flg/kg in coffee and 90 flg/kg in sausages. Levels of 1-10 flg/kg were detected in flour products, potatoes, cod liver oil, margarine, lard, fish, mussels, and milk; levels in most foodstuffs, however, were less than 1 flg/kg (Bauer, 1981 ). Chloroform was detected in about 90 of 300 samples in a market basket survey of 231 "table-ready" foodstuffs (prepared and cooked as normally served) in the USA, most often in fat-containing samples (Daft, 1988). In a later account (Daft, 1989), it was reported that chloroform concentrations of 2-830 flg/kg of food were detected in 68% of 549 samples of foodstuffs obtained in a market basket survey (average of71 flg!kg). Chloroform was not detected in composite samples of meat/fish/poultry or in composite samples of oil/fat in 39 different foods in the USA, although it should be noted that the quantification limits were higher (18-28 flg!kg) than those in the studies described above. However, a chloroform concentration of 17 flg/litre was found in the composite of dairy foods (Entz et al., 1982). Concentrations of chloroform in soft drinks range from 3 to 50 flg/litre, with levels for cola being at the upper end of the range (Abdel-Rahman, 1982; Entz et al., 1982; Wallace et al., 1984).
3.5 Estimated total exposure and relative contribution of drinking-water
3 Based on a daily inhalation volume for adults of 20 m , a body weight of 60 kg, the assumption that 20 out of 24 hours are spent indoors, and the mean levels of 3 chloroform in indoor air presented above (1-4 flg/m ), the mean intake of chloroform from indoor air for the general population is estimated to be 0.3-1.1 flg/kg of body weight per day. Individuals may be exposed during showering to elevated concentrations of chloroform from chlorinated tap-water (Jo et al., 1990a,b ). Based on experimental studies with humans, these authors concluded that the contribution of dermal exposure was approximately equivalent to inhalation exposure during showering and that the average intake of chloroform (inhalation and dermal absorption) was 0.5 flg/kg of body weight per shower for a person weighing 70 kg. Based on a review of relevant estimates, Maxwell et al. ( 1991) concluded that the ratio of the dose of chloroform received over a lifetime from inhalation to that received from ingestion of drinking-water is probably in the range of 0.6-1.5 but could be as high as 5. 7. The ratio of the dose received dermally compared with that received orally over a lifetime from drinking-water was considered to be approximately 0.3 but could be as high as 1.8. Based on daily consumption of 2 litres of drinking-water for adults, a body weight of 60 kg, and the mean levels of chloroform presented above (generally <20 flg/litre), the estimated mean intake of chloroform from drinking-water for the general population is less than 0.7 flg/kg of body weight per day. Actual levels of exposure may be less than those estimated on the basis of mean levels in drinking water, as most of the chloroform would be expelled from drinking-water that is heated before consumption (tea, coffee, soups, sauces). For example, approximately 96% of the total volatile halogenated hydrocarbon fraction was eliminated in water 260 DISINFECTANT BY-PRODUCT boiling for 5 minutes, whereas 50-90% was eliminated upon heating at 70-90°C (Bauer, 1981 ). Owing to the wide variations in concentrations of chloroform in water supplies, intake from drinking-water could be considerably greater than estimated here for some segments of the general population. Based on a daily ingestion of solid foodstuffs for adults of 1.5 kg (WHO, 1994b ), a mean body weight of 60 kg, and the mean level and percent detection of chloroform in foodstuffs in the market basket survey reported by Daft ( 1989), the estimated intake of chloroform from foodstuffs is approximately 1 ~g/kg of body weight per day. Based on estimates of mean exposure from various media, therefore, the general population is exposed to chloroform principally in food, drinking-water, and indoor air, in approximately equivalent amounts. The estimated intake from outdoor air is considerably less. The total estimated mean intake is approximately 2-3 ~g/kg of body weight per day. For some individuals living in dwellings supplied with tap water containing relatively high concentrations of chloroform, estimated total intakes from drinking-water through ingestion, inhalation, and dermal contact are up to 10 ~g/kg of body weight per day. Pools are also an important source of exposure to chloroform for swimmers. Based on an experimentally determined relationship, Levesque et al. (1994) estimated that the daily dose of chloroform resulting from a 1-hour swim (65 ~g/kg of body weight per day) in conditions commonly found in public swimming pools is 141 times greater than that for a 10-minute shower and 93 times greater than that for tap-water ingestion.
4. KINETICS AND METABOLISM IN LABORATORY ANIMALS AND HUMANS
The kinetics and metabolism of chloroform were reviewed in WHO (1994a). Chloroform is well absorbed in animals and humans after oral administration, but the absorption kinetics are dependent upon the vehicle of delivery. After inhalation exposure in humans, 60-80% of the inhaled quantity is absorbed, with kinetics being dependent upon concentration and species-specific metabolic capacities. Chloroform is readily absorbed through the skin of humans and animals, and significant dermal absorption of chloroform from water while showering has been demonstrated (Jo et al., 1990a). Hydration of the skin appears to accelerate absorption of chloroform. Chloroform distributes throughout the whole body. with levels being highest in the fat. blood, liver, kidneys, lungs, and nervous system. Distribution is dependent on exposure route; extrahepatic tissues receive a higher dose from inhaled or dermally absorbed chloroform than from ingested chloroform. Placental transfer of chloroform has been demonstrated in several animal species and humans. Unmetabolized chloroform is retained longer in fat than in any other tissue. The oxidative biotransformation of chloroform is catalysed by cytochrome P-450 to produce trichloromethanol. Loss of hydrogen chloride from trichloromethanol produces phosgene as a reactive intermediate. Phosgene may be detoxified by reaction with water to produce carbon dioxide or with thiols, including glutathione and cysteine. to produce adducts. The reaction of phosgene with tissue CHLOROFORM 261 proteins is associated with cell damage and death. Little binding of chloroform metabolites to DNA is observed. Chloroform also undergoes cytochrome P-450 catalysed reductive biotransformation to produce the dichloromethyl radical, which becomes covalently bound to tissue lipids. A role for reductive biotransformation in the cytotoxicity of chloroform has not been established. In animals and humans exposed to chloroform, carbon dioxide and unchanged chloroform are eliminated in the expired air. The fraction of the dose eliminated as carbon dioxide varies with the dose and the species. The rate of biotransformation to carbon dioxide is higher in rodent (hamster, mouse, rat) hepatic and renal microsomes than in human hepatic and renal microsomes. Also, chloroform is biotransformed more rapidly in mouse than in rat renal microsomes.
5. EFFECTS ON EXPERIMENTAL ANIMALS AND IN VITRO TEST SYSTEMS
5.1 Acute exposure
The liver is the target organ for acute toxicity in rats and several strains of mice. Liver damage is characterized mainly by early fatty infiltration and balloon cells, progressing to centrilbbular necrosis and then massive necrosis. The kidney is the target organ in male mice of other more sensitive strains. The kidney damage starts with hydropic degeneration and progresses to necrosis of the proximal tubules. Significant renal toxicity has not been observed in female mice of any strain. Acute toxicity varies depending upon the strain, sex, and vehicle. In mice, the oral LD50 values range from 36 to 1366 mg/kg of body weight; for rats, they range from 450 to 2000 mg/kg ofbody weight (WHO, 1994a).
5.2 Short-term exposure
The most universally observed toxic effects of chloroform are liver and kidney damage. The severity of these effects per unit dose administered depends on the species, vehicle, and method by which the chloroform is administered. Many histological and biochemical parameters were examined in female and male CD1 mice (7-12 per sex per group) administered chloroform at 0, 50, 125, or 250 mg/kg of body weight per day in water by gavage for 14 and 90 days (Munson et al., 1982). After 90 days, a depression in the number of antibody-forming cells was found at the highest dose level in both sexes. In females at the highest dose level, a decrease in cell-mediated-type hypersensitivity was observed. Liver weight was increased after 90 days of exposure at all dose levels in females and at the highest dose level in males. After 90 days of exposure, the animals demonstrated tolerance against a challenging dose of 1000 mg/kg of body weight. Histological changes observed in the kidneys of all dosed animals included small intertubular collections of chronic inflammatory cells; in the liver, they included generalized hydropic degeneration of hepatocytes and occasional small focal collections of lymphocytes. In females, small amounts of extravasated bile were occasionally noted in the sinusoidal Kupffer cells. 262 DISINFECTANT BY-PRODUCT
Jorgenson & Rushbrook (1980) administered chloroform to female B6C3F1 mice for 90 days at concentrations of 0, 200, 400, 600, 900, 1800, or 2700 mg/litre of drinking-water (equal to 0, 34, 66, 92, 132, 263, and 400 mg/kg of body weight). In the first week of the experiment, some mice in the highest dose group died of dehydration as a result of reduced water consumption. Concentration-related depression of the central nervous system occurred. The only treatment-related histopathological findings consisted of a mild adaptive and transitory fatty change in the livers of animals dosed with 66 mg/kg of body weight per day or more and a mild lymphoid atrophy of the spleen at 92 mg/kg of body weight per day and higher dose levels. There is evidence that the vehicle in which chloroform is administered significantly affects its toxicity. Bull et al. (1986) reported that chloroform administered by gavage in corn oil was significantly more hepatotoxic than equivalent doses administered in an aqueous emulsion (2% Emulphor, polyoxyethylated vegetable oil, GAF Corp.) to male and female B6C3F 1 mice (10 per sex per group) administered doses of 0, 60, 130, or 270 mg/kg of body weight per day for 90 days, based on examination of serum hepatic enzymes and hepatic histopathology.
In female B6C3F 1 mice exposed for 1 week to chloroform vapour at 3 concentrations ofO, 4.9, 14.7, 49, 147,490, or 1470 mg/m , there was increased cell proliferation in the nasal passages. This response was markedly less than that in F344 rats exposed to similar concentrations (Mery et al., 1994). Palmer et al. (1979) exposed 10 male and 10 female SPF Sprague-Dawley rats to chloroform by intragastric gavage (in toothpaste) daily for 13 weeks. Dose levels were 0, 15, 30, 150, or 410 mg/kg of body weight per day. At 150 mg/kg of body weight per day, there was "distinct influence on relative liver and kidney weight" (significance not specified). At the highest dose, there was increased liver weight with fatty change and necrosis, gonadal atrophy in both sexes, and increased cellular proliferation in bone marrow. In a 90-day study by Chu et al. (1982), groups of 20 male and 20 female Sprague-Dawley rats were exposed to chloroform via drinking-water at dose levels of 0, 0.17, 1.3, 12, or 40 mg/day for males and 0, 0.12, 1.3, 9.5, or 29 mg/day for females; this was followed by 90 days of recovery. Water and food intake were reduced in the highest dose group. At the 40 mg/day level, mortality was increased. Upon histological examination, there were mild liver and thyroid lesions, especially in the highest dose group. In livers of both males and females, there was an increase in cytoplasmic homogeneity, density of the hepatocytes in the periportal area, mid-zonal and centrilobular increase in cytoplasmic volume, vacuolation due to fatty infiltration, and occasional nucleic vesiculation and hyperplasia of biliary epithelial cells. Thyroid lesions consisted of a reduction in follicular size and colloid density, increase in epithelial cell height, and occasional collapse of follicles. Liver and thyroid lesions diminished in severity during the 90-day recovery period. Jorgenson & Rushbrook (1980) administered chloroform in drinking-water to male Osborne-Mendel rats for 90 days at concentrations of 0, 200, 400, 600, 900, or 1800 mg/litre (equal to 0, 20, 38, 57, 81, and 160 mg/kg of body weight per day). A concentration-related central nervous system depression was seen. Body weights in CHLOROFORM 263 the highest dose group were reduced throughout the study. In biochemical investigations of serum, there were no important deviations from control values other than a dose-related increase in cholesterol at dose levels of 38 mg/kg of body weight per day or more after 60 days and a decrease in triglycerides in the highest dose group from 30 days onwards. After 90 days of administration, however, these parameters were affected in the two highest dose groups only. No dose-related histopathological changes were reported. Exposure of male F-344 rats for 1 week to chloroform vapour at concentrations ofO, 4.9, 14.7, 49, 147, 490, or 1470 mg/m3 resulted in concentration dependent lesions in the nasal passages. Changes included increased epithelial mucosubstances in the respiratory epithelium of the nasopharyngeal meatus and a complex set of responses in specific regions of the ethmoid turbinates. The NOEL 3 for these responses ranged from 14.7 to 490 mg/m , with histological changes and induced cell proliferation being the most sensitive parameters (Mery et al., 1994).
5.3 Long-term exposure
Beagle dogs (eight per sex per dose) were given chloroform in a toothpaste base in gelatin capsules, 6 days/week for 7.5 years, at doses of 15 or 30 mg/kg body weight (Heywood et al., 1979). After 6 weeks of treatment, there were significant increases in serum glutamate-pyruvate transaminase (SGPT) levels in dogs given the high dose. At the low dose level, significant increases were observed at 34 weeks and after. Similar effects were not observed in the vehicle control (16 dogs of each sex) or untreated control (8 dogs of each sex) groups. "Fatty cysts" of the liver were observed in both dose groups at the end of this study. The LOAEL in this study was 15 mg/kg of body weight per day.
5.4 Reproductive and developmental toxicity
Adverse effects on reproduction in ICR mice have been observed, but only at doses that are hepatotoxic (Borzelleca & Carchman, 1982). There are also some limited data to suggest that chloroform is toxic to the fetus in studies in Sprague Dawley rats, but only at doses that are maternally toxic (>20 mg/kg of body weight per day) (Thompson et al., 1974; Ruddick et al., 1983).
5.5 Mutagenicity and related end-points
The results of genotoxicity assays for chloroform must be considered in light of two important aspects that could potentially compromise interpretation. First, there is the possibility that ethylcarbonate and diethylcarbonate, produced by reaction of phosgene with the ethanol that is routinely added to USP chloroform, could generate false-positive results. Secondly, owing to its volatility, chloroform must be tested in a sealed system to avoid false-negative results. In three separate studies conducted under sealed conditions to ensure chloroform retention, results were negative in the Ames assay. In validated systems in Salmonella typhimurium and Escherichia coli, results have been negative both 264 DISINFECTANT BY-PRODUCT with and without metabolic activation. In only one uncommon test with Photobacterium phosphorum were positive results reported (Wecher & Scher, 1982). Chloroform did not induce gene mutations in V79 Chinese hamster cells (Sturrock, 1977) or chromosomal aberrations in human lymphocytes in vitro (Kirkland et al., 1981). In three of four micronucleus tests in mice, results were negative (Gocke et al., 1981; Salamone et al., 1981; Tsuchimoto & Matter, 1981). In the fourth micronucleus test, a weakly positive result was reported (San Augustin & Lim-Sylianco, 1978). The same authors found a positive effect in the mouse host mediated assay with Salmonella typhimurium TAI537 but not with TA1535. In investigations of the potential of chloroform to induce unscheduled DNA synthesis in vitro and in vivo (single doses of 238 and 477 mg/kg of body weight by gavage in corn oil) at the most sensitive site for tumour formation, the female mouse liver, results have been uniformly negative and are consistent with the suggestion that neither chloroform nor its metabolites react directly with DNA in vivo. No increase in DNA repair was observed in freshly prepared primary cultures of human hepatocytes from discarded surgical material from four different individuals exposed to concentrations as high as I mmol of chloroform per litre (Butterworth et al., 1989). Given the large number of sensitive assays to which chloroform has been submitted, it is noteworthy that the reported positive responses are so few. Furthermore, these few positive responses were randomly distributed among the various assays with no apparent pattern or clustering for any test system. Taken together, the weight of evidence indicates that neither chloroform nor its metabolites would appear to interact directly with DNA or possess genotoxic activity (WHO, 1994a).
5.6 Carcinogenicity
Roe et al. (1979) administered 0, 17, or 60 mg ofUSP-grade chloroform per kg of body weight per day in a toothpaste base (vehicle) to ICI mice (control group 104 animals per sex, dose groups 52 animals per sex) by gavage, 6 days/week for 80 weeks, followed by a 16-week observation period. The control toothpaste did not contain eucalyptol and peppermint oil, whereas the toothpaste containing chloroform did contain these substances. Treatment with chloroform resulted in slightly increased survival, especially in the males. The most common cause of death was respiratory failure. A slightly increased incidence of fatty degeneration was observed among the chloroform-treated animals, and there was a slight increase in total tumours in male mice. Renal tumours (3 hypernephromas and 5 cortical adenomas) were reported in 8 out of38 males in the high-dose group. In a second experiment by Roe et al. ( 1979), the influence of peppermint oil, eucalyptol, and chloroform was determined separately. In this study, male ICI mice received 60 mg of chloroform per kg of body weight daily, in a similar manner as in the study reported above. The vehicle control (toothpaste without chloroform, eucalyptol, and peppermint oil) and dose groups consisted of 260 and 52 male animals, respectively (the groups receiving a dose of peppermint oil or eucalyptol also consisted of 52 animals). Again, the survival in the chloroform-dosed group was CHLOROFORM 265 better than in the control group. However, administration of chloroform resulted in a kidney tumour frequency of 9/49 (2 hypernephromas and 7 adenomas), compared with 6/240 in controls. In a third study by Roe et al. ( 1979), a chloroform dose of 60 mg/kg of body weight in toothpaste (containing eucalyptol and peppermint oil) was administered daily to male mice (52 per group) of the ICI, CBA. C57BL, and CF1 strains for a period of 80 weeks. The chemical was also administered in arachis oil to male mice of the ICI strain. Each strain had its own control group. Terminal sacrifice was at 93, 97-99, 104, and 104 weeks for the CF1, ICI, C57BL, and CBA strains, respectively. In this study, a treatment-related increase in survival was found in all strains tested, except for the CF1 strain. Treatment with chloroform resulted in a higher incidence of renal changes in the CBA and CF1 strains but not in the C57BL strain. The cause of death in all four strains was renal neoplasia in combination with respiratory and renal disease. In the C57BL, CBA, and CFI strains, no changes in tumour frequencies were observed. In the ICI mice, after treatment with chloroform in either the toothpaste vehicle or arachis oil, there was an increase in the incidence of malignant kidney tumours.
In an NCI carcinogenicity study, B6C3F 1 mice (20 animals per sex in the control group; 50 animals per sex in the dosed groups) received USP-grade chloroform stabilized with ethanol (0.5-1 %) in corn oil 5 times a week by gavage (NCI, 1976a,b ). Dosing was stopped after 78 weeks, and the animals were sacrificed after 92 weeks. The dose levels changed after 18 weeks, resulting in time-weighted average dose levels of 138 (low) and 277 (high) mg of chloroform per kg of body weight per day for male mice and 238 (low) and 477 (high) mg of chloroform per kg of body weight per day for female mice. Administration of the higher dose of chloroform reduced survival in the female mice. Causes of death were related to the observed liver tumours, pulmonary inflammation, and cardiac thrombosis. This latter lesion was not observed in either the control or the low-dose group. There was a dose-related increased incidence of hepatocellular carcinomas in males and females. Mice presented clinical signs of illness (i.e. reduced food intake and an untidy appearance), but clear information on non-neoplastic lesions was not provided. Increases in tumour incidence were not observed in a carcinogenicity bioassay in which female B6C3F 1 mice were exposed to 0, 200, 400, 900, or 1800 mg of chloroform per litre of drinking-water (number of animals: 430, 430, 150, 50, and 50, respectively) for a period of 2 years. These concentrations (monitored by analysis) correspond to time-weighted average daily chloroform doses of 0, 34, 65, 130, and 263 mg/kg of body weight (Jorgenson et al., 1985). Matched controls (50 females) received an amount of water without chloroform equal to that consumed by the 1800 mg/litre group. Initially, 25% of the animals in the two highest dose groups died, but later on the death rate was more or less equal to that in the control group. After 3 months, livers of animals exposed to chloroform concentrations of 65 mg/kg of body weight per day or more had a higher fat content than those of the controls (as examined by chemical techniques). After 6 months, liver fat content was increased in all exposed groups. Data on organ weights were not provided. No treatment-related effects on either liver or total tumour incidence were observed. 266 DISINFECTANT BY-PRODUCT
Palmer et al. (1979) exposed 50 male and 50 female SPF Sprague-Dawley rats to 60 mg of chloroform per kg of body weight per day for 80 weeks, then sacrificed them after 95 weeks. The same number of controls was gavaged with toothpaste containing essential oils. Respiratory disease was observed in both sexes. At week 95, survival was 32% for exposed males and 22% for controls. For females, survival was 26% for exposed animals and 14% for controls. There were no significant differences in any tumour incidence. There was a significant decrease in plasma cholinesterase. Although there was a significant decrease in relative liver weight, there was no histological evidence of toxicity. Osbome-Mendel rats (20 animals per sex in the control group; 50 animals per sex per dose group) received USP-grade chloroform stabilized with ethanol (0.5-1 %) in corn oil 5 times a week by gavage (NCI, 1976a,b ). Dosing was stopped after 78 weeks, and the animals were sacrificed after Ill weeks. The dose levels changed after 23 weeks, resulting in time-weighted average daily dose levels of 90 and 180 mg of chloroform per kg of body weight for males and 100 and 200 mg of chloroform per kg of body weight for females. Administration of chloroform reduced survival in male and female rats in all dose groups. A clear pathological reason for this effect in the rats was not given. In male rats, dose-related increased frequencies of kidney epithelial tumours were observed. In the females, a non significant increase in the frequency of thyroid tumours was found. Rats presented clinical signs of illness (i.e. reduced food intake and an untidy appearance). However, clear information on non-neoplastic lesions was not provided. Reuber (1979) re-evaluated the histological sections of the NCI (1976a,b) study and reported the same neoplastic lesions as the NCI. In addition, he noted that chloroform-dosed female rats developed liver lesions that were not seen in the control females. Male Osborne-Mendel rats were exposed to 0, 200, 400, 900, or 1800 mg of chloroform per litre of drinking-water (number of animals: 330, 330, 150, 50, and 50, respectively) for a period of 2 years (Jorgenson et al., 1982). These concentrations (monitored by analysis) correspond to time-weighted average daily chloroform doses ofO, 19, 38, 81, and 160 mg/kg of body weight (Jorgenson et al., 1985). Matched controls received an amount of water without chloroform equal to that consumed by the 1800 mg/litre group. Additional rats were used for intermediate biochemical and histopathological examination. As a probable consequence of reduced drinking and reduced body weights, death rate was reduced with increasing chloroform dosage and in the matched control group. Biochemical examination of blood after 6, 12, and 18 months showed that chlorine, potassium, total iron, and albumin levels and the albumin/globulin ratio tended to be increased after chloroform treatment, whereas levels of cholesterol, triglycerides, and lactate dehydrogenase were decreased in all treated groups. These deviations were also observed in the matched controls, but the decreases in serum triglycerides and cholesterol levels were more severe at the two highest dose levels than in the matched control group. Data on org~n weights were not provided. The only clear dose-related effect was an increase in renal tubular cell adenomas and adenocarcinomas. From 38 mg/kg of body weight per day upwards, the increase in the frequency of all kidney tumours was statistically significant. Although increased CHLOROFORM 267 incidence of non-neoplastic histopathological effects was not reported, histological sections have been re-evaluated recently (ILSI, in preparation). In a range of initiation/promotion assays in the liver, chloroform has not initiated tumour induction. Although chloroform had promotional activity in several studies, particularly when administered in corn oil (Capel et al., 1979; Deml & Oesterle, 1985, 1987), it inhibited the growth or formation of precancerous or cancerous cells in the majority of the assays in liver (Pereira et al., 1985; Herren Freund & Pereira, 1986; Klaunig et al., 1986; Reddy et al., 1992) and in several in the gastrointestinal tract (Daniel et al., 1989, 1991).
5.6.1 Mechanism of carcinogenicity
The weight of the available evidence indicates that chloroform has little, if any, capability of inducing gene mutation or other types of direct damage to DNA. Moreover, chloroform does not appear capable of initiating hepatic tumours in mice or of inducing unscheduled DNA synthesis in vivo (WHO, 1994a). The pattern of chloroform-induced carcinogenicity in rodents in bioassays conducted to date can be summarized as follows. Chloroform induced hepatic tumours in B6C3F1 mice (males and females) when administered by gavage in corn oil at doses in the range of 138-477 mg/kg of body weight per day (NCI, 1976a,b). However, when similar doses were administered to the same strain in drinking water, hepatic tumours were not increased (Jorgenson et al., 1985). Liver tumours are observed, therefore, only in mice following administration by gavage in corn oil. This observation is consistent with those in initiation/promotion assays in which chloroform has promoted development of liver tumours, particularly when administered by gavage in a corn oil vehicle. Chloroform also induces renal tumours, but at lower rates than liver tumours in mice. Chloroform induced kidney tumours in male Osborne-Mendel rats at doses of 90-200 mg/kg of body weight per day in corn oil by gavage (NCI, 1976a,b ). However, in this strain, results were similar when the chemical was administered in drinking-water, indicating that the response is not entirely dependent on the vehicle used (Jorgenson et al., 1985). It should be noted, however, that at the higher doses in this study, there were significant reductions in body weight. In an early, more limited investigation, kidney tumours were increased in ICI mice but not in CBA, C57BL, or CF1 mice administered chloroform by gavage in toothpaste (Roe et al., 1979). Therefore, the tumorigenic response in the kidney, although observed in both rats and mice (males), is highly strain-specific. To investigate the possible role of replicative proliferative effects in the carcinogenicity of chloroform, a wide range of studies have been conducted in which replicative proliferative effects have been examined in similar strains of rats and mice exposed to similar doses or concentrations of chloroform, although for shorter periods, as in the principal carcinogenesis bioassays (Larson et al., 1993, 1994a,b,c, 1995a,b, 1996; Lipsky et al., 1993; Pereira, 1994; Templin et al., 1996a,b,c). Most of these studies involved evaluation of histopathological changes and cell proliferation in the kidney and liver, the latter determined as a BrdU labelling index in histological tissue sections. 268 DISINFECTANT BY-PRODUCT
The weight of evidence on genotoxicity, sex and strain specificity, and concordance of cytotoxicity, regenerative proliferation, and tumours is consistent with the hypothesis that marked cytotoxicity concomitant with a period of sustained cell proliferation may underlie a secondary mechanism for the non-linear induction by a metabolite of chloroform-related tumours. The weight of evidence is strongest in this regard for hepatic and renal tumours in mice but more limited for renal tumours in rats. Specifically, histopathological effects and regenerative proliferation were observed consistently following administration by gavage in corn oil but not following administration by gavage in drinking-water or continuous administration in drinking-water in the liver ofB6C3F1 mice (Larson et al., 1994a; Pereira, 1994) and kidney of F344 rats (Lipsky et al., 1993; Larson et al., 1994b). These data are consistent with the hypothesis that chloroform-induced cancer following gavage administration is secondary to events associated with induced cytolethality and cell proliferation. Results of available studies also indicate that the proliferative response is less when exposure is not continuous (e.g. inhalation for 5 days/week vs 7 days/week) (Larson et al., 1996; Templin et al., 1996c) and returns to baseline following a recovery period. Correlations for renal tumours in the rat are less clear primarily due to the relative paucity of data on intermediate end-points and renal cancer in various strains and both sexes. There is some evidence of variations in sensitivity among strains, with those with most severe renal disease being most susceptible; however, available data are limited in this regard. Tumours occur at the site of renal injury following exposure via both gavage in corn oil and drinking-water. Based on studies conducted primarily in a strain of rats in which tumours have not been observed, a mode of action for carcinogenicity in the kidney based on cytotoxicity and tubular cell regeneration is plausible. In a recent review article (Chiu et al., 1996), lack of observation of histopathological effects in test groups with tumours and lack of correlation between cytotoxicity and tumours in the kidneys of individual male rats at autopsy in the study of Jorgenson et al. (1982) are considered to be inconsistent with the above mentioned seemingly plausible mechanism. Slides from this bioassay have recently been re-evaluated (ILSI, in preparation). Based on studies conducted primarily in the F344 rat, available data are consistent with a mode of action for carcinogenicity in the kidney based on tubular cell regeneration. Studies in this strain indicate that chloroform causes damage and increases cell replication in the kidney at doses similar to those that induce tumours in Osborne-Mendel rats following gavage in corn oil for periods up to 3 weeks (Larson et al., 1995a,b). However, there has been no clear dose-response for renal damage or proliferation in F344 rats exposed to concentrations in drinking-water that were similar to those that induced tumours in Osborne-Mendel rats in the carcinogenesis bioassay of Jorgensen et al. (1985) (Larson et al., 1995b). In a single study in which the proliferative response was compared in F344 and Osborne Mendel rats at 2 days following a single gavage administration, it was concluded that these strains were about equally susceptible to chloroform-induced renal injury. although a statistically significant increase in labelling index was observed at a much CHLOROFORM 269 lower dose in the Osborne-Mendel rat (10 mg/kg of body weight) than in the F344 rat (90 mg/kg of body weight); this latter observation may have been a function of the low value in controls for the Osborne-Mendel rats. Data on the proliferative response in the strain in which renal tumours have been observed (Osborne-Mendel rats) are limited to examination at 2 days following a single administration by gavage in corn oil (Templin et al., 1996b ); studies in which the proliferative response was examined in Osborne-Mendel rats following administration in drinking-water have not been identified. Although the results of this study are not inconsistent with a mode of action of induction of tumours based on tubular cell regeneration, they are considered inadequate in themselves to quantitatively characterize the dose-response relationship for an intermediate end point for cancer induction.
5. 7 Other special studies
Balster & Borzelleca (1982) administered chloroform in water to male ICR mice (8-12 per group) and examined their performance in a battery of neurobehavioural tests (several exposure periods and several dose levels). The only effect observed was a reduced achievement in an operant behaviour test after dosing with 100 and 400 mg of chloroform per kg of body weight in water for 60 days. At the chloroform level of 400 mg/kg of body weight, about half the treated animals died. No adverse effects on behaviour were observed after 90 days of dosing with 31 mg of chloroform per kg of body weight in water. According to Burkhalter & Balster (1979), oral administration of chloroform to mice from 3 weeks before mating until the end of the lactating period (in both sexes the dose was 31 mg/kg of body weight) did not result in retardation of the development of responses to a battery of neurobehavioural tests in the pups.
5.8 Factors modifying toxicity
The in vivo toxicity of chloroform is modified by a range of factors. The rate of its biotransformation is a significant determinant of its toxicity. Hence, factors that increase or decrease chloroform biotransformation may alter the intensity of chloroform-induced toxicity. Moreover, the activities of the enzymes that metabolize chloroform may be increased or decreased by exposure to chemicals, and exposure to chloroform itself may alter chloroform metabolism. In addition to differences in the rates of chloroform bioactivation, treatments that alter susceptibility are also important determinants of chloroform-induced toxicity. Cellular glutathione concentrations are an important determinant of susceptibility, and perturbations of glutathione homeostasis may markedly affect the toxicity of chloroform. In experimental studies, the drinking-water contaminants monochloroacetic acid, dichloroacetic acid, and trichloroacetic acid have potentiated the toxicity of chloroform, although mechanisms were not elaborated (Davis, 1992; Davis & Berndt, 1992). 270 DISINFECTANT BY-PRODUCT
6. EFFECTS ON HUMANS
In general, chloroform elicits the same symptoms of toxicity in humans as in animals. In humans, anaesthesia may result in death as a result of respiratory and cardiac arrhythmias and failure. Renal tubular necrosis and renal dysfunction have also been observed in humans. The lowest levels at which liver toxicity due to occupational exposure to chloroform has been reported are in the range of 80-160 mg/m3 (with an exposure period of less than 4 months) in one study and in the range of 10-1000 mg/m3 (with exposure periods of 1--4 years) in another study (WHO, 1994a). The mean lethal oral dose for an adult is estimated to be about 45 g, but large interindividual differences in susceptibility occur. Analytical epidemiological studies of potential associations between ingestion of chlorinated drinking-water and colorectal cancer have been conducted in the USA in Wisconsin (Young et al., 1990), New York (Lawrence et al., 1984 ), and Iowa (Cantor et al., 1996). Similar investigations of bladder cancer have been conducted in Colorado, USA (McGeehin et al., 1993), Ontario, Canada (King & Marrett, 1996), and Iowa, USA (Cantor et al., 1996). Cumulative exposure to THMs was slightly higher in New York than in Wisconsin, but no increased colon cancer risk associated with THM exposure was observed in either study. Data reported thus far from a study in Iowa indicate that colon cancer was not associated with estimates of past exposure to chlorination by-products, but rectal cancer risk was associated with measures of exposure to chlorination by-products, duration of exposure to chlorinated water, and increasing amounts of lifetime exposure to THMs. Levels of THMs were not reported. At this time, the evidence for an association between exposure to THMs in drinking-water and rectal cancer must be considered inconclusive. No evidence is available to suggest an increased risk of colon cancer, but studies have been conducted in areas where cumulative exposures are generally low. Data reported thus far from a study in Iowa indicate that risk of bladder cancer is not associated with estimates of past exposure to chlorination by-products, except among men and smokers, for whom bladder cancer risk increased with duration of exposure after control for cigarette smoking. No increased risk of bladder cancer was associated with exposure to THMs in Colorado; cumulative exposure in this study was similar to that in New York and Wisconsin, where there was no increased risk for colon cancer. In Ontario, King & Marrett (1996) found an increased bladder cancer risk with increasing duration of exposure to THM levels; however, only after 35 or more years of exposure is the association statistically significant and of higher magnitude. The bladder cancer incidence was about 40% higher among persons exposed to greater than 1956 (Jlg/litre )-years THMs in water compared with those exposed to less than 584 (Jlg/litre)-years. Although it is not possible to conclude on the basis of available data that this association is causal, observation of associations in well conducted studies where exposures were greatest cannot be easily dismissed; however, the degree of evidence should be considered to be limited. In addition, it is not possible to attribute these excesses to chloroform per se, although it is generally the disinfection by-product present at highest concentration in drinking-water. CHLOROFORM 271
There have also b'een a smaller number of epidemiological investigations of associations between consumption of chlorinated drinking-water and developmental/reproductive effects. In a recent review of relevant information, Reif et al. ( 1996) concluded that the existing database was inadequate to determine an association between exposure to disinfection by-products and adverse reproductive and developmental effects, drawing attention particularly to the variability in exposure assessments and examined end-points and potential for exposure misclassification and confounding.
7. GUIDELINE VALUE
The weight of evidence for genotoxtctty is considered negative (WHO, 1994a). The weight of evidence for liver tumours in mice is consistent with a threshold mechanism of induction. Although it is plausible that kidney tumours in rats may similarly be associated with a threshold mechanism, there are some limitations of the database in this regard. A guideline value can therefore be developed on the basis of a TDI for threshold effects. The most universally observed toxic effect of chloroform is damage to the centrilobular region of the liver. The severity of these effects per unit dose administered depends on the species, vehicle, and method by which the chloroform is administered. The lowest dose at which liver damage has been observed is 15 mg/kg of body weight per day administered to beagle dogs in a toothpaste base over a period of7.5 years. Effects at lower doses were not examined. Somewhat higher doses are required to produce hepatotoxic effects in other species. Effects in the proximal tubules of the kidney cortex have been observed in male mice of sensitive strains and in both male and female rats of several strains. Levels inducing adverse histopathological effects in the range of 30 mg/kg of body weight per day have been reported in some studies in sensitive strains. Based on the study by Heywood et al. (1979) in which slight hepatotoxicity (increases in hepatic serum enzymes and fatty cysts) was observed in beagle dogs ingesting 15 mg/kg of body weight per day in toothpaste for 7.5 years, and incorporating an uncertainty factor of 1000 ( 100 for inter- and intraspecies variation and 10 for use of a LOAEL rather than a NOAEL and a subchronic study), a TDI of 13 ).lg per kg of body weight per day (corrected for 6 days/week dosing) is derived. Allocation of 50% of total intake to drinking-water is a reasonable default based on estimates of mean exposure that indicate that the general population is exposed to chloroform principally in food, drinking-water, and indoor air in approximately equivalent amounts and that most of the chloroform in indoor air is present as a result ofvolatilization from drinking-water. Moreover, the population is additionally exposed dermally to chloroform in drinking-water during showering. Based on an average body weight of 60 kg and daily ingestion of 2 litres of drinking-water, the guideline value is 200 ).lg/litre (rounded figure). It is noted that a drinking-water concentration for a 1o-s excess lifetime cancer risk estimated on the basis of the default linearized multistage model for renal tumours in rats is similar to the value developed on the basis of non-neoplastic effects. 272 DISINFECTANT BY-PRODUCT
It is cautioned that where local circumstances require that a choice be made between meeting microbiological guidelines or guidelines for disinfection by products such as chloroform, the microbiological quality must always take precedence. Efficient disinfection must never be compromised.
8. REFERENCES
Abdel-Rahman HS (1982) The presence of trihalomethanes in soft drinks. Journal of apphed toxicology, 2: 165-166. A TSDR (1993) Toxicological profile for chloroform (update). Atlanta, GA, US Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and Disease Registry. Balster RL, Borzelleca JF (1982) Behavioral toxicity of trihalomethane contaminants of drinking water in mice. Environmental health perspectives, 46:127-136. Biitjer Ketal. (1980) Chloroform emission into urban atmosphere. Chemosphere, 9:311-316. Bauer U (1981) [Human exposure to environmental chemicals- Investigations on volatile organic halogenated compounds in water, air, food and human tissues. Ill. Communication. results of investigations.] Zentralblatt fur Bakteriologie, M1krobiologie und Hygiene, 1 Abtedung Originate B. Hygiene, 174:200-237 (in German). Borzelleca JF, Carchman RA (1982) Effect of selected orgamc drmking water contammants on male reproduction. Research Triangle Park, NC, US Environmental Protection Agency (EPA 600-/1-82-009; PB82-25984 7). Budvari S, O'Neill M, Smith A, eds. (1989) The Merck index: an encyclopedia of chemicals, drugs and biologicals, IIth ed. Rahway, NJ, Merck and Co. Bull RJ et al. (1986) Enhancement of the hepatotoxicity of chloroform in B6C3F, mice by corn oil: implications for chloroform carcinogenesis. Environmental health perspectives, 69:49-58. Burkhalter JE, Balster RL (1979) Behavioral teratology evaluation of trichloromethane in mice Neurobehavioral toxicology, I: 199-205. Butterworth BE et al. ( 1989) Use of primary cultures of human hepatocytes in toxicology studies. Cancer research, 49.1075-1084. Cantor KP, Lynch CF, Hildeshiem M (1996) Chlorinated drinking water and nsk of glioma: a case control study in Iowa, USA. Epidemiology, 7(4) (Suppi.):S83. Cape! ID et al. (1979) The effect of chloroform ingestion on the growth of some murine tumours. European journal of cancer, 15:1485-1490. Chiu N et al. (1996) Characterization of cancer risk associated with exposure to chloroform. Journal of environmental sc1ence and health, Cl4:81-l 04. Chu I et al. (1982) Toxicity of trihalomethanes. 11. Reversibility of toxicological changes produced by chloroform, bromodichloromethane, chlorodibromomethane and bromoform in rats. Journal of environmental science and health, B 17:225-240. Daft JL (1988) Fumigant contamination during large-scale food sampling for analysis. Archives of environmental contamination and toxicology, 17:177-181. Daft JL (1989) Determination of fumigants and related chemicals in fatty and non-fatty foods. Journal of agncu/tura/ and food chemtstry, 37:560-564. Daniel FB et al. (1989) Chloroform inhibition of I ,2-dimethylhydrazine-induced gastrointestinal tract tumors in the Fisc her 344 rat. Fundamental and applied toxicology, 13:40--45. Daniel FB et al. (1991) Site-specific modulation of carcinogen-induced gastrointestinal tract nuclear anomalies in B6C3F, mice by chloroform. Anticancer research, !I :665-670. Davis ME (1992) Dichloroacetic acid and tnchloroacetic acid increase chloroform toxicity. Journal of toxicology and environmental health, 37:139-148. Davis ME, Berndt WO (1992) Sex differences in monochloroacetate pretreatment effects on chloroform toxicity in rats Fundamental and applied toxicology, 18:66-71. Dernl E, Oesterle D (1985) Dose-dependent promoting activity of chloroform in rat liver foci bioassay. Cancer letters, 29:59-63. CHLOROFORM 273
Deml E, Oesterle D (1987) Dose-response of promotion by polychlorinated biphenyls and chloroform m rat hver foci bioassay Archtves oftoxzcology, 60:209-211. den Hartog JC (1980, 1981) [Analysts of gaseous organic components 111 the outdoor azr.] Rijswijk, TNO, Prins Maurits Laboratory (Reports PML-TNO Nos. 1980-182, 1981-149, 1981-155, 1981-158) (in Dutch). Entz RC, Thomas KW, Diachenko GW (1982) Residues of volatile halocarbons in foods using headspace gas chromatography. Journal of agrrcultural and food chemtstry, 30:846-849. Environment Canada (1992) Volatzle organic compounds measurements 111 Canadtan urban and rural areas· /989-/99/ Ottawa, Ontario, Environment Canada, Technology Development Branch (Internal Report 92-75.16). Gocke E et al. ( 1981) Mutagenicity of cosmetics mgredients licensed by the European Communities. Mutation research, 90:91-109. Health Canada (1995) A natwnal survey of chlorination byproducts 111 Canadzan dnnkmg water. Ottawa, Ontano, Health Canada, Environmental Health Directorate (95-EHD-197). Herren-Freund SL, Pereira MA ( 1986) Carcinogenicity of by-products of disinfection m mouse and rat liver. Envtronmental health perspectives, 69:59-65. Heywood R et al. (1979) Safety evaluation of toothpaste containing chloroform. Ill. Long-term study in beagle dogs. Journal ofenvtronmental pathology and toxicology, 2:835-851. IARC (1979) Chloroform. In. Some halogenated hydrocarbons. Lyon, International Agency for Research on Cancer, pp. 401--427 (IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 20) ILSI (in preparation) An evaluatwn of EPA 's proposed gwdelmes for carcinogen nsk assessment using chloroform and dichloroacetate as case studzes. Report of an expert panel Washmgton, DC, International Life Sciences Institute, Health and Envtronmental Sciences Institute. ISO (1997) Water quality - Determination of htgh~v volatile halogenated hydrocarbons - Gas chromatographic methods. Geneva, International Organization for Standardization (ISOI0301. 1997 (E)). Jo WK, Weisel CP, Lioy PJ (1990a) Routes of chloroform exposure and body burden from showering with chlorinated tap water. Risk analysts, I 0:575-580. Jo WK, Weisel CP, Lioy PJ (1990b) Chloroform exposure and the health risk associated with multiple uses of chlorinated tap water. Risk ana~VSIS, I 0:581-585. Jorgenson TA, Rush brook CJ (1980) Effects of chloroform 111 the drinkmg water of rats and mzce Nmety-day subacute toxicity study Washington, DC, US Environmental Protection Agency (EPA-600/1-80-030; NTIS PB80-2191 08). Jorgenson TA, Rushbrook CJ, Jones DC (1982) Dose-response study of chloroform carcinogenesis in the mouse and rat: status report. Envzronmental health perspecttves, 46·141-149. Jorgenson TA et al. (1985) Carcmogemcity of chloroform in drinking water to male Osborne-Mendel rats and female B6C3F, mice. Fundamental and applied toxzcology, 5:760-769. Kerfoot HB ( 1987) Shallow-probe soil-gas sampling for mdication of ground-water contamination by chloroform. !nternatwnal journal ofenvzronmental analyttcal chemtstry, 30.167-181. Kmg WD, Marrett LD (1996) Case-control study of bladder cancer and chlorination by-products in treated water. Cancer causes and control, 7:596-604. Kirkland DJ, Smith KL, van Abbe NJ ( 1981) Failure of chloroform to induce chromosome damage or sister-chromatid exchanges in cultured human lymphocytes and fmlure to induce reversion in Eschericlua colt. Food and cosmettcs toxicology, 19:651-656 Klaunig JE, Ruch RJ, Pereira MA (1986) Carcmogenicity of chlorinated methane and ethane compounds admmistered in drinking water to mice. Envtronmental health perspectzves, 69:89-95. Krasner SW et al. (1989) The occurrence of disinfection by-products in U.S. dnnking water. Joumal of the Amencan Water Works Assoctatwn, 81:41-53. Lahl U et al ( 1981 a) Distribution and balance of volatile halogenated hydrocarbons in the water and mr of covered swimmmg pools using chlorine for water disinfection. Water research, 15:803-814. Larson JL, Wolf DC, Butterworth BE (1993) Acute hepatotoxic and nephrotoxic effects of
chloroform in male F-344 rats and female B6C3F 1 mice. Fundamental and apphed toxzcology, 20:302-315. 274 DISINFECTANT BY-PRODUCT
Larson JL, Wolf DC, Butterworth BE (1994a) Induced cytotoxicity and cell proliferation in the hepatocarcinogenicity of chlorofonn in female B6C3F I mice: comparison of administration by gavage in corn oil vs ad llbllum in drinking water. Fundamental and applied toxicology, 22:90--102. Larson JL, Wolf DC, Butterworth BE (1994b) Induced cytolethality and regenerative cell proliferation in the livers and kidneys of male B6C3F, mice given chlorofonn by gavage. Fundamental and applied toxicology, 23(4):537-543.
Larson JL et al. (1994c) The toxicity of !-week exposures to inhaled chorofonn in female B6C3F1 mice and male F-344 rats. Fundamental and applied toxicology, 22:431-446. Larson JL et al. (1995a) Toxicity and cell proliferation in the liver, kidneys and nasal passages of female F-344 rats, induced by chloroform administered by gavage. Food and chemical toxicology, 33(6):443-456. Larson JL, Wolf DC, Butterworth BE (1995b) Induced regenerative cell proliferation in livers and kidneys of male F-344 rats given chloroform in corn oil by gavage or ad libitum in drmking water. Toxicology, 95:73-86.
Larson JL et al. (1996) A 90-day chlorofonn inhalation study in female and male B6C3F 1 mice: implications for cancer risk assessment. Fundamental and applied toxicology, 30(1 ): 118-137. Lawrence CE et al. (1984) Trihalomethanes in drinking water and human colorectal cancer. Journal of the Natwnal Cancer Institute, 72:563-568. Levesque B et al. (1994) Evaluation of dermal and respiratory chloroform exposure in humans. Environmental health perspect1ves, I 02(12): I 082-1087. Lipsky MM, Skinner M, O'Connell C (1993) Effects of chloroform and bromodichloromethane on DNA synthesis m male F344 rat kidney. Environmental health perspectrves, 101 (Suppl. 5):249-252. Maxwell NI, Burmaster DE, Ozonoff D ( 1991) Trihalomethanes and maximum contaminant levels: the significance of inhalation, dermal exposures to chloroform in household water. Regulatory toxicology and pharmacology, 14:297-312. McGeehin MA et al. (1993) Case~ontrol study of bladder cancer and water disinfection in Colorado. American journal ofeprdemwlogy, 138(7):492-50 I.
Mery Set al. (1994) Nasal toxicity of chlorofonn in male F-344 rats and female B6C3F1 mice following a !-week inhalation exposure. Toxtcologyandappliedpharmacology, 125:214-227. Munson AE et al. (1982) Toxicology of organic drinking water contaminants: trichloromethane, bromodichloromethane, dibromochloromethane and tribromomethane. Environmental health perspectives, 46:117-126. NCI (1976a) Carcinogenesis bioassay of chloroform. Bethesda, MD, National Cancer Institute (NTIS PB-2640 18/AS). NCI (1976b) Report on carcmogenesis bioassay of chloroform Bethesda, MD, National Cancer Institute (NTIS PB-264-0 18). Oliver BG (1983) Dihaloacetonitriles in drinking water. Algae and fulvic acid as precursors. Environmental science and technology, 17:80-83. Otson R, Fellin P, Whitmore R (1992) A national pilot study of airborne VOCs 111 res1dences destgn and progress. Presented at the 1992 US Environmental Protection Agency/Air & Waste Water Management Association Symposium on Measurement of Toxic and Related Air Pollutants, Durham, NC, 4-8 May. Pittsburg, PA, Air & Waste Water Management Association. Palmer AK et al. (1979) Safety evaluation of toothpaste containing chloroform. II. Long term studies in rats. Journal of envrronmental pathology and toxtcology, 2·821-833. Pereira MA (1994) Route of admimstration detennines whether chlorofonn enhances or mhibits cell
proliferation in the liver ofB6C3F1 mice. Fundamental and applied toxicology, 23(1):87-92. Pereira MA, Knutsen GL, Herren-Freund SL (1985) Effect of subsequent treatment of chloroform or phenobarbital on the incidence of hver and lung tumors, initiated by ethylnitrosourea in 15 day old mice. Carcinogenesis, 6:203-207. Reddy TV et al. (1992) Chlorofonn inhibits the development of diethylnitrosamine-initiated, phenobarbital-promoted gamma-glutamyltranspeptidase and placental form glutathiOne S transferase positive foci in rat liver. Carcmogenes1s, 13:1325-1330. CHLOROFORM 275
Reif JS et al. (1996) Reproductive and developmental effects of disinfection by-products in drinking water. Environmental health perspectives, I 04: I 056-1061. Reuber MD ( 1979) Carcinogenicity of chloroform. Environmental health perspectives, 31: 171-182. Roe FJC et al. (1979) Safety evaluation of toothpaste containing chloroform. I. Long-term studies in mice. Journal ofenvironmental pathology and toxicology, 2:799-819. Ruddick JV et al. (1983) A teratological assessment of four trihalomethanes in the rat. Journal of environmental science and health, Bl8:333-349. Salamone MF, Heddle JA, Katz M (1981) Mutagenic activity of 41 compounds in the m v1vo micronucleus assay. In: de Serres FJ, Ashby J, eds. Evaluation of short-term tests for carcmogens. Report of the international collaborative study. Amsterdam, Elsevier North/Holland, pp. 686--697 (Progress in Mutation Research, Vol. I). San Augustin J, Lim-Sylianco CY (1978) Mutagenic and clastogenic effects of chloroform. Bulletm of the Philippine Biochemical Society, I: 17-23. Sturrock J (1977) Lack of mutagenic effect of halothane or chloroform on cultured cells using the azaguanine test system. British journal ofanaesthesia, 49:207-210. Templin MV et al. (1996a) Chloroform-induced cytotoxicity and regenerative cell proliferation in the kidneys and liver of BDF I mice [abstract]. The toxicologist, 30 (I part 2): 134. Templin MV et al. (1996b) Comparison of chloroform-induced toxicity in the kidneys, liver, and nasal passages of male Osborne-Mendel and F-344 rats. Cancer letters, 104:71-78. Templin MV et al. (1996c) A 90-day chloroform inhalation study in F-344 rats: profile of toxicity and relevance to cancer studies. Fundamental and applied toxicology, 32: I 09-125. Thompson DJ, Warner SD, Robinson VB (1974) Teratology studies on orally administered chloroform in the rat and rabbit. Toxicology and applied pharmacology, 29:348-357. Toxicology and Environmental Health Institute of Munich ·Technical University (1992) [Recommendations on chloroform emission data BUA plenary session in Merseburg, 24-25 August 1992] Berlin, Federal Office of the Environment (BUA) (in German). Tsuchimoto T, Matter BE (1981) Activity of coded compounds in the micronucleus test. In: de Serres FJ, Ashby J, eds. Evaluatwn of short-term tests for carcinogens Report of the international collaborative study Amsterdam, Elsevier North/Holland, pp. 705-711 (Progress in Mutation Research, Vol. I). Ullrich D (1982) [Organic halogen compounds in the air of some Berlin indoor swimming pools.] WoBoLu-Berlin, 1:50-52 (in German). US EPA (1992) Disinfection by-products field study data base Washington, DC, US Environmental Protection Agency, Office of Ground Water and Drinking Water, Technical Support Division. Versteegh et al. (1996) The quality of dnnkmg water in the Netherlands m 1994 Nieuwegein, KIWA (Report 1996/105, VROM 83231133). Wallace LA (1987) The total exposure assessment methodology (TEAM) study· Summary and analysis Vol I. US Environmental Protection Agency (EPA/600/6-87/002a). Wallace LA et al. (1984) Personal exposure to volatile organic compounds. I. Direct measurements in breathing-zone air, drinking water, food and exhaled breath. Environmental research, 35:293-319. Wecher RA, Scher S (1982) Bioassay procedures for identifying genotoxic agents using light emitting bacteria as indicator organisms. In: Seno M, Pazzagli M, eds. Luminescent assays: perspectives in endocrinology and climcal chenustry New York, NY, Raven Press Ltd., pp. 109-113. WHO ( 1994a) Chloroform. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Criteria 163). WHO ( 1994b) Assessmg human health nsks of chemicals· denvatwn of guidance values for health based exposure limits. Geneva, World Health Organization, International Programme on Chemical Safety (Environmental Health Cntena 170). Williams DT et al. (1980) Tnhalomethane levels in Canadian drinking water Environmental science research, 16.503-512. Young TB et al. ( 1990) Case-control study of colon cancer and volatile organics in Wisconsin municipal groundwater supphes In: Jolley RL et al., eds. Water chlorinatiOn Chemistry. environmental Impact und health effects Vol 6 Chelsea, Ml, Lewis Publishers, Inc., p. 373
Annex 1 List of participants in preparatory meetings
Coordinating Committee Meeting for the updating of the WHO Guidelines for drinking-water quality, WHO, Geneva, 13--15 December 1995
Members
P. Chambon, Department of Hygiene, Pasteur Institute of Lyon, Lyon, France
A. Havelaar, Microbiological Laboratory of Health Protection, National Institute of Public Health and the Environment (RIVM), Bilthoven, Netherlands
T. Hayakawa, Head, Drinking Water Quality Management Office, Department of Water Supply and Environmental Sanitation, Ministry of Health and Welfare, Tokyo, Japan
A. Jensen, Water Quality Institute (VKI), Horsholm, Denmark
G. Klein, Director, Institute for Water, Soil and Air Hygiene (WaBoLu), Federal Environment Agency, Bad Elster, Germany (Vice-Chairman)
Y. Magara, Director, Department of Water Supply Engineering, Institute of Public Health, Tokyo, Japan
J. Orme-Zavaleta, Associate Director, Health and Ecological Criteria Division, United States Environmental Protection Agency, Washington, DC, USA
T. Simons, European Commission, Directorate-General XI, Environment Quality and Natural Resources, Brussels, Belgium
P. Toft, Director, Bureau of Chemical Hazards, Health Canada, Ottawa, Canada
N. Yoshiguti, Section Head for Water Supply Engineering, Water Supply Division, Department of Water Supply and Environmental Sanitation, Ministry of Health and Welfare, Tokyo, Japan
Secretariat
H. Abouza"id, Environmental Health Risk Assessment and Management, Division of Environmental Health, WHO Regional Office for the Eastern Mediterranean, Alexandria, Egypt 278 ANNEXES
J. Bartram, Manager, Water and Wastes, WHO European Centre for Environment and Health, Rome, Italy (Co-Rapporteur)
H. Galal-Gorchev, Urban Environmental Health, Division of Operational Support in Environmental Health, World Health Organization, Geneva, Switzerland (Co-Rapporteur)
R. Helmer, Chief, Urban Environmental Health, Division of Operational Support in Environmental Health, World Health Organization, Geneva, Switzerland
J. Kenny, Urban Environmental Health, Division of Operational Support in Environmental Health, World Health Organization, Geneva, Switzerland
M. Y ounes, Chief, Assessment of Risk and Methodologies, International Programme on Chemical Safety, World Health Organization, Geneva, Switzerland
Working Group Meeting on Chemical Substances in Drinking water, WHO, Geneva, 22-26 April1997
Members
P. Chambon, Health Environment Hygiene Laboratory of Lyon, Lyon, France
J. Fawell, Water Research Centre, Medmenham, England
O.D. Hydes, Drinking Water Inspectorate, Department of the Environment, London, England
J. Kielhorn, Fraunhofer Institute for Toxicology and Aerosol Research, Hannover, Germany (Co-Rapporteur)
U. Lund, Water Quality Institute (VKI), Horsholm, Denmark
M.E. Meek, Environmental Health Centre, Health Canada, Ottawa, Canada (Chairperson)
E. V. Ohanian, Health and Ecological Criteria Division, United States Environmental Protection Agency, Washington, DC, USA (Co-Rapporteur)
K. Petersson Grawe, National Food Administration, Uppsala, Sweden
T. Simons, European Commission, Directorate-General XI, Environment Quality and Natural Resources, Brussels, Belgium ANNEXES 279
G.J.A. Speijers, Head of the Public Health Section of the Advisory Centre of Toxicology, National Institute of Public Health and Environmental Protection (RIVM), Bilthoven, Netherlands (Vice-Chairman)
B.H. Thomas, Acting Chief, Monitoring and Criteria Division, Bureau of Chemical Hazards, Environmental Health Centre, Health Canada, Ottawa, Canada
J. Tuomisto, National Public Health Institute, Division of Environmental Health, Kuopio, Finland
Observers
K. Bentley, Principal Adviser, Centre for Environmental Health, Woden, Australia
G. Ethier, Executive Director, Health and Environmental Science, International Council on Metals and the Environment, Ottawa, Canada
R. Gaunt, Occupational Physician, International Council on Metals and the Environment, Ottawa, Canada
W. Graham (representing GIF AP), European Affairs, Registration Manager, Europe Africa, Monsanto, Brussels, Belgium
B. Lang (representing GIF AP), Novartis Crop Protection AG, Basle, Switzerland
Y. Magara, Director, Department of Water Supply Engineering, Institute of Public Health, Tokyo, Japan
J. Meheus, Director, Water Quality Department, International Water Supply Association, Antwerp, Belgium
F.J. Murray (representing International Life Sciences Institute), Murray & Associates, San Jose, CA, USA
M. Richold (representing European Centre for Ecotoxicology and Toxicology [ECETOC]), Unilever, Environmental Safety Laboratory, Bedford, England
H. Tano, Section Chief, Drinking-water Quality Management, Department of Water Supply and Environmental Sanitation, Ministry of Health and Welfare, Tokyo, Japan
R. Uauy, INTA University of Chile, Santiago, Chile
A.M. van Dijk-Looijaard, KIWA N.V. Research and Consultancy, Nieuwegein, Netherlands 280 ANNEXES
Secretariat
J. Bartram, Manager, Water and Wastes, WHO European Centre for Environment and Health, Rome, Italy
H. Galal-Gorchev, Urban Environmental Health, Division of Operational Support in Environmental Health, World Health Organization, Geneva, Switzerland (Secretary)
R. Helmer, Chief, Urban Environmental Health, Division of Operational Support in Environmental Health, World Health Organization, Geneva, Switzerland
J. Kenny, Urban Environmental Health, Division of Operational Support in Environmental Health, World Health Organization, Geneva, Switzerland
G. Moy, Food Safety Unit, Programme of Food Safety and Food Aid, World Health Organization, Geneva, Switzerland
M. Sheffer, Scientific Editor, Ottawa, Canada
P. Toft, Associate Director, International Programme on Chemical Safety, World Health Organization, Geneva, Switzerland
F.X.R. van Leeuwen, WHO European Centre for Environment and Health, Bilthoven, Netherlands
M. Younes, Chief, Assessment of Risk and Methodologies, International Programme on Chemical Safety, World Health Organization, Geneva, Switzerland Annex2
Tables of guideline values
The following tables present a sununary of guideline values for chemicals in drinking water. Individual values should not be used directly from the tables. The guideline values must be used and interpreted in conjunction with the information contained in the text.
Notes to tables
(P) Provisional guideline value. This term is used for constituents for which there is some evidence of a potential hazard but where the available information on health effects is limited; or where an uncertainty factor greater than 1000 has been used in the derivation of the tolerable daily intake (TDI). Provisional guideline values are also recommended: (!) for substances for which the calculated guideline value would be below the practical quantification level, or below the level that can be achieved through practical treatment methods; or (2) where disinfection is likely to result in the guideline value being exceeded.
For substances that are considered to be carcinogenic, the guideline value is the concentration in drinking-water associated with an excess lifetime cancer risk of w-s (one additional cancer per I 00 000 of the population ingesting drinking-water containing the substance at the guideline value for 70 years). Concentrations associated with estimated excess lifetime cancer risks of 10-4 and 10..,; can be calculated by multiplying and dividing, respectively, the guideline value by 10. In cases in which the concentration associated with an excess lifetime cancer risk of w-s is not feasible as a result of inadequate analytical or treatment technology, a provisional guideline value is recommended at a practicable level and the estimated associated excess lifetime cancer risk presented. It should be emphasized that the guideline values for carcinogenic substances have been computed from hypothetical mathematical models that cannot be verified experimentally and that the values should be interpreted differently from TDI-based values because of the lack of precision of the models. At best, these values must be regarded as rough estimates of cancer risk. However, the models used are conservative and probably err on the side of caution. Moderate short-term exposure to levels exceeding the guideline value for carcinogens does not significantly affect the risk. 282 ANNEXES
Table I. Chemicals of health significance in drinking-water
A. Inorganic constituents
Guideline value Remarks (mgllitre)
boron 0.5 (P)"
copper 2 (P) Based on acute gastrointestinal effects
nickel 0.02 (P)
50 (acute) The sum of the ratio of the concentration of each to its respective (acute) guideline value should not exceed 1
nitrite (as Non 3 (acute) 0.2 (P) (chronic)
uranium 0.002 (P)
B. Organic constituents
Guideline value Remarks (llgllitre)
benzo[a]pyrene 0.7b For excess risk of 10-5
edetic acid (EDTA) 600 Applies to the free acid
microcystin-LR 1 (P) Applies to total microcystin LR (free plus cell-bound) ANNEXES 283
C. Pesticides
Guideline value Remarks (J.lg/litre)
bentazone 300
carbo fur an 7
cyanazine 0.6
1 ,2-dibromoethane 0.4-15b (P) For excess risk of 10-5
2,4-dich1orophenoxyacetic 30 acid (2,4-D)
1,2-dichloropropane 40 (P) (1,2-DCP)
diquat 10 (P)
pentachlorophenol 9b (P) For excess risk of 10-5
terbuthylazine (TBA) 7
D. Disinfectant by-product
Guideline value Remarks (J.lg/litre)
chloroform 200
Table 2. Chemicals not of health significance at concentrations normally found in drinking-water
Chemical Remarks
fluoranthene u
glyphosate u
U - It is unnecessary to recommend a health-based guideline value for these compounds because they are not hazardous to human health at concentrations normally found in drinking-water. This Addendum to Volume 2 of the second edition of the Guidelines fordrinking-water quality contains evaluations of a number of chemical substances that may be found in drinking- water, including inorganic substances (aluminium, boron, copper, nickel, nitrate, nitrite and uranium), organic substances (edetic acid, microcystin-LR, polynuclear aromatic hydrocarbons), pesticides (bentazone, carbo fur an, cyanazine, 1,2-dibromoethane, 2,4-D, 1,2-dichloropropane, diquat, glyphosate, pentachlorophenol and terbuthylazine), and a disinfectant by-product (chloroform). In this Addendum, certain guideline values established in the second edition of the Guidelines have been reviewed and, where necessary, updated in the light of new scientific information. Evaluations of chemical substances given in this Addendum supersede evaluations previously published in the second edition of the Guidelines. In addition, some chemical substances not included in the second edition are evaluated. It is emphasized that the guideline values recommended are not mandatory limits. Such limits should be set by national or regional authorities, using a risk-benefit approach and taking into consideration local environmental, social, economic and cultural conditions.