Drinking Water and Health, Volume 7: Disinfectants and Disinfectant By-Products (1987)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4 Chemistry and Toxicity of Selected Disinfectants and By-Products Volume 3 of the Drinking Water and Health series examined the toxicity of several major disinfectants and many of the by-products formed during drinking water disinfection. This chapter updates that material by assessing current research data. Recommendations for future research are also provided. Risk quantification is an essential tool for rationalizing regulatory ac- tions. Where sufficient data are available, quantitative risk assessments are calculated for the substances reviewed in this chapter. Quantitative risk assessment includes four distinct components: hazard identification, exposure assessment, dose-response assessment, and characterization of human risk at projected levels and patterns of exposure. Following a thorough review of the toxicological data, compounds were classified according to whether they were or were not known (or suspected) carcinogens. For carcinogens, the multistage model was chosen for ex- trapolating from the high doses used in animal studies to the lower doses common in the environment of humans. This model appears to have a greater biological basis than most other models and in most cases is more conservative, usually producing higher estimates of risk at low doses. It incorporates the reasonable assumption of background additivity and is thus linear at low doses. For carcinogens, the risk to humans is expressed as the probability that persons weighing 70 kg would develop cancer sometime in their lives as a consequence of ingesting 1 liter of water containing 1 fig of the substance daily over a lifetime of 70 years. Although risks to a 10-kg child were not calculated, the disproportionately high intake of drinking water by children compared with that of adults would place them at greater risk. 80

Selected Disinfectants and By-Products {31 For substances not identified as known or suspected carcinogens and for which there were adequate toxicity data from prolonged ingestion studies in humans or animals, the subcommittee calculated a suggested no-adverse- effect level (SNARL), using methods developed in earlier volumes of Drink- ing Water and Health and estimating dose-response relationships when data were sufficient. This conventional approach was taken by default in the absence of suitable low-dose extrapolation models and because a "safe" level has not been demonstrated for these noncarcinogenic effects. The SNARL was derived by estimating a no-observed-effect level (NOEL) for any given compound and then dividing it by an uncertainty or safety factor. Because of the pitfalls encountered in estimating NOELs, evidence supporting such a level in any given study was carefully weighed. Safety factors should be properly interpreted to indicate levels of confidence in the underlying studies. For some substances, the data base was adequate to permit an estimate of the magnitude of interspecies or intraspecies variability and to suggest a safety factor based on that estimation. Where such an estimate was not possible, safety factors devised in the first volume of Drinking Water and Health (1977, pp. 803-804) were used: 10 when satisfactory data from chronic epidemiological or clinical studies were used; 100 for well-conducted long-term animal studies; and 1,000 for short-term studies or studies with some potential inadequacies. Ingestion may not be the sole route of exposure to substances in drinking water that are examined in this review. Cooking, showers, bathing, swim- ming, and other activities could theoretically provide important toxic con- tributions. Given the absence of data on these noningestion routes, this report does not include specific estimates of their contribution to total exposure. Further, drinking water is not the only medium or source of exposure to many of the substances evaluated here. To allow for exposures through other routes, the subcommittee generally assumed that drinking water provides 20% of the total exposure to a given substance. CHLORINE CAS No. 7782-50-5 C12 Chlorine was previously reviewed in Volume 2 of Drinking Water and Health (NRC, 1980, pp. 18, 39, 144-1661. At room temperature, chlorine is a greenish-yellow gas. It has a melting point of-102°C and a boiling point of-35°C. Chlorine is widely used as a water supply disinfectant as well as an oxidizing or chlorinating agent in producing chlorinated organic compounds. When chlorine is added to water, the following reactions occur:

82 DRINKING WATER AND HEALTH Cl2 + H2O ~ HOCl + H+ + C1 HOC1 ~ H + + OCl Very little chlorine is available in molecular form (C12) at pH values greater than pH 3.0. The hypochlorous acid (HOC1) that is formed may further ionize to produce hypochlorite ion (OC1- ~ and hydrogen ion (H + ). The dissociation of hypochlorous acid to hypochlorite and hydrogen ions is dependent on the pH of the solution. Before analytical methods were capable of distinguishing free chlorine (HOCl + OC1- ~ from combined chlorine (chloramines), it was recognized that chlorine residuals depended on chlorine dose in a complex manner. Low concentrations of added chlorine produce an equivalent amount of chlorine residual (oxidant), but subsequent addition of chlorine causes a reduction in the residual. After the loss of most of the residual chlorine, a point is reached beyond which additional chlorine produces a chlorine residual that is clearly more effective as a bactericide but less stable with time of contact. This point is known as the "breakpoint." Thus, chlori- nation of drinking water is practiced in two distinct ways, depending on the level of nitrogen present and the level of chlorine added. Marginal chlorination, or the use of the first residual produced, is really chlor- amination. Breakpoint chlorination, or addition of chlorine beyond the dip in the curve of residual produced versus chlorine added, is free residual chlorination (White, 19721. HEALTH ASPECTS Based on relatively early literature, the American Conference of Govern- mental Industrial Hygienists (ACGIH, 1986) recommends a threshold limit value (TLV' expressed as an 8-hour time-weighted average for workroom air of 1 ppm (approximately 3 mg/m3) for occupational exposures to protect against chronic lung changes, accelerated aging, and corrosion of teeth. Observations in Humans Occupational and domestic poisonings to chlorine gas have been re- ported (Philipp et al., 19851. No other recent studies were found. Observations in Other Species . . Potential mutagenicity to germ cells was studied by Meter et al. (1985~. Oral administration of chlorine (pH 8.5) to B6C3F1 mice at 4 and 8 ma/ kg of body weight (bw) per day for 5 weeks induced significant increases in sperm-head abnormalities. In another study by Chang and Barrow

Selected Disinfectants and By-Products 83 (1984), sensory respiratory tolerance was shown to develop in F-344 male rats after repeated exposure up to 10 days to chlorine gas at 2.5 and 10 ppm. As indicated earlier in this report, the principal health concerns, other than exposure directly to chlorine gas, arise from the use of chlorine as a disinfectant in drinking water supplies, when various chlorinated by- products such as the trihalomethanes are formed. However, Vogt and coworkers (1979) showed in vivo production of chloroform after ingestion of sodium hypochlorite; several halogenated by-products produced in vivo were found in blood plasma and stomachs of rats 1 hour after NaOC1 injection; and various halogenated organics are known to be produced by chlorination of amino acids (Trehy and Bieber, 1981), nucleic acids (Oli- vieri et al., 1980), uracil (Dennis et al., 1978), and nucleotides (Hoyano etal.,1973~. Revis and co-workers (1986) studied the prevention, by calcium ion in drinking water, of atherosclerotic plaques and the effects on serum cho- lesterol concentrations induced in pigeons by lead and cadmium. Chlorine, chlorine dioxide, chlorite, and monochloramine were added individually to the drinking water of separate groups of pigeons at 2 and 15 mg/liter for 3 months. The chlorine was tested at two pH levels (6.5 and 8.5) to provide conditions under which essentially 99% HOCl and 99% OCl- were being administered. While methods were not specifically stated, apparently the investigators examined major blood vessels for lipid-con- taining material, counted the number of plaques, and measured their area on the vessel wall. Plaque formation was reduced but low-density- lipoprotein (LDL) cholesterol levels increased. The sample size in this study was too small with proportionately too few controls to produce statistically significant results. Marked effects were observed on serum thyroid T3 and T4 levels at very low doses of the disinfectants. The marked increases in serum T3 and T4 need to be further evaluated, as well as the relevance of the pigeon as an animal model. This study is also described in the section on chlorine dioxide. CHLORINE DIOXIDE CAS No. 10049-04-4 C1O2 Chlorine dioxide is a reddish-yellow gas that freezes at-59.5°C, boils at 10°C, and is explosive in air at concentrations of about 4% or more. It decomposes in water and dissolves in alkalies, forming a mixture of chlorite and chlorate. In addition to its utility in water treatment, it is used as a bleach for wood pulp, fats, and oils; a maturing agent for flour; a biocide; and an

84 DRINKING WATER AND HEALTH odor controller. The toxicity of chlorine dioxide was reviewed in Drinking Water and Health, Volume 4 (NRC, 1982, pp. 174-178~; the following material updates and reevaluates information on this disinfectant. METABOLISM Abdel-Rahman and coworkers (1980a, 1982) studied the absorption of 3 ml of a solution of 100 mg 36ClO2/liter of water given per os to male Wistar rats. The rate constant for absorption of the labeled C1 was 3.8/ hour. Within 72 hours, 31% of the label was excreted in the urine and 10% in the feces. The for in which Cl was excreted in feces was not identified; however, in buccal scrapings from monkeys, chlorine dioxide was reduced to a nonoxidizing substance rapidly (Bercz et al., 1982), suggesting that chlorine dioxide is rapidly altered after ingestion. Most of the labeled C1 found in the urine by Abdel-Rahman and associates was as chloride ion, with some chlorite and chlorate. Less than 5% of the administered dose of labeled Cl was found 72 hours after gavage in plasma, kidney, lung, stomach, duodenum, ileum, liver, spleen, and bone marrow. Moore and Calabrese (1980) administered 100 mg chlorine dioxide per liter of drinking water to A/J and C57L/J mice (18 mg/kg bw per day) for 30 days and studied possible effects on blood components, including glucose-6-phosphate dehydrogenase (G-6-PD), red blood cells, hemato- crit, mean corpuscular volume, mean corpuscular hemoglobin, mean cor- puscular hemoglobin concentration, reticulocyte count, and osmotic fragility. They found no significant effects on any of these parameters, though the G-6-PD activity in C57L/J mice was said to be reduced slightly. Bercz et al. (1986) summarized the conclusions from some of their studies of endocrine effects in Wistar and Sprague-Dawley rats and African Green monkeys, especially of chlorine dioxide on thyroid function. In both species ingestion of chlorine dioxide affects the mucosal surfaces of the alimentary tract and the chemical composition of nutrients and hor- mones within it, apparently by oxidation and covalent binding of bio- available iodide, which is ubiquitous in the digestive tract. Absorption of the iodinated molecules may be the mechanism for inhibition of activity by the thyroid and for an accelerated decrease in the concentration of thyroxin in the blood. HEALTH EFFECTS Observations in Humans Lubbers et al. (1981, 1982, 1983) exposed human volunteers to chlorine dioxide in a triphasic study; statistical analysis of the data was reported

Selected Disinfectants and By-Products 85 by Lubbers and Bianchine (1984) and Lubbers et al. (19841. In phase I, the acute effects of increasing doses were investigated (Lubbers et al., 19811. Ten normal volunteers ingested two 500-ml portions of water containing chlorine dioxide 4 hours apart every fourth day for a total of 6 dosing days. Each portion was consumed within 15 minutes. The con- centration of chlorine dioxide was increased in steps on each of the 5 subsequent dosing days, from 0.1 mg/liter on the first day to 24 mg/liter on the final day. The maximum dose for a 70-kg person was 0.34 mg/kg bw. Hematic, blood chemical, urinary, and other values of the volunteers who ingested chlorine dioxide did not differ significantly from those of 10 volunteers who ingested plain water. Nevertheless, the investigators did not rule out the possibility that effects might become significant upon increased exposure. In phase II, 10 volunteers drank 500 ml of a solution containing chlorine dioxide concentrations of 5 mg/liter of water daily for 12 weeks. Weekly physical examinations and laboratory studies of blood and urine showed a statistically significant group-time interaction (p < 0.05) for group mean urea nitrogen values in the volunteers who ingested chlorine dioxide, but the investigators questioned the clinical significance of these changes. Phase III was concerned only with the effects of consumption of water containing sodium chlorite on G-6-PD-deficient subjects, which will be described in the section on that compound. Michael et al. (1981) conducted a prospective study of 197 inhabitants of a rural village using water disinfected with chlorine dioxide. They compared the hematological profiles of 87 males and 110 females (23 being less than 15 years of age) from this village with those of a group of 1 12 people (48 males and 64 females, 12 of whom were under 15 years of age) using unchlorinated water. The chlorinated water contained chlor- ine dioxide concentrations of 0.25-1.11 mg/liter and free-chlorine con- centrations of 0.45-0.91 mg/liter during the 12-week period of the study. The concentrations of chlorite and chlorate in the water were 3.2-7.0 ma/ liter and 0.87-1.8 mg/liter, respectively. The water treatment plant op- erated for only ~ hours a day, which was responsible, at least in part, for the variable concentrations measured. Neither the exposed group nor the comparison group of 118 persons showed any significant changes in he- matocrit, hemoglobin, erythrocyte count, white-cell count, reticulocyte count, mean corpuscular volume, methemoglobin level, serum creatinine, or serum total bilirubin from the preexposure levels to those measured after 115 days of exposure. Only blood urea nitrogen (BUN) was changed, the mean values of the test population at the end of the experiment being lower than at the beginning, with a slight but opposite trend in the com- parison group.

86 OR ~ N K' NG WATER AN D H EALTH The researchers believed that they had not ruled out the possibility of transient effects, such as a transient hemolytic anemia. They pointed out the lack of racial and ethnic diversity in their study population. Only one person exhibited a G-6-PD deficiency; his erythrocyte count, hemoglobin, and hematocrit declined by the end of the 3 months of exposure, but without a noteworthy change in methemoglobin or in any other effect studied. These variables had returned at least partly to normal by 90 days after the end of the exposure. Analysis of variance showed a small but significant association of met- hemoglobin with sex, of total bilirubin with age, of BUN with exposure (as mentioned earlier), of hemoglobin with age, and an exposure-sex-age association with erythrocyte count. The investigators thought that BUN changes might reflect a mild dehydration during the summer weather prevailing at the end of the study and pointed to a similar pattern in the BUN/creatinine ratio consistent with this interpretation. They recom- mended that further research on chlorine dioxide disinfection focus on high-risk persons who might be especially susceptible to oxidants. Tuthill et al. (1982) examined a population that had used water disin- fected with chlorine dioxide in the 1940s, comparing its morbidity and mortality with those of a neighboring community. The chlorine dioxide- exposed group had a significantly greater proportion of premature births, but this difference between the two communities disappeared when the effect of the age of the mother on premature parturition was taken into account. The only other significant difference after compensating for var- iations in feeding methods was a greater postnatal loss of weight by infants born into the exposed population. Three older studies deserve mention here: those by Gloemme and Lund- gren (1957), Elkins (1959, pp. 89-90), and Ferris et al. (19671. Gloemme and Lundgren studied 12 men who had experienced acute symptoms usu- ally related to the respiratory tract, and who worked in a Swedish sulfite pulping plant during the mid-1950s. They were exposed usually to less than 0.1 ppm chlorine dioxide and of Cl, with occasional exposures to low concentrations of sulfur dioxide or to comparatively large concentra- tions of chlorine dioxide. Slight chronic bronchitis was identified in seven of these men. In one case, a detected bronchitis disappeared, proving reversibility of the lesion. There were complaints among the 12 men also of irritation of eyes, respiratory tract, and G.I. tract. Elkins found that 5 ppm of chlorine dioxide in air was definitely an irritant to the respiratory and the G.I. tracts. A concentration of 19 ppm of chlorine dioxide in the air within a bleach tank was reported by Elkins to have caused the death of a man assigned to work within the tank. Ferris et al. compared two populations of workers: one of 124 individuals working in a plant making Kraft paper and the other of 147 people working in a sulfite pulp mill and

Selected Disinfectants and By-Products {37 with an average duration of possible exposure to chlorine dioxide, chlorine, and sulfur dioxide of 6 times that of the workers in the paper mill. Ferris and coworkers found no significant differences in ventilatory function or respiratory symptoms between the two populations. Observations in Other Species Acute Effects No studies of acute effects were found. Subchronic Elects Bercz et al. (1982) studied possible hematological effects in African Green monkeys given water that contained 100 mg chlorine dioxide/liter to yield a daily dose of 9.5 mg/kg bw during a 6- week period. This dose was without effects on measured components, but the monkeys were said not to have tolerated higher doses. Monkeys given chlorine dioxide at this dose level, but not those given 3.5 mg/kg bw per day, had decreased serum levels of thyroxine. The effect on secretions of thyroxine appears not to have been due to formation of chlorite or chlorate, as daily doses of 43 or 44 mg/kg bw of these compounds did not alter the concentration of thyroxine in the serum. Abdel-Rahman et al. (1980b) and Couri and Abdel-Rahman (1980) gave male Swiss Webster mice and Sprague-Dawley rats 1, 10, 100, or 1,000 mg chlorine dioxide per liter of drinking water. If the rats drank 0.1 ml/ kg bw per day and the mice 0.18 ml/kg bw per day, the first three concentrations would yield daily doses of 0.1, 1, and 10 mg/kg bw for rats and 0.18, 1.8, and 18 mg/kg bw for mice. (Because the highest concentration was expected to induce a decrease in water consumption, a dose corresponding to that concentration was not estimated.) At 2, 4, 6, and 12 months of exposure, the following were assayed: glutathione reductase, glutathione peroxidase, catalase, glutathione, and methemo- globin. There were various statistically significant changes, as judged by use of multiple t-tests, but there was no consistency with respect to dose or to period of exposure, except for an increase in both species in catalase activity at 1,000 mg/liter and at 10 and 100 mg/liter in mice. At no time nor at any dose was there an increased methemoglobinemia in either species. In another experiment from the same laboratory (Abdel-Rahman et al., 1984), osmotic fragility was studied in rats given drinking water containing 10 to 1,000 mg chlorine dioxide/liter (about 1 to 100 mg/kg bw per day). There was a dose-related increase in resistance of erythrocytes to hemolysis in hypotonic media (i.e., decreased fragility) in animals given water con- taining 10 to 100 mg/liter. When the concentration was increased from 100 to 1,000 mg/liter, the increase in resistance was not proportional to

88 DRINKING WATER AND HEALTH the earlier ones, perhaps because of the markedly decreased water con- sumption mentioned above. Mutagenicity Meter et al. (1985) observed no increase in sperm-head or chromosomal aberrations or in micronuclei formation in CD-1 mice administered aqueous chlorine dioxide by Savage (3.2, 8, or 16 mg/kg bw per day) for 5 days. The sperm heads were examined at 1, 3, and 5 weeks after the last dose, to detect effects at all stages of spermatogenesis. No evidence that Clot induces mutational change was found. Carcinogenicity No investigation of possible carcinogenicity of C1O2 was found. Reproductive Toxicity Abdel-Rahman et al. (1984) found a dose-re- lated decrease in testicular uptake of 3H-thymidine in male rats given 10 or 100 mg of chlorine dioxide/liter or 10 mg/kg bw per day) in their drinking water (65% and 38% of control levels, respectively). This sug- gests a reduction in cell division in the testes but does not indicate what effect, if any, there was on spermatogenesis or on hormone production. Teratogenicity Orme et al. (1985) gave female rats drinking water containing chlorine dioxide concentrations of 2, 20, or 100 mg/liter starting 2 weeks prior to mating and continuing through lactation, 21 days after parturition. At the highest concentration, 100 mg/liter (14 mg/kg bw per day), there was a significant depression of the concentration of thyroxine in serum in the pups, but not in the dams, at the time of weaning. No significant thyroid effects were seen in pups from the group ingesting 20 mg of chlorine dioxide/liter. There was a decrease (p = 0.08) in ex- ploratory and locomotor activity in pups born to dams given 100 mg/liter but not in pups from dams given 20 mg/liter (3 mg/kg bw per day). In a second experiment, pups born to dams drinking plain water were admin- istered chlorine dioxide at concentrations of 14 mg/kg bw by stomach tube per day between days 5 and 20 after birth. There was a larger depression in serum thyroxine and a greater and more consistent delay in development of exploratory and locomotor activity (p < 0.05) than in the first exper- iment. Taylor and Pfohl ( 1985) observed a significant reduction in cell number, as judged by total DNA content, in the cerebella of rat pups born to dams given water containing chlorine dioxide concentrations of 100 mg/liter through gestation and lactation. Pups given 14 mg/kg bw per day by stomach tube had reduced numbers of cells in both cerebellum and fore- brain at 11 days postpartum and exercised less than normal on a voluntary running wheel at 50-60 days postpartum (though administration of chlor- ine dioxide had ended at 20 days of age).

Selected Disinfectants and By-Products 89 Sub et al. (1983) examined the effects of chlorine dioxide and its metabolites on fetal development in the rat. Female rats were administered chlorine dioxide at 0, 1, 10, or 100 mg/liter of drinking water for 2.5 months prior to and throughout gestation. The small number of dams bred (six to nine per test group), an unusually high percentage of abnormal control fetuses (Gino), and uncertainty as to the unit of statistical com- parison (fetus or litter) preclude statistical significance of observed skeletal variations and decreased number of implants per dam. A statistically significant increase in fetal body weight at the 100-mg/liter dose level may be a consequence of the reduced litter size observed at this dose. Sub et al. (1983) gave groups of six to eight female Sprague-Dawley rats drinking water containing chlorine dioxide concentrations of 0, 1, 10, or 100 mg/liter (O. 0.1, 1, or 10 mg/kg bw per day) for 2.5 months, before they were bred with untreated males. Exposure was continued throughout gestation. At the highest dose, there were reductions in mean number of implants and mean number of live fetuses per dam. The small number of litters and the unusually high percentage of abnormal fetuses among the control litters (approximately 31%) preclude statistical significance of skeletal variations observed and invite questions as to the validity of the study. Other Effects Revis et al. (1986) investigated the effects of drinking water containing chlorine dioxide at 2 or 15 ppm on thyroid function and on plasma cholesterol in rabbits and pigeons. In pigeons supplied drinking water containing 15 ppm chlorine dioxide for 3 months, concentrations of T4 in the plasma were reported to be significantly reduced whether they were on a normal or a high-cholesterol diet, as compared with those of controls. In most of the groups, T4 levels were reported to be significantly lower after imbibing water containing a 2-ppm concentration of chlorine dioxide. Increases in plasma cholesterol were seen frequently in groups with the lower T4 levels, especially in those given the high-cholesterol diet and the water at 15 ppm chlorine dioxide. Revis et al. suggest that these effects are mediated by-products formed by the reaction of chlorine dioxide, hypochlorite, and monochloramine with organic matter in the upper gastrointestinal tract. The significance of this study for humans is unknown because little information exists on pigeon thyroid function. Further, the statistical and hormone measurement methods used in this study appear inappropriate to the experimental design. CONCLUSIONS AND RECOMMENDATIONS Chlorine dioxide produces hematological effects in both humans and laboratory animals. The mechanism of these effects is not known; how- ever, it is believed to be related to the oxidant properties of chlorine

90 DRINKING WATER AND HEALTH dioxide and its aqueous reaction products, chlorite and chlorate. In ad- dition, thyroid and developmental neurological effects have been observed in laboratory animals. The thyroid effects of chlorite are thought to be caused by its oxidation of dietary iodide in the gastrointestinal tract. The oxidized iodide then binds to either food or tissue and is unavailable for absorption. The mechanisms of the neurological anomalies are unknown. Orme et al. (1985) found that levels of 14 mg/kg bw per day produced drops in T4 levels and abnormal neurological development in rat pups born to dams exposed during gestation and lactation. The investigators were able to show a no-observed-effect level (NOEL) of 3 mg/kg bw per day. In addition, Bercz et al. (1982) were able to show a NOEL of 3.5 mg/kg bw per day for thyroid effects in monkeys, supporting the more recent results by Orme et al. (19851. The committee selected the NOEL of 3.0 mg/kg bw per day and an uncertainty factor of 100 to estimate a chronic suggested no-adverse-effect level (SNARL) assuming that a 70-kg human consumes 2 liters of water daily, which contributes 20% of total intake: 3 mg/kg low/day x 70 kg x 0.2 0.21 mg/liter, or = 100 x 2 liters 210 ~g/liter. A SNARL may also be estimated for a 10-kg child consuming 1 liter of water daily, which contributes 20% of total intake: 3 mg/kg low/day x 10 kg x 0.2 100 x 1 liter CHLORAMINES Monochioramine CAS No. 10599-90-3 NH2C _ 0.06 mg/liter, or 60 ~g/liter. Pure monochloramine is a colorless, unstable, and pungent liquid with a freezing point of-66°C. It decomposes above-50°C and forms ni- trogen, chlorine, and nitrogen bichloride (Cotton and Jones, 1955; Ko- vacic et al., 19701. Monochloramine is used as an intermediate in the Raschig process for the industrial production of hydrazine. However, aqueous solutions of monochloramine formed by the chlorination of natural waters containing ammonia hold the primary environmental significance of the compound.

Selected Disinfectants and By-Products 91 Some confusion exists over the use of the term chloramine because it implies simply that a compound, organic or inorganic, contains both a chlorine atom and an amino-nitrogen functional group. This would then include the highly carcinogenic nitrogen mustards, which are not formed when water containing ammonia or organic amines is chlorinated. Fur- thermore, monochloramine should be recognized as different from the commercial products known as chloramine B. chloramine T. and dichlo- ramine T. which are organic compounds made by chlorinating benzene- sulfonamide or para-toluenesulfonamide. The reactions of chloramine with organic materials have been exten- sively reviewed (Kovacic et al., 19701. In general, monochloramine is a less potent oxidant than chlorine, with a standard oxidation potential of -1.16 V compared with-1.49 V for chlorine (Rosenblatt, 19751. Chlor- amine is able to transfer either the chlorine or the nitrogen onto reactive organic substrates. For example, chloramine reacts with phenols under certain conditions in dilute aqueous solution to form chlorophenols (Burtt- schell et al., 1959; Carlson and Lin, 1985), or it can form phenolamine coupling reaction products (first observed by Berthelot in 18591. In ad- dition, C. Le Cloirec and Martin (1985) have shown that chloramine can react with aldehydes to form nitrites. However, it has not been shown that these products are formed in chloraminated natural waters. Even though monochloramine can substitute a chlorine atom into organic com- pounds present in water, it does so to a much lesser extent than chlorine. It is for this reason that it has been recommended as an alternative dis- infectant to limit the formation of trihalomethanes in disinfected water (EPA, 19831. METABOLISM Few studies have been reported on the fate of monochloramine following ingestion. Stomach fluid contains high concentrations of proteins and amino acids, and monochloramine reacts with some amino acids (Jacan- gelo and Olivieri, 19851. Sulfur-containing amino acids and proteins are readily oxidized by monochloramine, and their presence may cause re- duction of the monochloramine to innocuous products. Under conditions that the reducing mechanisms in stomach fluid may overcome (Scully et al., 1986), the transfer of a chlorine atom from monochloramine to organic amines or amino acids to form organic chloramines may be important (Isaac and Morris, 1980, 1983, 1985; Snyder and Margerum, 19821. In the absence of reducing mechanisms, the concentrations of organic amino- nitrogen compounds in stomach fluid (Scully et al., 1985, 1986) are sufficient to allow half of the monochloramine in a 2-mg/liter (as C12) solution to transfer its chlorine atoms to amino acids in less than 1.5

92 DRINKING WATER AND HEALTH minutes at pH 6 (concentrations from Scully et al., 1985, 1986; rate data from Isaac and Morris, 19851. In this case the toxicological effects of monochloramine may be related to the organic chloramines formed by this chlorine transfer reaction. Absorption of an orally administered or- ganic chloramine (N-chloropiperidine) into the blood of Sprague-Dawley rats has been shown (Scully et al., 1985, 1986), but the full toxicological implications of these results are still to be determined. Under conditions designed to simulate the gastrointestinal tract, Bercz and Bawa (1986) found that monochloramine caused covalent binding of radioiodide to nutrient biochemicals. Monochloramine was believed to oxidize iodide to iodine, which subsequently reacted with nutrient chem- icals to form iodinated organic compounds. Tyrosine, 4-aminobenzoic acid, arachidonic acid, and folio acid were among the compounds that became iodinated under the conditions of the experiment. Some of the reactions were carried out under conditions of pH that were somewhat basic. The observed percent binding was generally higher for reactions carried out at higher pH levels. Although complex mixtures of nutrients, such as gastric juice and saliva, appeared to bind iodine in dilute aqueous solution, it is important that these results be correctly extrapolated to physiological pH before their significance is fully understood. Grisham et al. (1984) published evidence that through the oxidation of chloride by the myeloperoxidase/hydrogen peroxide/chloride system, hu- man neutrophilic leukocytes may produce organic chloramines. The pro- posed function of the organic chloramines is to provide a reserve of oxidizing equivalents for killing bacteria. The organic chloramines are then believed to react with ammonium ions to form the more bactericidal monochloramine. However, because no transfer of chlorine from organic chloramines to ammonia has ever been demonstrated, the validity of this mechanism has not yet been proven. Abdel-Rahman et al. (1983) administered doses of 3 ml of freshly synthesized aqueous chloramineL36Cl] (370 mg/liter) orally to four male Sprague-Dawley rats that had been fasted overnight. Suh and Abdel- Rahman (1983) reported a study of the control compound chloride-36. The rate of absorption of the radiolabel into blood was considerably faster in the case of chloramineL36Cll (T,/2 = 2.5 hours) than it was in the case of chloride-36 (T,/2 = 19.2 hours). The plasma maximum was reached in ~ hours for both compounds, followed in each case by a plateau in the plasma concentration. Following this plateau, the rate of elimination of the label from plasma was similar in both studies (T,/2 = 38.8 hours for chloramineL36Cll and T,/2 = 51.9 hours for 36Cl-~. In their study of chloramineL36Cll, Abdel-Rahman et al. ~ 1 983) collected the urine, feces, and expired air over 4- and 5-day periods from four male Sprague-Dawley rats. During the first 24 hours after administration of chloramineL36Cll, only 0.40% and 0.08% of the total dose administered

Selected Disinfectants and By-Products 93 were eliminated in the urine and feces, respectively. The proportion elim- inated through the urine and feces at the end of the 120-hour study period was 25.15% and 1.98%, respectively. By comparison, Sub and Abdel- Rahman (1983) found that over twice as much of the 36Cl-label was eliminated over the 120-hour study period when 36Cl-labeled chloride was administered. They found that 57.2% of the administered chloride-36 was eliminated in urine and 3.0% was eliminated in feces. Unlike chlor- amineL36Cll, a considerable amount of the chloride-36 was eliminated in the first 24 hours: 16.1% in urine and 0.92% in feces. After 48 hours, radioactivity was eliminated with a half-time similar to that found after 24 hours in the chloramineL36Cl] study (To = 24 hours). Only 27.1% of the administered amount of the label was excreted in the chloramine study over 5 days, whereas 60.2% of the radiolabel was excreted in the chloride- 36 study. Abdel-Rahman et al. (1983) reported that the principal (88%) excreted metabolite of chloramineL36Cll is chloride-36. However, the iden- tity of the 73% of the label retained by the body is unknown. The subcellular distribution of 36Cl activity in rat liver preparations was similar at 24 hours following oral administration of either chloramineL36Cl] or chloride-36 (Abdel-Rahman et al., 1983; Sub and Abdel-Rahman, 1983~. Abdel-Rahman et al. (1983) analyzed plasma for metabolites at 120 hours after administration of chloramineL36Cl] to rats. Neither 36Cl-labeled chlorite nor chlorate was detected in rat plasma. Most of the total 36Cl was identified as chloride-36, which, according to the authors, indicated that the chlorine moiety was eliminated primarily in this form. The control reagent for a pharmacokinetic study of chloramineL36Cll is 36Cl-. It is not likely that an oxidant as strong as NH2C1 will survive intact absorption, distribution, and excretion from an animal. If the chlor- amine is rapidly detoxified in the stomach to 36Cl-labeled chloride and ammonia, the observed pharmacokinetics will be identical to the control kinetics. There are aspects of the pharmacokinetics of both chlor- amineL36Cll and chloride-36 that are similar, but there are also notable differences. The fate of much of the 36C1 label is unknown. If the chlor- amine acts as a chlorinating agent, other, more stable chlorinated organic compounds may form. The observed kinetics of the radioactivity seen after administration of chloramineL36Cll appear to be quite complex, pos- sibly owing to a combination of chloride-36, 36Cl-chlorinated compounds, and their metabolites. HEALTH EFFECTS Observations in Humans Acute Effects One case has been reported of a woman who was over- come by inhalation of gaseous monochloramine in a poorly ventilated

94 DRINKING WATER AND HEALTH bathroom (Laakso et al., 19821. Monochloramine and dichloramine were formed when household ammonia was mixed with chlorine bleach (5% sodium hypochlorite). The vapors caused burning of the eyes and throat, dyspnea, coughing, and vomiting. The resultant pneumonitis did not lead to permanent pulmonary damage. In another reported instance, treatment of a dental abscess with a 2% solution of monochloramine resulted in the development of a type I allergic reaction (Beck, 19834. Subchronic Elects Lubbers et al. (1981) administered drinking water containing chlorine, chloramine, chlorine dioxide, sodium chlorite, or sodium chlorate to adult male human subjects. Sixty volunteer subjects were randomly assigned to six treatment groups of 10 subjects each; the members of one group (the control group) received untreated water, while the members of the other groups each received one of the disinfectants or disinfectant reaction products. In phase I, the acute effects of increasing doses were investigated; in phase II, the effects of ingestion of the dis- infectants or reaction products at a concentration of 5 mg/liter of water for 12 consecutive weeks were investigated. A third phase of the study did not include monochloramine. In the first phase, each subject ingested 1 liter of the untreated water in two half-portions; the second SOO-ml portion was given 4 hours after the first. Each portion was consumed within 15 minutes. Following this day of disinfectant administration, there were 2 days free of such admin- istration, during which blood and urine were collected and physical ex- aminations were conducted. Five such consecutive 3-day segments constituted this phase of the monochloramine study, and doses were in- creased from 0.01 to 24.0 mg/liter. Assays of a number of serum chemical components and of blood cells were performed, as well as analyses of urine and special tests, such as G-6-PD, thyroid hormones, and electro- cardiograms (ECGs). Physical examinations included observations of sys- tolic and diastolic blood pressures, respiratory and pulse rates, and oral temperature. Using analysis of variance techniques, group main effects (G), time main effects (R), and group-time interaction (RG) were esti- mated. In all instances, group mean values were and remained within normal ranges. Any trends identified by the analysis of variance were concluded not to be of clinical importance. In phase II, the group of 10 subjects administered monochloramine was divided into three subsets; each subset entered the study sequentially to facilitate management of the experiment. Each subject consumed 500 ml of water containing monochloramine in concentrations of 5 mg/liter daily for 12 weeks. Physical examinations, collection of blood and urine, and taste evaluations were performed weekly during the treatment period and for 8 weeks afterward. Compared with the common control group, sig

Selected Disinfectants and By-Products 95 nificant RG values (p < 0.05) were found in group mean corpuscular hemoglobin in the case of chlorine. No linear trends were detected by linear regression analysis for monochloramine. The findings with regard to other disinfectants and their by-products are discussed in their respective sections. Several studies of the effect of monochloramine on blood components and the risk to hemodialysis patients have been reported (Eaton et al., 1973; Kjellstrand et al., 19741. When tap water containing chloramines was used for dialysis baths, two major effects were observed: oxidation of hemoglobin to methemoglobin and denaturation of hemoglobin. The amount of oxidative damage was proportional to the amount of mono- chloramine formed. Furthermore, exposure of red blood cells to mono- chloramine also inhibited the hexose monophosphate shunt that protects red cells from oxidative damage. Kjellstrand et al. suggested that chlor- amine-induced hemolysis might be reduced by the addition of ascorbic acid to the treatment water. Observations in Other Species Acute Effects Abdel-Rahman et al. (1984) investigated the toxicity of monochloramine in male Sprague-Dawley rats. Acute exposure to a single dose (3 ml) at 10, 20, or 40 mg/liter induced a significant increase in blood glutathione levels within 30 minutes of administration of an aqueous solution by gavage. Subchronic Elects Bercz et al. (1982) studied the subchronic toxicity of monochloramine administered to African Green monkeys in drinking water following 30- to 60-day subchronic, exponentially rising step doses. At 100 mg/liter, monochloramine had no detectable effect in 18 hema- tological tests on the 12 monkeys, including red-cell glutathione (GSH) levels. No evidence of thyroid suppression was detected in serum. In a draft of a report of the Gulf South Research Institute for the National Toxicology Program, peer-reviewed by the committee (GSRI, 1981), Fisher 344 rats and the B6C3F~ mice were given monochloramine in drinking water at concentrations of 0, 25, 50, 100, 200, and 400 mg/liter for 90 days. The investigators reported decreased body weight gain and liver damage including increased mitotic figures, cellular hypertrophy, and unusual chromatin patterns in the liver cells of mice exposed to chlor- amines in concentrations of 100, 200, and 400 mg/liter. They also observed decreased body weight gain and decreased relative liver weight in male and female rats and increased protein excretion in male rats given 200 and 400 mg/liter. The investigators stated that 100 mg/liter "appeared to

96 OR ~ N K! NG WATER AN D H EALTH be the threshold level for lower toxicity in mice." So a no-observed-effect level would be 50 mg/liter, approximately 8.3 mg/kg bw per day. Chronic Effects Moore and Calabrese (1980) reviewed the health ef- fects studies previously conducted on monochloramine. Few drinking water studies had been reported. Maziarka et al. (1976) had found no observable effects in rats exposed for 12 months to chloramine at concentrations of 9.0 mg/liter. Moore et al. (1980) exposed male A/J mice for 30 days to monochloramine in bicarbonate-buffered solution (pH 8.9) at concentra- tions of 2.5 to 200 mg/liter. They found no significant changes in nine measured blood parameters. Following chronic treatment at 1, 10, or 100 mg/liter, they found no significant changes in measured blood parameters. Chronic treatment at 1, 10, or 100 mg/liter doses in drinking water induced a significant decrease in glutathione levels after 4 months of treatment at the 1-mg/liter and 100-mg/liter dose levels (Abdel-Rahman et al., 19841. Results varied over the 12-month study period, but at 6 and 12 months after initiation of the study, statistically lower glutathione levels were observed at all dosages. After 3 months of treatment, significant decreases in red-blood-cell count and hematocrit were observed at the higher dosage levels. However, there was an apparent lack of dose-re- sponse or time-dependent response in both glutathione levels and hema- tological parameters. Bull (1980) reported the results of a 45-day study in which mono- chloramine in drinking water was administered to laboratory rats. Body weight gain and hematological parameters in exposed animals did not differ significantly from those observed in control animals. The only significant finding was a decrease in the amount of methemoglobin present in the blood the opposite result of what was expected. Mutagenicity In a study by Shih and Lederberg (1976), chloramine reacted with Bacillus subtilis deoxyribonucleic acid (DNA) in vivo and in vitro; it was shown to be weakly mutagenic to a strain of B. subtilis, causing reversion of trpC to trp+. Shih and Lederberg studied the bio- logical and physical effects of chloramine on B. subtilis after treating the bacterial cells (in vivo) and the bactenal DNA (in vitro). Both kinds of treatment resulted in single-strand breaks and a few double-strand scissions (at higher chloramine doses) with loss of DNA-transforming activity. Chloramine used as a bactericide may target DNA since some DNA-repair mutants seem to be more sensitive to chloramine. Fetner (1962) found that distilled water containing monochloramine produced chromosome breakage when it was used for soaking Vicia faba seeds. A 1-hour exposure to 10-4 M monochloramine produced 24%

Selected Disinfectants and By-Products 97 abnormal anaphases. Monochloramine produced chromosome breakage at concentrations that exhibited little evidence of tissue damage. On the other hand, Cheh et al. ~ 1980b) found that the organic concentrate from drinking water treated with chloramine produced only half as many revertants in the Ames Salmonella mutagenicity assay as the same water treated with chlorine. The addition of sulfite to reduce chemically the oxidants in chloramine-treated drinking water sharply decreased the mu- tagenic response (Cheh et al., 1980a,b; Wilcox and Denny, 19851. Carcinogenicity In two reports (Bull, 1980; Bull et al., 1982), settled, coagulated, and sand-filtered Ohio River water was treated with mono- chloramine (3 mg/liter). The residual disinfectant was dissipated within 48 hours. The water was then concentrated by reverse osmosis, and the concentrate was subjected to a mouse skin initiation-promotion assay in Sencar mice. Lesions macroscopically observed at autopsy included papil- lomas, squamous carcinomas, and lung adenomas. In these studies, mono- chloramine was not the primary carcinogen but was believed to be reacting with trace organics in the water and producing carcinogenic by-products. Bull observed variations in the quality of the source water (which produced tumors in one case prior to disinfection) and cautioned against extrapo- lation of these results until further examples have been examined. Herren-Freund and Pereira (1986) used a bioassay that involved an increased incidence of ~y-glutamyl transpeptidase (GOT) foci as an indi- cator of carcinogenicity. This bioassay consisted of administration of a candidate initiator to rats 18 to 24 hours after removal of two-thirds of their livers. Seven days after this initiation, the rats were given a candidate promoter in drinking water for at least 10 weeks. The assay detected such initiators as 2-acetylaminofluorene, aflatoxin Be, diethylnitrosamine, di- methylhydrazine, and urethane. In this assay, chloramine was not found to act as an initiator when a dose of 14.75 mg/kg bw was administered by gavage 1 day after partial hepatectomy and promotion with 500 ppm phenobarbital in the drinking water was begun 7 days after the dose of chloramine and continued for 10 weeks. The rats were killed 1 week after the end of promotion. Diethylnitrosamine was used as a positive control for initiation at 0.3 mole/kg. Teratogenicity Abdel-Rahman et al. (1982) investigated the effects of monochloramine administered in drinking water to mature virgin female Sprague-Dawley rats. Six animals per group were administered mono- chloramine daily in concentrations of 0, 1, 10, or 100 mg/liter of drinking water, both 2.5 months prior to and throughout gestation. Sacrifice of the rats on the twentieth day of gestation was performed for soft-tissue and skeletal examination. Monochloramine did not produce any significant

98 OR ~ N K! NG WATER AN D H EALTH changes in rat fetuses at any dose level; in fact, there was a slight increase in fetal weight in all monochloramine groups compared with controls. Meter et al. (1985) evaluated the ability of 40-, 100-, and 200-mg/liter solutions of monochloramine and other oxidants to induce chromosomal aberrations and micronuclei in the bone marrow of CD- 1 mice and sperm- head abnormalities in B6C3F~ mice. Monochloramine showed no evidence of any significant effects in any of the tests. Carlton et al. (1986) administered chloramine by intragastric catheter at doses of 0, 2.5, 5.0, or 10 mg/kg bw per day to male and female Long- Evans rats that were 4 to 6 weeks old at the beginning of the experiment. Males were treated for 56 days and females for 14 days prior to mating; the administration was continued during the 10-day mating period, and thereafter females were given chloramine daily through gestation and lactation. Males were necropsied at the end of the mating period. Their sperm were examined for normalcy, and microscopic changes in the anat- omy of the reproductive tract were sought. Dams and some offspring were necropsied at weaning, 21 days after birth. Other offspring were admin- istered chloramine after weaning until they were 28 or 45 days old; these pups were evaluated for vaginal patency and thyroid hormone levels. No differences between control and exposed rats were found in fertility, viability, litter size, day of eye opening, or day of vaginal patency. There were no alterations in sperm count, direct progressive sperm movement, percent mobility, or sperm morphological characteristics in adult males. Weights of male and female reproductive organs were not significantly different among test and control groups, and there were no significant morbid anatomic changes evident on tissue examination. There were no signs of toxicity, changes in blood counts, or body weight suppression in adult rats of either sex at any dose level. The mean weight of the pups was unchanged from that for control litters. Other Effects Revis et al. (1986) studied the effects of monochlora- mine on thyroid function and on plasma cholesterol in rabbits and pigeons. In pigeons supplied drinking water containing monochloramine at 2 ppm for 3 months, concentrations of plasma thyroxine (T4) and plasma cho- lesterol were increased, although a clear dose-response effect for plasma cholesterol was not observed. The sample size in this study was too small with proportionately too few controls to produce statistically significant results. The marked increase in serum T4 should be further evaluated, as well as the relevance of the pigeon as an animal model. CONCLUSIONS AND RECOMMENDATIONS Monochloramine can produce hepatocellular changes in laboratory an- imals. A subchronic bioassay (GSRI, 1981) showed decreased body weight

Selected Disinfectants and By-Products 99 gain and liver toxicity in exposed mice. A no-observed-effect level (NOEL) for liver toxicity from the GSRI study was 8.3 mg/kg bw per day. Using this as a basis, and assuming that a 70-kg human consumes 2 liters of water daily, which contributes 20% of total intake, the committee estimates a suggested no-adverse-effect level (SNARL) as: 8.3 mg/kg low/day x 70 kg x 0.2 0.581 mg/liter, or 100 x 2 liters = 581 ~g/liter. A SNARL may also be estimated for a 10-kg child consuming 1 liter of water daily, which contributes 20% of total intake: 8.3 mg/kg low/day x 10 kg x 0.2 100 x 1 liter CH LORITE Sodium chlorite CAS No. 7758-19-2 CH LORATE Sodium chlorate, AtIacide CAS No. 7775-09-9 NaClO2 NaClO3 0.166 mg/liter, or 166 ~g/liter. Sodium chlorite can be crystalline, flake, or powder in form. It is slightly hydroscopic but does not cake. In crystalline form its density is 2.468 g/cm3. At 17°C its solubility is 39 g/100 ml of water and at 60°C, 55 g/ 100 ml. The melting point of sodium chlorite is between 1 80°C and 200°C. Chlorite is used for the on-site production of chlorine dioxide, in water purification, in paper pulp, and as a bleaching agent for textiles. It is also used in shellacs, varnishes, waxes, and straw products. Sodium chlorate is colorless, odorless, and crystalline in form with a cooling saline taste. It is soluble in cold water (79 g/100 ml water at 0°C) and in 0.5 ml of boiling water (230 g/100 ml water at 100°C); its melting point is 248°C, and its boiling point is 122°C with a density of 2.490. Sodium chlorate is used as an oxidizing agent and as a bleach (especially to make chlorine dioxide). It is also used in paper pulps, matches, ex

TOO DRINKING WATER AND HEALTH plosives, flares and pyrotechnics, ore processing, herbicides and defol- iants, medicine, and as a substitute for potassium chlorate. Both sodium chlorite and sodium chlorate were reviewed in Volumes 3 and 4 of Drinking Water and Health (NRC, 1980, pp. 193-202; 1982, pp. 174-1761. The following information updates and reevaluates what is known about these compounds. METABOLISM As a part of an investigation of the kinetics of C1O2, Abdel-Rahman et al. (1980) investigated the effects of chlorite and chlorate on blood components. Male rats weighing 150 to 170 g and white leghorn roosters weighing 250 to 300 g were given drinking water that contained chlorite or chlorate in concentrations of 10 or 100 mg/liter for 20 hours/day, 7 days/week for 4 months. In all groups of rats and chickens intoxicated by chlorite or chlorate there were decreases in blood glutathione, in osmotic fragility of erythrocytes, and in the morphology of erythrocytes from rats and chickens. No methemoglobin was detected. Abdel-Rahman and coworkers (1984a) administered 36Cl-labeled so- dium chlorite and sodium chlorate to rats in a study of their absorption, distribution, and excretion Groups of four male rats drank 3 ml of so- lutions of chlorite (10 mg/liter) or chlorate (5 mg/liter); some groups were kept for periodic drawing of blood before being killed at 72 hours for analysis of various organs and tissues, while others were kept in metab- olism chambers for collection of excreted air, urine, and feces. Peak plasma levels of 36Cl were reached 2 hours after administration of chlorite; the half-life for elimination of the labeled C1 was 35 hours. Peak plasma levels of labeled C1 from chlorate were reached in 30 minutes, with a half-life for rapid elimination of about 6 hours, followed by a slower phase with a half-life of 36.7 hours. At 72 hours after administration of the chlorite, 36Cl was highest in whole blood, followed by packed cells, plasma, stomach, testes, skin, lungs, kidney, duodenum, carcass, spleen, ileum, brain, bone marrow, and liver. But with chlorate, 36C1 was highest in plasma, followed by whole blood, stomach, testes, lungs, kidney, skin, duodenum, spleen, brain, packed cells, ileum, carcass, liver, and bone marrow. In the 72-hour period following administration of chlorite or chlorate, during which the rats were kept in metabolism chambers for collection of urine, feces, and expired air, 39% of the 36C1 from administered chlorite was recovered, about 35% in the urine and about 5% in the feces. No labeled C1 was detected in expired air. In the case of chlorate, about 43% of the labeled C1 was collected in the 72-hour period, about 39% in the

Selected Disinfectants and By-Products 101 first 24 hours. As in the case of chlorite, no labeled C1 was found in . . . expired air. Labeled C1 excreted after chlorite administration was in the form of chlorite or chloride; labeled Cl excreted after chlorate administration was in the form of chlorate, chlorite, or chloride. The investigators pointed to the prolonged retention of chlorite or chlor- ate metabolites in the testes as evidence of possible action at this site. Abdel-Rahman and coworkers (1980, 1982) also studied the tissue dis- tribution of 36Cl-labeled chlorite and chlorate. The amounts found in various fluids and tissues, expressed as percentage of the initial dose after 72 hours, were as follows: for chlorite, 0.55% was found in the plasma, 0.63% in packed cells, 0.64% in whole blood, and a total of about 3% in kidneys, lungs, stomach, duodenum, ileum, liver, spleen, bone marrow, testes, skin, and carcass, with the highest concentrations (about 0.4%) being found in skin, testes, stomach, and lungs; for chlorate, 0.68% was found in plasma, 0.23% in packed cells, 0.57% in whole blood, with a total of 3.6% in the same tissues listed for chlorite, with about 0.4% in each of several tissues, namely, kidney, lungs, stomach, testes, and skin. Chlorite was administered as a 10-mg/liter solution and chlorate as a 5- mg/liter solution. According to Abdel-Rahman (1985), the amounts of labeled chlorine remaining in various tissues after 72 hours can be largely accounted for as chloride. Together with the other data, this suggests that neither chlorite nor chlorate bioaccumulates. Unlike chlorite the excretion of 36Cl from chlorate was biphasic. The first phase of decay from the plasma of rats had a half-life of 6 hours, and a second phase had a half-life of 36.7 hours (Abdel-Rahman et al., 1982, 1984a). Most of the intact chlorate excreted in the urine appeared within the first ~ hours after administration; thereafter, no intact chlorate was detected. Bercz et al. (1982) observed that one effect of ClO2 in African Green monkeys was a decrease in serum thyroxine. On the chance that this might reflect an effect of chlorite or chlorate resulting from C1O2 transformation, they administered chlorite or chlorate to these monkeys at concentrations equivalent to doses of up to 60 mg/kg bw per day without developing a depression of serum thyroxine. HEALTH ASPECTS Observations in Humans Because of its use as a weed killer, there are many case reports of chlorate intoxication. Most of these were available at the time of the

i02 DRINKING WATER AND HEALTH publication of Volume 3 of Drinking Water and Health (NRC, 19801. Several recent ones add to that information. Stavrou et al. (1978) described the effects in a 13-year-old boy who ingested an unknown amount of NaClO3 by licking the crystals adhering to his moistened finger placed in a bag of crystals. He became very ill and was seen by the family physician the next day; on the third day he was admitted to the hospital. He was cyanotic, passing little urine, and febrile. His liver was enlarged, he was jaundiced, and he felt tenderness in the epigastrium and loins. His blood was brown and Heinz bodies were seen. Methemoglobin was found in the plasma, and protein was detected in the urine. He had renal failure, managed by peritoneal dialysis, for 21 days. Bloxham et al. (1979) described a 29-year-old man who had ingested about 20 g NaClO3 (230 mg chlorate/kg bw). He became cyanotic, and his hemoglobin dropped to 11 g/100 ml within 24 hours; methemoglobin and methemoalbumin were detected in his plasma. He was anuric for 14 days, then gradually improved, and he was released from the hospital after 6 weeks. Helliwell and Nunn (1979) reported on 14 cases of NaClO3 poisoning. The patients' ages ranged from 3 to 55 years. Doses estimated to be in excess of 100 g or 79 g as chlorate ion were uniformly fatal. One 46- year-old woman given supportive therapy died 20 hours after a dose estimated to be 15 g (218 mg chlorate/kg bw). This was the lowest dose found to be fatal in these cases. Another female of unreported age died 5 days after ingesting 30 g (436 mg chlorate/kg bw), despite treatment with methylene blue, peritoneal dialysis, and exchange transfusion. How- ever, an 18-year-old male survived a dose estimated at 100 g (1.45 g chlorate/kg bw) after treatment with methylene blue, exchange transfusion, and hemodialysis. Cyanosis was seen in 50% of the patients, abdominal pain in 36%, diarrhea in 21%, dyspnea in 21%, anuria within 48 hours in 50%, coma in 12%, and methemoglobinemia in 93%. Sixty-four percent died. Steffen and Seitz (1981) saw a 26-year-old woman 5 hours after she ingested 150 to 200 g chlorate (2.145 g chlorate/kg bw). Methemoglobin was an early sign of intoxication; although the administration of methylene blue appeared to be helpful in treating the methemoglobinemia, it did not prevent massive hemolysis and disseminated intravascular coagulation. Renal function was said to be completely absent for 10 days, but dialysis was continued another month, and renal output then began to exceed 1,000 ml/day. Lubbers and associates (1981, 1982, 1983) conducted controlled clinical investigations of volunteer human subjects exposed to sodium chlorite and sodium chlorate (as well as to some other materials). Statistical analyses

Selected Disinfectants and By-Products 103 of the data were reported by Lubbers and Bianchine (1984) and Lubbers et al. (19841. The subjects were given repeated physical examinations and extensive laboratory tests. Described in more detail in the section on chloramine, Lubbers et al. (1981) gave groups of 10 male volunteer subjects sodium chlorite in drinking water in two separate phases. In the first phase, the 10 subjects drank 1 liter of water every fourth day for 6 days, over a 16-day period. The first day, the concentration of sodium chlorite in water was 0.01 ma/ liter, and this was increased each exposure day to a final concentration of 2.4 mg/liter (average 0.34 mg/kg bw per day). Subjects were given the test substance once every 3 days. Group mean values determined for 2 days following each exposure of all investigated effects, including many serum chemistry components and blood-cell counts as well as results from EGG analysis and physical examinations, were within normal ranges, and no trends or interactions judged to be of clinical significance were found. In a second phase, 10 volunteers drank 500 ml of water containing 5 mg sodium chlorite/liter daily for 12 weeks (average 0.034 mg/kg bw per day). Significant (p < 0.05) group-time interaction was found in the case of group mean corpuscular hemoglobin; however, linear trend regression analysis did not show a significant linear trend, and the physiological significance of this interaction was doubted. In a third phase of the study on sodium chlorite, three subjects found to be deficient in G-6-PD were given 500 ml of drinking water containing 5 mg chlorite/liter of water every day for 12 weeks, as in the second phase. The rationale for this phase was that those deficient in G-6-PD might be expected to be especially susceptible to oxidative stress. The small number of subjects made some of the otherwise desirable statistical procedures of questionable use, and linear regression analysis was chosen. There were several laboratory analyses with a significant probability (p < 0.05) of a change with respect to time over the 12-week treatment period; these were in albumin/globulin ratio, T4 radioimmunoassay (RIA), free thyroxine, mean corpuscular hemoglobin concentration, and methe- moglobin values. The authors cautioned that these were only trends, that caution should be used in interpreting their significance because of the small number of subjects and the possibility of "laboratory drift," and that attribution of physiological consequence would be premature. Lubbers and coworkers (1981) also administered chlorate in drinking water to groups of male volunteer subjects in the same experiments, described more fully in the section on chloramine. In phase I, 10 subjects drank 1 liter of chlorate-containing water for 6 days, with 2 days inter- vening between each exposure day, so that the six exposures occurred over a span of 16 days. The concentration of sodium chlorate rose gradually during the experiment; it was 0.01 mg/liter the first day and rose to 2.4

]04 DRINKING WATER AND HEALTH mg/liter by the sixth day (average 0.34 mg/kg bw per day). There were changes concluded to be statistically significant but not clinically signif- icant in serum bilirubin (total) and iron. In a second experiment, 10 subjects drank 500 ml of water containing sodium chlorate at 5 mg/liter every day for 12 weeks (average 0.034 ma/ kg bw per day). There was a significant (p < 0.05) group-time interaction in the case of group mean corpuscular hemoglobin and of group mean blood urea nitrogen, but linear regression analysis did not show a signif- icant linear trend in the means of these effects, and they were judged not to be clinically significant. Observations in Other Species Acute Elects There were no studies available. Subchronic Effects Bercz et al. (1982) administered drinking water containing NaClO2 or NaClO3 as well as ClO2 or NH2C1 to African Green monkeys for 30-60 days in a study of possible thyroid effects. The drinking water contained chlorite or chlorate in concentrations of 25, 50, 100, 200, or 400 mg/liter in a rising-dose experiment. Equivalent doses were 4, 7.5, 15, 30, or 58.4 mg/kg bw per day. Neither the chlorite nor the chlorate induced any thyroid depression at any dose. The chlorite but not the chlorate induced a dose-dependent oxidative stress on hematopoiesis, re- sulting in a decreased hemoglobin and erythrocyte count and an increased me/hemoglobin. Serum glutamic pyruvate transaminase (SGPT) was in- creased in a statistically significant and dose-dependent manner; however, the investigators believed that the effect was not clinically important and pointed out that this elevation was not corroborated by elevations in other enzymes or in serum bilirubin. The blood changes in the chlorite-treated monkeys started to reverse before the end of the period of administration. Heffernan et al. (1979a) found that the hemoglobin in blood from both rats and human beings was oxidized to methemoglobin by the chlorite, the hemoglobin in rat blood being somewhat more sensitive to concen- trations of 10-3 to 10-2 M NaClO2 than that of human blood. Above a concentration of NaClO2 of about 4 x 10-2, the situation seemed to reverse, the hemoglobin of human blood being more sensitive to oxidation by chlorite than that of rat blood. Hemoglobin of rat blood was more sensitive to oxidation by NO2 than by chlorite. The concentration of glutathione in erythrocytes was reduced almost to zero (3.8% of original) by added NaClO2 to yield a concentration of 50 mM; added NaNO2 to yield the same concentration of added material reduced the concentration of glutathione in the erythrocytes to only 79.5% of the original value.

Selected Disinfectants and By-Products 105 Anaerobic conditions did not alter appreciably the reduction in glutathione concentration by chlorite. The finding that reduction of glutathione con- centration in erythrocytes preceded the production of measurable amounts of methemoglobin suggests that chlorite-induced oxidative injury of red cells involves targets other than hemoglobin. EM-visible distortion of the erthrocyte membrane appeared at a concentration of NaClO2 of 7.4 x 10-4 M and increased in number on single membranes and in populations of red blood cells as the concentration of NaClO2 in the suspension of erythrocytes was increased. At a concentration of NaClO2 of 1.5 x 10-2 M, the envelope of the red cell became completely permeable to both hemoglobin and me/hemoglobin. The activity of catalase in red blood cells, used as a measure of production of H2O2, decreased practically to zero as the concentration of NaClO2 in the erythrocyte suspension was increased from 10-4 to 10-3 M, indicating that the active center of the enzyme became almost completely occupied by endogenously produced H2O2. These various findings in vitro indicate that the chlorite ion belongs to the class of chemicals that induces production of H2O2 and that is likely to produce hemolytic anemia in vivo. In fact, Heffernan et al. (1979b) found that rats given drinking water containing NaClO2 at 100 mg/liter for 30 days had a mean concentration of glutathione in their red blood cells that was only 66. 1% of normal, that the concentration of Hgb in their blood was reduced to about 87% of normal, that their mean red-blood-cell (RBC) count was 96.9% of normal, and that their mean packed-cell volume was 93.5% of normal. Higher concentrations of NaClO2 in the drinking water, up to 500 mg/liter, pro- duced greater changes in these measures. Administration of these amounts of drinking water for longer periods of time, up to 90 days, generally yielded smaller terminal deviations from normal than the shorter period, indicating the marshaling of compensatory mechanisms. The single ex- ception to this generalization was the glutathione concentration within the red blood cell, which was insignificantly lower after 90 days of exposure to NaClO2 in drinking water than after 30 days. When single doses of NaClO2 were injected intraperitoneally into rats in amounts of 1, 10, 20, 30, and 50 mg/kg bw, there was a dose-related increase in the percentage of methemoglobin in total heme pigment and a less clearly dose-related decrease in total heme pigment. Oral doses of 20 and 64 mg NaClO2/kg bw given to cats resulted in fairly sharp (1.5 hours) increase in the per- centage of methemoglobin in total heme pigment to a mean of 22.8% in three cats given the smaller dose and of 46.7% in one cat given the larger dose. By 6 hours after the doses, these increases had decreased consid- erably from their maxima. Only the cat given the smaller dose that had had the least maximum increase in its percentage of methemoglobin (10%) had returned completely to normal within that time.

i06 DRINKING WATER AND HEALTH In a companion report, Heffernan et al. (1979a) described some of these effects further. They incubated sodium chlorite with blood from male rats and studied some of the reactions that occurred. They found chlorite to be slightly less potent than nitrite as an oxidant of hemoglobin and less specific in its oxidation of cellular constituents. As in the earlier paper, they found that chlorite depleted red-cell glutathione, accompanied by an increased generation of hydrogen peroxide. They described changes in the structure of erythrocyte membranes. They commented that chlorite acts primarily as a hemolytic agent and secondarily as an oxidant of hemoglobin. In an investigation of subchronic toxicity, Heffernan et al. (1979b) gave rats and cats NaClO2 in their drinking water. There was a dose-related decrease in RBC, hemoglobin, and packed-cell value at 30 and at 60 days in rats given water containing chlorite at 100 mg/liter, equivalent to 10 mg/kg bw per day, and higher. After 90 days, some adaptation had oc- curred. RBC glutathione concentrations were significantly decreased at chlorite concentrations as low as 50 mg/liter (5 mg/kg bw per day). Unlike the other blood effects studied, this effect on glutathione did not recover on longer exposure. Erythrocytes from rats earlier intoxicated with chlorite were less able than is normal to control generation of hydrogen peroxide when incubated in vitro with chlorite. The investigators attributed this to chlorite-induced depletion of glutathione in erythrocytes. In cats, there was a 20-30~o decrease in packed-cell volume and in hemoglobin concentrations. They were given drinking water containing NaClO2 at 500 mg/liter (7 mg/kg bw per day). Doubling the chlorite concentration greatly increased this effect. The use of 5~Cr-labeling of e~rocytes in cats given water containing O. 100, 250, or 500 mg NaClO2/ liter (0.6, 3.6, or 7 mg/kg bw per day) showed a dose-related increase in turnover of RBCs at levels of 100 mg/liter and above. (Water consumption at the highest dose was markedly decreased.) It was concluded that chlorite can induce hemolytic anemia. An increase in the ratio of kidney weight to body weight was observed in male CD rats given drinking water containing 500 mg NaClO2/liter (50 mg/kg bw per day) for 60 days (Heffernan et al., 1979b). This was not seen at lower water concentrations (100 and 150 mg/liter). Moore and Calabrese (1980) found an increase in mean corpuscular volume, in osmotic fragility and G-6-PD activity, and in the number of acanthocytes in A/J and C57L/J mice given drinking water containing 100 ppm (but not at 1 or 10 ppm) NaClO2, equivalent to 13-18 mg chlorite/ kg bw per day for 30 days. Moore and Calabrese concluded that the main effect of chlorite on erythrocytes is disruption of the cell membrane. Similarly, Heffernan et al. (1979a) had earlier described effects of chlorite on erythrocyte membranes of rats.

Selected Disinfectants and By-Products 107 Couri and Abdel-Rahman ( 1980) administered water containing chlorite at 100 mg/liter (10 mg/kg bw per day) to rats for 6-12 months and found marked decreases in RBC glutathione. There were smaller decreases at 10 mg/liter (1 mg/kg bw per day) at 6 but not at 12 months. This was more extensively studied by Abdel-Rahman et al. (1984b), but questions about the statistical treatment of the data need to be considered in light of the inconsistency of results. They reported a progressive decrease in RBC osmotic fragility as chlorite intoxication was extended beyond a few months. Hemoglobin, hematocrit, and RBC were decreased from control values after 9 months of exposure to chlorite at concentrations as low as 10 mg/liter. Chronic Expects Haag (1949) gave month-old rats sodium chlorite in their drinking water at concentrations of 1, 2, 4, 8, 100, and 1,000 ppm for 2 years. There were three groups of controls; two received distilled water and the third received water containing enough sodium chloride to be tonically equivalent to 1,000 ppm of sodium chlorite. There were seven female and seven male rats in each group except that no female rats were included in one of the drinking water controls. The alkalinity of the solution of 1,000 ppm sodium chlorite was adjusted by the addition of hydrochloric acid (HCl). At the end of the first year, all rats receiving 1, 2, or 4 ppm were switched to the 8-ppm solution, and they continued to receive sodium chlorite at that level throughout the second year. Rats of both sexes receiving sodium chlorite at 1,000 ppm consistently drank less water than did controls, while male rats receiving sodium chloride had an elevated fluid consumption. Consistent with the reduced ingestion of water, rats getting 1,000 ppm chlorite grew at a slower rate, whereas the male rats getting sodium chloride grew at a greater rate than did negative controls. There was no trend of increased mortality in any group. Microscopic studies of morbid anatomy were performed on represen- tative animals and on all rats with macroscopically evident tumors. The. only findings judged to be related to the treatment were in the kidneys of male rats that imbibed 100 or 1,000 ppm sodium chlorite and in male rats receiving sodium chloride in their drinking water. These changes consisted of a marked distention of the glomerular capsules by fluid as well as the filling of tubules with a pale, pink-staining material. The changes were most severe in the rats that imbibed sodium chloride. Haag (1949) wondered whether the affected rats might have developed a beginning nephritic condition at some point in the experiment that might have become exacerbated by the prolonged absorption of these salts. While he did not describe the source of his animals, it is doubtful that he had

i08 DRINKING WATER AND HEALTH specific-pathogen-free rats or other rats of excellent health in an experiment conducted about 40 years ago. Mutagenicity Eckhardt et al. (1982) reported in an abstract that NaClO3 was mutagenic to Salmonella typhimurium strain TA1535 in the presence of an S9 supernatant from an unspecified source. They also reported that NaClO3 was mutagenic in a Drosophila system but was inactive in the micronucleus assay; they did not give details. Meter et al. (1985) investigated the mutagenicity of NaClO2 and NaClO3 (as well as C1O2 and other disinfectants) in the mouse micronucleus assay, in the mouse sperm-head assay, and in the mouse bone-marrow chro- mosomal aberration assay. Swiss (CD-1) mice were used for the micro- nucleus and bone-marrow studies, and B6C3F~ mice were used for sperm- head studies. Doses, administered orally for 5 days, were 1 ml of a solution of 1 g/liter in water, or about 50 mg/kg bw if the mouse weighed 20 g. One group of animals was killed 6 hours after the last dose and examined for types and numbers of chromosomal aberrations. Another group, also killed at 6 hours, was examined for micronuclei in polychromatic eryth- rocytes. Sperm-head abnormalities were looked for in animals killed at 1, 3, or 5 weeks after the last dose, to ensure that all major stages of spermatogenesis would be included. Evidence of mutagenicity was not found in any of these tests with either salt. Carcinogenicity Kurokowa et al. (1984) tested chlorite as a possible tumor promoter or as a complete carcinogen. In the test of complete carcinogenicity, 0.2 ml of a 20-mg/ml solution in acetone was applied to the shaved backs of female Sencar mice twice weekly for 51 weeks; this is 200 mg/kg bw if the mouse weighed 20 g. No tumors were detected in the group of 20 animals. In the test of tumor promotion, dimethylbenz- anthracene (DMBA) was applied once prior to the 51-week application of NaClO2 as in the test of complete carcinogenicity. Six of the 20 mice receiving the prior dose of DMBA developed skin tumors, whereas those that received DMBA followed by acetone alone did not develop tumors. Five of the mice had squamous-cell carcinomas, but this did not meet the investigators' criterion of statistical significance (p = 0.01) on Fisher's exact test or the chi-squared test. The investigators' conclusion on sodium chlorite was that a "potential promoting effect was suspected...." Developmental E~ects Suh et al. (1983) gave groups of six to nine pregnant rats drinking water containing chlorite or chlorate in concentra- tions of 0, 1, or 10 mg/liter for 2~/2 months prior to and throughout gestation in a teratogenicity study of these disinfectants. Some alterations in number of fetuses resorbed or with skeletal or visceral anomalies were observed,

Selected Disinfectants and By-Products 109 but the small number of litters and a high rate of abnormalities among control fetuses precluded statistical significance or clear interpretation of the results. Sub et al. (1983) also gave groups of eight to nine pregnant rats drinking water containing chlorate at the same concentrations (1 or 10 mg/liter). The incidence of these same skeletal anomalies was 52% in the group receiving 1 mg chlorate/liter and 55% in the group receiving 10 mg/liter, compared with 31% in controls. Again, statistical significance was not demonstrated. An animal at 10 mg/liter was found to have hydronephrosis. Reproductive Elects Moore and Calabrese (1982) gave female A/J mice water containing 100 ppm NaClO2 as soon after mating as vaginal plugs, indicative of conception, were seen. There were no significant differences from controls, as judged by a l-test, in litter size, number alive at weaning, gestation time, number of stillbirths, number of pups dying between birth and weaning, average litter birth weight, and the weight and age of dams. However, there were significant decreases (p < 0.03) in the case of offspring of chlorite-treated dams, in body weights at wean- ing, and in the growth rate between birth and weaning. Couri et al. (1982) found fetal resorptions, sometimes involving all of a litter, in dams given high levels of NaClO2 (20,000 mg/liter, or 212 mg/kg bw per day). However, resorptions were not induced at 5,000 ma/ liter (122 mg/kg bw per day). The investigators suggested that these resorptions might be a consequence of anemia-induced hypoxemia. Sub et al. (1984) gave rats water containing chlorite or chlorate in concentrations of 100 mg/liter for 3 weeks. They found incorporation of 3H-thymidine (2-hour labeling period) into nuclei of testes was inhibited 50% in rats given chlorite at 100 mg/liter but was not decreased in rats given chlorate at that concentration. After exposure for 3 months, there was an inhibition of thymidine uptake (8-hour labeling period) at drinking water concentrations of 10 mg chlorate/liter (1 mg/kg bw per day) in another study (Abdel-Rahman et al., 1984b). Carlton and Smith (1985) described effects of chlorite on several aspects of reproduction. Long-Evans rats ingested water containing sodium chlor- ite at 0, 1, 10, or 100 ppm for 66-76 days. Males were treated for 56 days and females for 14 days before being allowed to mate, and the treatment was continued during the 10-day breeding period; other females were continued on the chlorite regimen throughout gestation and lactation. Following breeding, males were killed and evaluated for sperm changes and for morbid anatomic changes in the reproductive tract. Dams and some pups were necropsied at weaning time, 21 days after birth. No differences between control and test animals were seen with respect to litter size and viability or days of vaginal patency and of eye opening.

1lO DRINKING WATER AND HEALTH Body weights of adults were not affected. Additional male rats imbibed sodium chlorite in drinking water at 0, 100, or 500 ppm for 72-76 days; and in these rats given 500 ppm, a 28% decrease in water consumption occurred. Methemoglobin levels were not increased nor were other chorite-in- duced changes observed in blood elements. Reproductive tract anatomy was not altered. Organ weights and ratios of organ to body weights were not changed. A slight trend toward decreasing motility of sperm was seen in rats given 10 ppm and higher, but this was not statistically significant. There was a decrease in sperm drive range in rats given 10 ppm and higher, and this decrease was statistically significant (p < 0.01) in groups imbibing 100 or 500 ppm. Increases in abnormal sperm forms (p < 0.001) occurred in rats given 100 or 500 ppm. CONCLUSIONS AND RECOMMENDATIONS Both chlorate and chlorite produce damage in erthyrocytes and produce me/hemoglobin. It is thought that these effects are related to their oxidative properties. Hematological effects have been observed both in humans and in laboratory animals. As in the case of chlorine dioxide, decreased thyroid function has been associated with exposure to chlorate and chlorite. None- theless, this has not been confined by other studies. The decreased thyroid function may be related to oxidation of dietary iodide in the gastrointestinal tract, which then binds to food or tissue and is unavailable for absorption. Further research is necessary before this association can be accepted. A previous Safe Drinking Water Committee (NRC, 1980) has recom- mended a suggested no-adverse-effect level (SNARL) for chlorite of 0.21 mg/liter based on work by Heffernan et al. (1979b), where the no-ob- served-effect level (NOEL) for hematological effects in cats was 0.6 ma/ kg bw per day. There are no new studies in humans or animals that would show lower NOEL values, with the exception of the clinical studies by Lubbers and coworkers (1981), where a NOEL of 0.034 mg/kg bw per day was indicated for both chlorate and chlorite. An observed-effect level was not determined in these studies, therefore the relevance of a dose of 0.034 mg/kg bw per day to a threshold for hematological effects in humans cannot be determined. Nonetheless, the committee prefers to use human data when they are available. Thus a SNARL may be calculated for chlorate and chlorite using a NOEL of 0.034 mg/kg bw per day, and assuming that a 70-kg human consumes 2 liters of water daily, which contributes 20% of the total intake: 0.034 mg/kg/day x 70 kg x 0.2 _ 0.024 mg/liter, or 10 x 2 liters 24 ~g/liter.

Selected Disinfectants and By-Products 111 A SNARL may also be estimated for a 10-kg child consuming 1 liter of water daily, which contributes 20% of total intake: 0.034 mg/kg/day x 10 kg x 0.2 _ 0.007 mg/liter, or _ 7 ~g/liter. TR I HALOM ETHAN ES ChIoroform CAS No. 67-66-3 DibromochIoromethane CAS No. ~ 24-48- ~ 10 x 1 liter CHC13 CHClBr2 = Chloroform (trichloromethane) is a colorless, highly refractive, heavy, sweet-tasting liquid that has a light, crisp odor. Its vapor pressure is 100 mm of mercury at 104°C, and it is very slightly soluble in water (0.8 g/ g of water at 20°C). Liquid chloroform is very volatile but nonflammable; its gaseous form is capable of burning. Chloroform is used as a grain fumigant and as a general solvent for adhesives, pesticides, fats, oils, rubbers, alkaloids, and resins. It is also registered for use in the United States as an insecticidal fumigant on stored barley, corn, oats, popcorn, rice, rye, sorghum, and wheat; as a dry-cleaning agent; as an extraction and purification solvent for penicillin; as a soil fumigant and insecticide; and as a mildew preventive for tobacco seedlings. It was used in the past as a component of cough syrups, toothpastes, liniments, and toothache compounds. Chloroform was reviewed in Volumes 1, 3, and 4 of Drinking Water and Health (NRC, 1977, pp. 713-717; NRC, 1980, pp. 203-204; NRC, 1982, pp. 206-2091; the following is principally an examination of data that were not considered in the earlier volumes. However, some data reviewed in earlier volumes that are relevant to the current review are included here for completeness. Dibromochloromethane is a heavy, colorless-to-pale-yellow liquid used as a chemical intermediate in the manufacture of fire extinguishing agents, aerosol propellants, refrigerants, and pesticides. Its boiling point is about 118°C; its specific gravity is 2.38 and it has a density of 2.451 at 20°C. It is soluble in alcohol, ether, acetone, benzene, and organic solvents. Dibromochloromethane is formed during chlorination from naturally oc

]~2 DRINKING WATER AND HEALTH curring humic substances in raw water. The health effects of this by- product of drinking water disinfection were reviewed in Volume 3 of Drinking Water and Health (NRC, 1980, pp. 205-2064; the following material updates and reevaluates that information. METABOLISM In a paired l-test, Withey et al. (1983) compared gastrointestinal ab- sorption of chloroform administered by gavage in a mixture with either water or corn oil. Fifteen male Wistar rats weighing about 400 g were given doses of chloroform in corn oil at 75 mg/kg bw, and 15 additional rats of comparable weight and identical sex and strain were given the same dose in an aqueous solution. The peak concentration of chloroform in blood was reached at about the same time with either solvent (5.6 minutes for water, 6.0 minutes for corn oil), but because the bioavailability of chloroform was 8.7-fold higher when given in the aqueous solution, much more chloroform reached the blood after administration in an aqueous medium (39.3 ~g/ml of water and 5.9 ~g/ml of corn oil). Pohl et al. (1977) demonstrated the formation of 2-oxothiazolidine-4- carboxylic acid (OTZ) during incubation of chloroform with a liver mi- crosome preparation. Cysteine inhibited the binding of labeled chloroform to microsomal protein in vitro and apparently reacted with phosgene to form OTZ. When the incubation was performed in an atmosphere of labeled oxygen, oxygen-labeled phosgene was formed. Pohl et al. (1977) suggested that the formation of unstable trichloromethanol through the action of a cytochrome P450 monooxygenase yields phosgene with spon- taneous elimination of hydrochloric acid. That compound can react with cysteine and macromolecules. Pohl and Krishna (1978) developed data suggesting that deuterium-labeled chloroform was less hepatotoxic and less readily metabolized than unlabeled chloroform, indicating that cleav- age of the C-H bond is the rate-limiting step in the process that renders chloroform hepatotoxic. Additional studies have examined renal metabolism of chloroform and have found nephrotoxicity to be linked to chloroform metabolism, as is hepatotoxicity. Renal homogenate subfractions form OTZ and carbon monoxide when incubated with chloroform (Branchflower et al., 19841. Formation of reactive intermediates is decreased by manipulations that inhibit the cytochrome P450-mediated metabolism, such as incubation in the presence of CO, or addition of inhibitors such as SKF-525A, piperonyl butoxide, or metyrapone to the incubation mixture (Smith and Hook, 19841. These authors showed renal bioactivation of chloroform to be greater than that of liver, when activity is expressed relative to cytochrome

Selected Disinfectants and By-Products Il3 P450; emphasizing that while metabolism in the kidney may be of lesser importance for overall deactivation of compounds, the kidneys possess adequate enzyme activity to produce sufficient amounts of reactive me- tabolites to result in an adverse effect locally either toxicity, as dem- onstrated, or possibly carcinogenicity. A deuterium isotope effect for nephrotoxicity has also been reported (Ahmadizadeh et al., 19811. In an abstract of a study of possible free-radical formation by chloroform (Ruch et al., 1986), the effects of several substances on the toxicity of chloroform to hepatocytes isolated from mice were reported. The possible modifiers investigated were SKF-525A, an inhibitor of mixed-function oxidase; the antioxidants vitamin E and N,N'-diphenyl-p-phenylenedi- amine (DPPD); and a depletor of cellular glutathione, diethylmaleate. Chloroform by itself had a dose-related toxicity for hepatocytes from male B6C3F~ mice at 1 and 5 mM. The toxicity of chloroform for the hepatocytes was decreased at 2, 4, and 20 hours after a dose of SKF-525A. Vitamin E and DPPD did not affect the toxicity of chloroform. Diethylmaleate potentiated its toxicity at the three time periods mentioned above. In viva studies have shown that diethylmaleate potentiates toxic effects of chlo- roform on liver (Stevens and Anders, 1981) and kidney (Kluwe and Hook, 19811. Hewitt et al. (1980) presented evidence from organ weights, biochem- ical changes, and both qualitative and quantitative microscopic studies of tissue changes that n-hexane, 2-hexanone, 2,5-hexanedione (the metab- olite of hexane and 2-hexanone), as well as acetone, can markedly increase the toxicity of chloroform to the liver and kidney. These potentiators by themselves did not produce marked liver injury, but when administered by gavage to male rats 18 hours before intraperitoneal administration of chloroform (0.5 ml/kg bw), they markedly increased the hepatotoxicity and nephrotoxicity of chloroform in rats. The authors hypothesized that prior exposure to ketones or substances which are metabolized to ketones enhanced the susceptibility of the liver and kidney to toxic actions of haloalkanes. Cresteil and co-workers (1979) compared metabolism of chloroform by microsomes prepared from rat or human liver tissue (human livers were obtained from donors for renal transplantation). Binding of radioactivity from [~4C]CHCl3 was used as an index of reactive metabolite formation. Human microsomes catalyzed binding at a lesser rate than did rats (0.2- 0.7 nmol/mg microsomal protein/5 min for human and 1.1 + 0.3 for rats); both the apparent spectral dissociation constant (Ks) and Michaelis- Menten constant (Km) were less for human tissue. Irreversible binding was decreased by addition of cysteine to the incubation mixture. This is an indication of phosgene formation during the metabolism of chloroform.

i 14 DRINKING WATER AND HEALTH This study is important because it indicates that chloroform is metabolized by humans along a pathway similar to that studied in rodents and that human liver activity is comparable with that of rats. Fry et al. (1972) studied the elimination of ingested chloroform in expired air. Eight normal subjects (five men, three women), 18 to 50 years of age and weighing 60 to 80 kg, took gelatin capsules containing 500 mg of chloroform dissolved in 1 ml of olive oil about 1.5 hours after eating breakfast. Expired air was collected during the succeeding 8 hours and was analyzed for chloroform content. Between 17.~% and 66.6% of the ingested compound was recovered (mean recovery, 40.3%~. Two subjects (a man and a woman) took similar capsules containing [~3C]- chloroform, and their expired air was examined for its content of i3CO2. During 8 hours the man excreted 50.6% of the ingested label as i3CO2; the woman excreted 48.5%. Ingested chloroform appears, therefore, to be excreted principally in the breath (a mean of about 90% during the 8 hours after ingestion). About 45% is excreted as unchanged chloroform and about 55% as carbon dioxide. The concentration of chloroform in the blood reached a maximum about 45 minutes after ingestion of the capsules and then decreased in a biphasic manner. The initial part of the biphasic elimination curve had a mean half- time (four subjects) of about 14 minutes after attaining the peak concen- tration; the mean half-time for the second portion of the elimination curve was attained about 90 minutes after the peak concentration. The rate of respiratory excretion of chloroform as such was a linear function of the concentration of that substance in the blood and was an inverse linear function of the deviation of the body weight of the subject from their ideal weight. Thus, obese subjects excrete a smaller proportion of chloroform through their lungs than subjects of normal weight. The slope of the line relating pulmonary excretion of chloroform to the deviation of the subject's body weight from the ideal was greater for men than for women, probably reflecting the greater reservoir of lipoid in the female body. HEALTH EFFECTS Observations in Humans Because chlorination can lead to the formation of chloroform or other chlorocarbons in drinking water, there has been interest in the cancer incidence among persons whose drinking water has been chlorinated as compared with those whose drinking water has not been so treated. Several epidemiological studies have indicated an association between water chlo- rination and increased mortality rates from cancer (Cantor et al., 1977, 1978, 1985; Cragle et al., 1985; Kuzma et al., 1977; Page et al., 1976~.

Selected Disinfectants and By-Products ~ ~ 5 The contribution of this information to an evaluation of the ability of chloroform to induce neoplasms is made uncertain by the presence of confounding variables. In a case-control study of the relationship between the incidence of colon cancer and water chlorination in North Carolina, Cragle et al. (1985) estimated exposure to chlorinated water through 25-year residence his- tories. They found a statistically significant relationship between chlori- nation and colon cancer above age 60 but not below that age. Cantor and coworkers (1978) studied the association of the use of drinking water containing trihalomethanes (THMs) with cancer mortality in 923 U.S. counties, more than half of which were urban. Mortality data were classified by county of usual residence and compared with THM data obtained from the Environmental Protection Agency (EPA) surveys of water supplies in those counties. Concentrations of chlorinated THMs were subtracted from total THMs to obtain brominated THM data. Using a weighted linear-regression model to predict sex- and site-specific cancer rates, they found positive correlations between THM levels and mortality from cancers of certain sites. Bladder cancer rates in both sexes showed the strongest and most consistent correlation with an index of THM ex- posure. After establishing controls for differences in social class, ethnic group, and residence (urban or rural, degree of industrialization of the area, and section of the United States), the authors also found correlations with brain cancers in both sexes and with non-Hodgkin's lymphoma and kidney cancer in males. Deaths from brain cancer were associated with brominated THMs in both sexes and were correlated with magnitude of exposure. There was a suggested association between chloroform levels and kidney cancer in males. On the other hand, Cantor et al. (1985) found no increased risk of bladder cancer among people living in areas with chlorinated surface water above that of those living in areas with unchlorinated groundwater. More- over, among smokers, there was a negative association of the incidence of bladder cancer with the number of years of drinking chlorinated surface water. The pattern among ax-smokers was variable. Young et al. (in press) described a population-based case-control study of possible association between colon cancer and exposure to THMs. They compared 372 cases of colon cancer and 1,451 controls with respect to estimated exposure to chloroform and other THMs from 1951 to 1981. Working from questionnaires, they used data on water sources and resi- dences to estimate THM exposure. No association between exposure to THMs and the occurrence of colon cancer was found. Two earlier studies recounted the effects of chloroform on human health. Challen et al. (1958) studied employees of a confectionery company that also manufactured medicinal lozenges. The process involved the use of

116 DRINKING WATER AND HEALTH chloroform as a component of the dough from which the lozenges were made. The dough was mixed in closed mixers, but there was major escape of vaporized chloroform for 1.5 to 2.0 minutes during removal of the dough from the mixer, which was said to occur no more frequently than four times per day. When it occurred, the concentration of chloroform in the air of the workplace may have reached levels as high as 1,163 ppm (5,582 mg/m3~. Workers involved in manufacture of the lozenges com- plained of lassitude during the latter part of their workday-an effect persisting after they returned to their homes and in some cases even throughout weekends. Improved ventilation of the work area lowered the concentration of chloroform in the ambient air to a mean of 28% of that existing before the improvement and enabled employees to perform their duties comfortably. Physical examinations of 10 employees believed to have been exposed to concentrations of chloroform vapor ranging from 77 to 237 ppm (370 to 1,138 mg/m3) in the air revealed that 9 of the 10 complained of severe symptoms: a sensation of having a ball in the stomach, nausea, anorexia, loss of ability to concentrate, staggering, depression, irritability, and sev- eral other bothersome and in some cases painful effects. Among 10 em- ployees who had worked only after improvement of the ventilation of the work area and who had been exposed to concentrations of vaporized chloroform ranging from 22 to 71 ppm (106 to 341 mg/m3), a complained of less severe symptoms: dry mouth, borborygmi, and frequent micturition (probably due at least in part to hyperhydration). Despite the symptoms listed, no evidence of definite liver damage was found by either physical examination or tests of liver function (e.g., concentration of bilirubin in serum and thymol turbidity). Challen et al. concluded that the complaints by the workers whom they studied indicated that the maximum allowable concentration of chloroform in the air of the workplace should be no more than 50 ppm (240 mg/m31. Bomski et al. (1967) studied from 62 to 68 employees of a pharma- ceutical company who had worked in an environment containing 0.01 to 1.0 mg/liter (2 to 205 ppm) of chloroform in the air for 1 to 4 years. The researchers examined the incidence of hepatomegaly, mean concentration of albumin in the blood, and mean levels of serum glutamic oxaloacetic transaminase (SOOT) and serum glutamic pyruvate transaminase (SGPT) in four groups. A separate group of employees was identified who had had less contact with chloroform (19 to 39 people) than the previous group, and an additional group was identified who had had viral hepatitis but had not worked with toxic substances (19 to 23 people). Finally, a control group was identified who had no history of jaundice and had not worked with toxic substances (86 to 164 individuals). (Ranges of numbers of people in the various groups are given in parentheses because different

Selected Disinfectants and By-Products 117 numbers of people in the groups volunteered to be tested for the separate parameters listed.) Hepatomegaly was found in 25% of the employees exposed to chlo- roform for a comparatively long time, in 12.8% of the employees exposed to chloroform for a shorter time, in 8.7% of the people who had had viral hepatitis, and in 3.7% of the control group. The lowest mean concentration of albumin in blood was found in employees with comparatively great exposures to chloroform, but the differences from the mean values for the other groups were not significant. The mean levels of the two transami- nases in the sera of the heavily exposed group of employees were lower than those of any of the other three groups. Electrophoretic fractionation of sera of the four experimental groups and three reactions that detect abnormal proteins in sera indicated that there was no significant change in the composition of the blood of the most heavily exposed group. This study suggests that exposure to chloroform may induce hepatic enlarge- ment without persistence of other indices of injury to the liver. However, administration of the bromsulfalein excretion test revealed that 61% of the 60 people exposed to chloroform for a comparatively long time ex- creted bromsulfalein at rates more than 5% lower than normal value. The relationship of these findings to the conclusions of Challen et al. (1958) is unclear, except for the premise that occupational exposure to chloroform may increase hepatomegaly with no evidence of continuing damage to the liver other than decreased biliary excretion of bromsulfalein. Observations in Other Species Acute Elects Earlier work established the susceptibility of the liver and kidney to chloroform, and more recent work has elaborated on these effects. Hill (1977) studied the effect of genetic and sex differences on the toxicity of chloroform in mice. He reported that many laboratory accidents involving the discharge of chloroform into quarters housing animals had been described in the literature and that males of some strains (but not females) had been found to die after exposure, whereas males and females of other strains survived. This suggestion of marked variations in suscep- tibility of specific strains led him to choose male mice of two inbred strains from the Jackson Laboratory as examples of two extremes of sensitivity to chloroform. DBA/2J was chosen as the sensitive strain and C57BL/6J as the resistant strain. The LD50 of chloroform administered orally in oil to DBA/2J mice (0.08 ml/kg bw) was about one-fourth that for C57BL/6J mice (0.33 ml/ kg bw). Male Fit hybrids (B6D2F~/J) had an LDso midway between those

1 1 ~ OR ~ N K'NG WATER AN D H EALTH of the two parent strains (0.20 ml/kg bw). No genotypic differences were found in the threshold or time courses of elevated enzyme activities fol- lowing administration of chloroform at several dose levels, nor were there different histological changes; each strain had centrilobular necrosis in the liver. However, the dose causing excessive loss of urinary glucose or protein in the DBA mice was about 60% of that causing the changes in C57BL mice. Again, the Fit hybrids were intermediate in their response. Twelve hours after oral administration of carbon-14-labeled chloroform, kidney homogenates from DBA mice contained significantly more radio- active carbon than did those of the C57BL mice; kidney homogenates from the hybrids were again intermediate. Hill (1977) referred to other studies indicating that such differences had not been found in whole-blood or liver samples and commented that female mice had survived in the accidental exposures that he had reviewed, suggesting a sex difference in addition to the genotypic difference. On the basis of his investigation as well as various studies described by others, Hill concluded that sensitivity to renal damage is the main factor in the sex and strain differences observed after administration of chloroform and that androgens play an important role in modulating at least some of these differences. Subchron~c Elects Chu et al. (1982a) gave rats drinking water con- taining chloroform concentrations of 5, 50, or 500 ppm for 28 days. Water intake was measured, and doses were calculated to be 0.13, 1.3, or 11 mglratlday. After the administration was ended, the rats were killed and examined grossly and microscopically; blood counts and analyses of serum were made, and hepatic microsomal and soluble enzymes were assayed. The only change seen was a decreased neutrophil count in rats given the highest dose. Chu et al. (1982b) gave weanling (94 to 100 g) Sprague-Dawley rats drinking water containing chloroform concentrations of 0, 5, 50, 500, or 2,500 ppm. Half of the animals in each group were killed at 90 days; the others were given tap water for another 90 days before being killed. At the highest dose, there were many deaths, decreased growth rate, and decreased food intake. Chu et al. also noted mild to moderate liver lesions that were not significantly different from those in controls and mild to moderate thyroid lesions that were significantly different from those in controls. The significantly different thyroid lesions were seen only in males at the highest dose, however, and there was some recovery within 90 days. There were no significant dose-related changes in biochemical or hematological elements. Munson et al. (1982) intubated groups of 14 to 24 CD-1 mice with chloroform and other halocarbons at 50, 125, or 250 mg/kg bw for 14 or

Selected Disinfectants and By-Products ~19 90 days. Possible effects that were investigated included changes in body weight, organ weight, blood count, bone marrow, clinical chemistry, hepatic microsomal enzyme assays, hexobarbital sleeping time, cell-me- diated and humoral immunity, and morbid anatomy. Creased liver weights, expressed both as absolute weights and as organ-to-body-weight ratios, were seen in males given 125 and 250 mg/kg bw that were killed on the fourteenth day. Liver-to-body-weight ratios in all female groups killed at 14 days were increased but were related to dose only at the two higher doses. There were increases of SGPT in high-dose males and females and increases of SCOT in high-dose females. There was a decrease in antibody- forming cells of the spleen in both males and females at all doses (p < 0.05), but hemagglutination titers were not affected and no alterations in cellular immunity were seen. Among rats treated for the full 90 days, there was a dose-related increase in liver weights (both absolute and organ- to-body-weight ratios) in high-dose males and in all female groups. Hepatic microsomal activities were decreased (p < 0.05) among high-dose males and in all treated groups of females; hexobarbital sleeping times were increased in all groups and significantly so in mid- and high-dose females. High-dose males had increased blood glucose and decreased humoral immunity. High-dose females had increased glucose and, together with mid-dose females, decreased humoral immunity. Cellular immunity was significantly decreased only in high-dose females. In the 90-day study, the absence of an increase in SOOT and SGPT activity, as seen in the 14-day study, suggested to the investigators that long exposure to chlo- roform may result in recovery from or development of a tolerance for- the hepatotoxic action of the chemical. Jorgenson and Rushbrook (1980) gave 6-week-old Osborne-Mendel male rats, weighing 190 g, chloroform in drinking water at concentrations of 200, 400, 600, 900, or 1,800 ppm. There were 30 animals per dose, and exposures were for 30, 60, or 90 days. One control group of 40 rats received water ad libitum, and a second control group of 30 rats had its water consumption matched with that of the group receiving the highest dose. From water intake data, dose levels of chloroform were calculated to be 0, 20, 3S, 57, 81, or 160 mg/kg bw per day. Changes in body weights, ratios of kidney fat to kidney weight, serum biochemical com- ponents, and macroscopic and microscopic anatomy were evaluated. The mean body weight of the highest-dose rats was decreased by 15%, and their mean body weight gain was decreased by 26.6%. The control rats, whose water intake was matched with that of the highest-dose rats, gained only 5.1% more weight than the highest-dose rats. A few biochemical changes were observed in rats given 400-ppm concentrations and higher, and these were attributed to reduced water intake. No effect on kidney

120 DRINKING WATER AND HEALTH fat was found. Pathological changes were slight or mild in severity, were not dose related, and were either sporadic or judged to be adaptive (ap- pearing in rats killed at 30 or 60 days but not in those killed at 90 days). Jorgenson and Rushbrook also performed a similar experiment with 6- week-old B6C3F~ female mice weighing an average of 19 g. Groups of 30 mice were given water containing 200-, 400-, 600-, 900-, 1,800-, or 2,700-ppm concentrations of chloroform. Two control groups similar to those in the rat experiment were included. From water intake data, doses were calculated to be 0, 32, 64, 97, 145, 290, or 436 mg/kg bw per day. Changes in body weights, ratios of organ fat to organ weight, and gross and microscopic anatomy were sought. Mice receiving 900 ppm and higher and those whose water intake was adjusted to that of the highest-dose group lost mean body weight during the first week, but thereafter their mean body weights were similar to those of controls. Several deaths occurred in some groups during the 13-week period of observation: 1 at 600 ppm, 2 at 900 ppm, and 4 at 2,700 ppm. There was much variation in water consumption. Ratios of liver fat to liver weight were significantly increased at 2,700 ppm. Macroscopically, there were occasional, very slight hemorrhages in mouse lungs at all doses. Microscopically, there were centrilobular fatty changes in mouse livers at the two highest doses, but they were mild and judged to be reversible. Extramedullary hema- topoiesis in the liver and lymphoid atrophy in the spleen were also observed but were judged to be unrelated to treatment. Chronic Elects Several investigations of chronic toxicity of chloro- form have been conducted (Heywood et al., 1979; Jorgenson et al., 1982; Palmer et al., 19791. Heywood et al. (1979) administered chloroform to beagle hounds. The chlorofon~ was mixed into a toothpaste base and administered in gelatin capsules. The administration continued for 7.5 years and was followed by a 20- to 24-week recovery period. A group of 16 males and females received toothpaste base without chloroform in doses of 0.5 ml/kg bw per day; another group of 8 of each sex was given another toothpaste base, also free of chloroforms; and 8 more of each sex were untreated. Groups of 8 of each sex received chloroform doses of 15 or 30 mg/kg bw per day in 0.5 ml of toothpaste. Eleven of the 64 dogs died during the study; two of them were chloroform-treated dogs. The only significant toxic change was a moderate rise in the activity in serum of such enzymes as SGPT in the high-dose group; this peaked in the sixth year and was believed to represent minimal liver damage. Fatty cysts were seen in the livers of several dogs. Although these could have been induced by chloroform, the incidence of the nodules did not seem to be dose related.

Selected Disinfectants and By-Products 121 Palmer et al. (1979) administered chloroform in a toothpaste base by gavage to Sprague-Dawley rats, 50/sex/dose level, at 0 or 60 mg/kg bw per day, for 6 days a week for 80 weeks. There was a marginal but consistent and progressive retardation in weight gain in rats of both sexes. The only significant (p < 0.01) change in organ weights was a decrease in relative liver weights of treated female rats. There was a decrease, reaching a maximum at week 52, in the cholinesterase activity of plasma but not in cholinesterase activity in erythrocytes among treated females. Differences in numbers or timing of deaths between control and treated rats were not significant. There were many minor microscopic hepatic changes but no severe fatty infiltration, fibrosis, or marked bile duct abnormality. There were slight increases in the incidence of moderately severe glomerulonephritis, but the significance of this effect was uncertain. Macroscopic or microscopic changes of a treatment-related nature were not seen in the brains. Tumor incidences among groups were not signif- icantly different. Jorgenson et al. (1982) gave chloroform in drinking water to male Osborne-Mendel rats at concentrations of 0, 200, 400, 900, or 1,800 ppm for 23 months. From data on body weights and water intake, which were monitored throughout the experiment, mean daily doses were calculated to be 0, 34, 66, 143, or 305 mg/kg bw per day. Body weight gains were inversely proportional to dose, but survival during 96 weeks was propor- tional to dose; that is, survival was lowest among the negative controls and highest in high-dose animals. It has been suggested that this result could reflect a beneficial effect of reduced body weight on survival; it could also reflect a beneficial effect of relative dehydration in view of the considerably greater survival among the matched controls than among the negative controls. Ten rats at each dose level were killed at 3 or 6 months, and liver triglyceride levels were measured. There was no increase in the mean percent of fat in the liver at any dose level except for an increase of 12% at the 1,800-ppm concentration at 6 months (p < 0.051. Groups of 20 rats/dose, killed at 6, 12, or 18 months, were examined for possible changes in blood elements. White-blood-cell (WBC) counts were de- creased in the high-dose group and in the matched controls at 6 and 12 months. Significant increases in red blood cells (RBCs) and hemoglobin at 12 months among the rats exposed to chloroform concentrations of 200, 900, and 1,800 ppm indicated some hemoconcentration, but significant differences were not seen among these same groups at 18 months. Serum chloride, potassium, phosphorus, bilirubin, alkaline phosphatase, total iron, albumin, and albumin/globulin ratio tended to be higher in treated than in control groups,-whereas cholesterol, triglycerides, lactic dehydro- genase, and globulin tended to be lower in treated than in control groups.

122 DRINKING WATER AND HEALTH The investigators attributed the changes to reduced consumption of food and water in the control groups. Mutagenicity Rapson et al. (1980) studied the mutagenicity of various compounds known or believed to be produced by treatment with chlorine or chlorine dioxide. They used the Ames mutagenicity assay with the TA100 strain of Salmonella typhimurium. Chloroform was tested at levels of 100 ng per plate; 1, 10, and 100 fig per plate; and 1 mg per plate and was not found to be mutagenic in this assay. On the other hand, Callen et al. (1980) showed that chloroform, as well as six other halogenated hydrocarbons (dichloromethane, halothane, carbon tetrachloride, tri- chloroethylene and tetrachloroethylene, and syn-tetrachloroethane) in duced genetic effects in yeast. Chloroform induced increased numbers of mitotic recombinants and of other genetic alterations in the ade2 locus of Saccharomyces cerevisiae strain D7. Frequencies of gene convertants at the trpS locus and of reverse mutations of the ilv1 marker also were increased, but without absolute increases in the numbers of induced events. Carcinogenicity Eschenbrenner and Miller (1945) administered chlo- roform by Savage at 0.15, 0.3, 0.6, 1.2, and 2.4 g/kg bw in olive oil and induced hepatomas in male mice. They used 3-month-old strain A mice from the National Cancer Institute, with a historical incidence of spon- taneous hepatomas of less than loo at 16 months of age. Dose groups included five animals of each sex. Chemical analysis of the chloroform was not indicated, but the compound was described as chemically pure. Doses were given every 4 days over 120 days for a total of 30 doses. The mice were killed for postmortem examination at 8 months of age 30 days after the last dose. They were given an additional dose of chloroform 24 hours before necropsy. None of the males of the three highest-dose groups (0.6, 1.2, and 2.4 g/kg bw) and none of the females of the highest-dose group survived; all deaths occurred 24 to 48 hours after the first or second administration. Liver necrosis was observed in both sexes at the three highest doses. Males in all groups developed dose-related renal necrosis, but this effect was not seen in females. All surviving females given doses of 0.6 or 1.2 g/kg bw developed hepatomas. Necrosis was not seen in tumor cells; hepatomas contained cords of enlarged liverlike cells that formed disorganized, anastomosed columns. The hepatomas were not judged to be invasive, and metastases were not found. The renal necrosis found in males was localized to the proximal and distal tubules; glomeruli and collecting tubules appeared not to be affected. These investigators confirmed the previously reported observa- tions that the Bowman's capsule of female mice is lined with squamous

Selected Disinfectants and By-Products 123 epithelium, whereas that of male mice is lined, at least partially, with cuboidal epithelium similar to that of the proximal convoluted tubule. The researchers also stated, however, that further research is necessary to determine whether this difference between the sexes has any relation to the tubular necrosis found in male mice gavaged with chloroform. Five other groups of mice, with one male and two females in each group, were given a single dose of chloroform at the same dose levels. There was a sharp distinction between normal and necrotic liver cells. At doses of 1.2 and 2.4 g/kg bw, there was extensive necrosis in all liver lobules, whereas at 0.6 g/kg bw there was necrosis in some lobes. Another bioassay of possible carcinogenicity was performed by the National Cancer Institute (NCI, 19769. Osborne-Mendel rats and B6C3F~ mice were intubated with chloroform in corn oil at two dose levels five times a week for 78 weeks. Male rats, 52 days old, were given 90 or 180 mg/kg bw. Female rats, 52 days of age, were given 125 or 250 mg/kg bw for the first 22 weeks; their doses were then reduced to 90 or 180 ma/ kg bw, so that their average doses were 105 or 171 mg/kg bw. Mice, 35 days old, were given 100 or 200 mg/kg bw (males3 or 200 and 400 ma/ kg bw (females); these doses were increased after 18 weeks to 150 or 300 mg/kg bw (males) and 250 or 500 mg/kg bw (females), so that the average doses were 133 or 265 mg/kg bw for males and 233 or 465 mg/kg bw for females. Rats were killed at 111 weeks. There was a decreased survival rate and weight gain in all treated groups. There were no renal tumors in control rats; but the low-dose group had an 8% incidence, and the high-dose group had a 24~o incidence (p = 0.00161. An increase in thyroid tumors, judged not to be biologically significant, was observed in female rats that had been given chloroform. Tumors were found mainly in the digestive, uri- nary, and endocrine systems; only a few tumors were found in other systems. Mice were killed after 92 or 93 weeks. Except for high-dose females, survival rates and weight gains were similar in all groups. Significant increases (p < 0.001) in hepatocellular carcinomas were observed in all treated groups; an incidence of 98% for males and 95% for females given the high dose and 36% for males and 80% for females given the low dose, compared with a 6% incidence in matched and in colony control males, 0 in matched control females, and 1% in colony control females. Many of the low-dose male mice that did not develop hepatocellular carcinoma had nodular hyperplasia in the liver. Roe et al. (1979) also investigated the carcinogenicity of chloroform, administered in toothpaste base or in arachis oil, to four strains of mice (C57BL, CBA, CF/1, and ICI). Ten-week-old mice were given chlorofo~ by stomach tube 6 days a week for 80 weeks, followed by a 16- to 24

124 DRINKING WATER AND HEALTH week rest period. The investigation was conducted in several parts: First, ICI mice, 52 of each sex per dose level, were given chloroform in tooth- paste at doses of 17 or 60 mg/kg bw per day; controls (104 of each sex) were given toothpaste without chloroform. Second, toothpaste without chloroform, peppermint oil, or eucalyptol was given to 260 male ICI mice; toothpaste with chloroform at levels equivalent to 60 mg/kg bw per day was given to a group of 52 mice; and a third group of 52 mice of each sex was untreated. Other groups of 52 males were given toothpaste aug- mented with peppermint oil or eucalyptol, or both, but not chloroform. In a third part of the investigation, four groups of 52 male mice, two groups/strain, were given chloroform in toothpaste at doses of 0 or 60 mg/kg bw per day, while a fifth group of 52 male ICI mice was given the same dose of chloroform in arachis oil. There were three control groups in this third phase: an untreated group of 100 ICI mice, another group of 52 ICI mice given toothpaste without chloroform, and a group of 52 ICI mice given arachis oil only. Body weights were recorded in each study, and food consumption was estimated in the second and third study phases. Adrenals, kidneys, livers, lungs, and spleens were weighed at the end of the study, and selected tissues and organs were examined microscopically. Liver and kidney weights were reported to be slightly lower in the chlorofo~-~-treated male ICI mice than in controls, and there were other significant organ weight changes that did not appear to fit any pattern; however, details were not provided in the report. Tumor incidences in control and chloroform-treated mice were not significantly different in male C57BL, CBA, CF/1, and female ICI mice. In male ICI mice treated with high doses of chloroform, there was an increased frequency of tumors of the renal epithelium, along with a greater incidence and severity of nonneoplastic renal disease. This effect was more pronounced in mice given chloroform in arachis oil in amounts equivalent to 60 mg/kg bw per day than in controls fed arachis oil. Malignant tumors were identified as hypernephromas and benign tumors as cortical adenomas. There was a significantly higher incidence (p < 0.001) of moderate to severe kidney lesions in CBA and CF/1 male mice and of moderate to severe kidney disease (p < 0.05) in male ICI mice treated with chloroform in arachis oil than that found in controls fed arachis oil. In addition, moderate to severe fatty degeneration of the liver was slightly more frequent among chloroform-treated mice than in controls. Jorgenson and coworkers (1985) administered chloroform in drinking water to male Osborne-Mendel rats and female B6C3F~ mice for 104 weeks at doses of 0, 200, 400, 900, or 1,800 mg/liter. Based on measured water intakes and body weights, the average doses were 0, 19, 38, 81, or 160 mg/kg bw per day for rats and 0, 34, 65, 130, or 263 mg/kg bw per day for mice. A second control group was paired for water intake with

Selected Disinfectants and By-Products 125 the high-dose groups; that is, their water intake was restricted to that of the companion test group. The administration was continued for 104 weeks. Group sizes were adjusted to achieve a detectable tumor response at low doses on the assumption that there was a linear relationship between tumor incidence and dose; thus, group sizes varied from 330 rats and 430 mice at doses of O and 200 mg/liter to 150 of each species at doses of 400 ma/ liter and 50 of each species at doses of 900 and 1,800 mg/liter. Monitoring of room air and animal feed showed that room air chloroform was consistently below 6 ppb and that there was no detectable amount in feed. The total amount of chlorinated hydrocarbon pesticides in feed was about 90 ppb. Malathion was also present at levels commonly below 50 ppb but sometimes as high as 220 ppb. High-dose male rats had a 14% incidence of adenomas and adenocar- cinomas of the renal tubules, compared with a 1% incidence in controls; this is consistent with, although slightly lower than, findings from a pre- vious investigation (NCI, 19761. In contrast to results of the previous investigation, however, an excess incidence of hepatocellular adenomas and carcinomas in mice did not occur. Jorgenson and coworkers (1985) suggested as an explanation that there may have been some interaction with the mode of administration (corn oil gavage) in the NCI study. Jorgenson et al. (1982) also reported that water consumption was de- creased in rats in a dose-related manner, but water intake tended toward normal at the two lower doses near the end of the study. There were decreases in body weight, from which it was inferred that food con- sumption had decreased; however, food consumption was not measured. Survival of rats was inversely related to dosage level of chloroform, and it was suggested that this might, in turn, be the result of the lower body weights of the rats given the higher doses. The pattern in mice was different, inasmuch as some mice at the two highest doses refused to drink water during the first week, and about 25% of each group died, as did 6% of those given doses of 400 mg/liter. After this initial period, survival did not differ significantly among groups. Neoplastic lesions other than the renal tumors in rats included increased incidences (as compared with controls) of neurofibromas, leukemias, lym- phomas, and circulatory system tumors and decreased incidences of renal tumors and thyroid C-cell adenomas. However, these were not believed to be the result of chloroform administration because they were not dose related or statistically significant; nor did they represent a natural pro- gression of tumors, such as from C-cell adenoma to carcinoma, which might have been expected because of the longer survival of animals to which chloroform was administered. Nonneoplastic lesions were frequent in all groups of rats. For example, there was a 91% incidence of nephropathy in the negative control rats and

]26 DRINKING WATER AND HEALTH 90% in the matched controls versus 92% to 100% in the rats exposed to chloroform, there was a 92% to 100% incidence. Pereira et al. (1985) administered ethylnitrosourea (ENU) intraperito- neally in doses of 0, 5, or 20 mg/kg bw to 15-day-old outbred CD- 1 Swiss mice. At weaning, at 5 weeks of age, 23 to 29 mice at each dose level were given drinking water containing 1,800-ppm concentrations of chlo- roform. Other groups of 25 to 36 of the ENU-treated mice were given sodium phenobarbital in water at concentrations of 500 ppm. After 46 weeks of exposure, that is, at 51 weeks of age, the mice were killed for necropsy. There was dose-related development of liver adenomas and hepatocellular carcinomas in males and of lung tumors (mostly adenomas) in both sexes that were attributable to ENU alone. The administration of chloroform seemed to inhibit tumor development in male mice that had received ENU. Phenobarbital seems to have increased the incidence of liver tumors among ENU-treated mice. A significant effect of chloroform or phenobarbital on lung-tumor incidence was not evident. As Jorgenson and his coworkers (1985) had done, Pereira et al. (1985) suggested that the differences between the results of their study, in which chloroform was found to inhibit the development of hepatic tumors, and those of the NCI (1976) study might be due to the different vehicles used: water versus corn oil. They speculated that toxicokinetic differences might have been engendered by the administration of chloroform as a bolus in corn oil or, alternatively, that there might be a synergistic interaction of chloroform and corn oil. In the study of Pereira and associates (1985), however, the only groups whose body weight gains were reduced were male mice given chlorofo~. Further pursuing this hypothesis, Bull et al. (1986) examined more specifically the effect of the vehicle on chloroform toxicity. Male and female B6C3F~ mice were given chloroform either in corn oil or in 2% Emulphor an emulsifying agent used to produce aqueous emulsions of lipophilic chemicals in water-by Savage at doses of 0, 60, 130, or 270 mg/kg bw per day; 10 mice of each sex were tested at each dose level in each vehicle, for a total of 80 mice of each sex. Administration was continued for 91 to 94 days, after which all animals were killed for necropsy. Urine was collected from metabolism cages the day before the mice were killed, and blood was obtained by cardiac puncture at necropsy. Brains, livers, spleens, lungs, thymus glands, kidneys, hearts, and gonads were removed and weighed, after which liver tissue sections were prepared and a liver lobe was processed for lipid analyses. At 270 mg/kg bw per day, chloroform caused an increase in SOOT when given in corn oil but not when given in Emulphor. There was no effect on lactic dehydrogenase (LDH). There was a small increase in blood

Selected Disinfectants and By-Products 127 urea nitrogen (BUN) in those given the corn oil solutions but not in those receiving the Emulphor suspensions. There was a decrease in body weight and an increase in liver weight from chloroform treatment regardless of the vehicle, but the effects were greater in those given the corn oil. Chloroform in corn oil caused a significant amount of diffuse parenchymal degeneration in the liver and mild to moderate early cirrhosis. Corn oil alone or chloroform in Emulphor did not cause significant pathological changes. Thus, chloroform in corn oil caused more marked hepatotoxic effects in mice than did chloroform in Emulphor or, in comparison with other investigations (Jorgenson et al., 1985), than did chloroform in water. This seems to suggest that the discrepancy between the incidences of liver tumors in the investigations of NCI (1976) and of Jorgenson et al. (1985) can be accounted for by the vehicle of administration, chloroform in corn oil causing more marked hepatic effects than chloroform in other vehicles. Moore et al. (1982) also developed data on some possible effects of vehicle. Toothpaste containing 0.0325%, 0.94%, or 3.59% of chloroform and corn oil solutions containing 0.045%, 1.8%, or 7.2% chloroform were given to male mice by Savage in amounts to yield chloroform doses of about 15, 60, and 240 mg/kg bw. Three days later, the investigators measured the incorporation of t6-3H]-labeled thymidine into liver and kidney during 24 hours and examined plasma heparinized to determine its urea concentration and glutamic oxalacetic transaminase (GOT) and glutamic pyruvic transaminase (OPT) activities. Control groups received only toothpaste without chloroform or plain corn oil but otherwise were given everything that the experimental groups received. The mean concentration of urea in plasma was increased to 5 to 8 times the mean control value after the highest doses of the chloroform-containing preparations were administered; the smaller doses had no definite effect on urea in plasma. The only clearly significant effect on urea levels in the plasma was observed among the mice given chloroform in corn oil. This group also had the only clearly significant increases (to 2.2 times the control value) in the activity of SGPT. The uptake of labeled thymidine by kidney and the weight of the kidney increased significantly after the largest dose of chloroform was given in either vehicle. The intermediate dose of chloroform in corn oil also resulted in a significant increase in kidney uptake of labeled thymidine, whereas the largest dose of chloroform in corn oil resulted in significant increases in the uptake of thymidine by the liver. No dose of chloroform in the toothpaste altered significantly the uptake of labeled thymidine into the liver. The largest doses of chloroform in either vehicle-induced tubular ba- sophilia and necrosis in the kidneys. The livers of mice that had received

128 DRINKING WATER AND HEALTH the largest dose of chloroform in corn oil had undergone centrilobular enlargement and some necrosis, whereas the largest dose of chloroform in toothpaste resulted in about 0.6 as much centrilobular enlargement and no evidence of necrosis of liver cells. The intermediate dose of chloroform in corn oil resulted in basophilia and necrosis of the tubules of the kidneys of some mice; the intermediate dose in toothpaste produced little evidence of liver or kidney toxicity. The agreement between the effects of chloroform administration on the concentration of urea in plasma and on renal integrity is striking. Similarly, the agreement between the effects of chloroform on the activity of OPT in the plasma and on the morphology of the liver is impressive. Further- more, the uptake of thymidine into the kidney corroborated the histological finding of kidney necrosis. Moore et al. (1982) discussed the results obtained in an earlier study (Roe et al., 1979), in which a daily chloroform dose of 60 mg/kg bw had been found to induce the formation of renal tumors but not hepatic tumors and a smaller daily dose of chloroform (17 mg/kg bw) had not resulted in renal tumors. Moore and his colleagues concluded that early acute cellular damage and subsequent repair are requisite for the production of tumors in target organs for this trihalogenated hydrocarbon. Herren-Freund and Pereira (1986) administered diethylnitrosamine (DENA) to male rats 18 hours after partial hepatectomy and 7 days before they were given drinking water containing 1,800-ppm concentrations of chloroform for 10 weeks. No evidence of the ability to promote carcin- ogenesis, such as an increased incidence of ~y-glutamyl transpeptidase (GOT) foci, occurred. However, when chloroform was given in the drink- ing water at the same time as weekly doses of DENA, there was an increased incidence of liver tumors. In a further attempt to study tumor promotion by chloroform, 15-day-old Swiss mice were given ENU; at weaning, they were given either chloroform (1,800 ppm) or sodium pheno- barbital in drinking water (500 ppm), the latter compound being used as a positive control. The chloroform and barbiturate were given until the mice were 51 weeks old, when they were necropsied. Administration of ENU at 5 and 20 mg/kg bw caused a dose-related increase in liver tumors. Chloroform inhibited the development of spontaneous and ENU-induced liver tumors in male mice, but phenobarbital enhanced this development in mice of both sexes. Klaunig et al. (1986) also investigated tumor promotion by chloroform. Thirty-five 4-week-old male B6C3F~ mice imbibed drinking water con- taining DENA in doses of 10 mg/liter for 4 weeks, after which they were given various chlorinated hydrocarbons, including chloroform, in their drinking water at the maximum tolerated dose (MTD) or at one-third of the MTD (1,800 or 600 mg of CHCl3/liter of water). One group of controls

Selected Disinfectants and By-Products 129 received DENA followed by untreated drinking water. Another positive control group received drinking water containing phenobarbital in doses of 500 mg/liter. Mice were killed at 6 months (10/group) and at 12 months (25/group) and were examined grossly and microscopically for liver tu- mors. There was no increased incidence of tumors among mice given chloroform after the course of DENA exposure, but phenobarbital en- hanced the production of liver tumors when given after exposure to DENA. The carcinogenicity of dibromochloromethane was assessed by the Na- tional Toxicology Program in a 2-year rodent bioassay (Dunnick et al., 1985; NIP, 1985~. Both sexes of Fischer 344 rats and B6C3F~ mice were Savaged 5 days per week for 104 weeks (rats) or 105 weeks (mice). Solutions were made up with corn oil and administered in doses equivalent to 5 ml/kg bw. The doses used were 0, 40, and 80 mg/kg bw for rats and 0, 50, and 100 mg/kg bw for mice. Mice on the dose equivalent to 50 mg/kg bw were given approximately 7 times the target dose at week 58. All females survived and were retained in the study; 35 of the males died, leaving an insufficient number (15) for analysis. There were no statistically significant increased incidences of neoplastic lesions in either male or female rats. Both sexes of rats showed dose-related liver toxicity, as fatty change, a ground-glass appearance of hepatocytes, and decreased cyto- plasmic basophilia. Both sexes of mice showed increased incidence of nonneoplastic liver lesions and the males a greater incidence of nephrosis. The significant liver lesions in males were hepatocytomegaly (12/50 versus 0/50 in controls) and necrosis (9/50 versus 2/50 in controls), and in females were calcification (7/50 versus 0/50 in controls) and fatty metamorphosis (28/50 versus 7/50 in controls). Liver carcinomas were significantly in- creased in males (19/50 versus 10/50 in controls) but not in females, whereas adenomas were increased in females (11/50 at the dose equivalent to 100 mg/kg bw versus 2/50 in controls) but not in males. The combined incidence of adenoma or carcinoma was significant for females by inci- dental tumor test and life-table analysis, but for males the combined incidence was significant only by life-table analysis. CONCLUSIONS AND RECOMMENDATIONS The evidence that THMs are capable of inducing cancer gives rise to a complex set of issues in risk assessment. Convincing evidence that the THMs, per se, are carcinogenic to humans is lacking. There is some evidence that associates increased cancer risk with the chlorination of drinking water (see "Epidemiological Studies" in Chapter 3), but THMs are not the only mutagenic and potentially carcinogenic by-products that are formed (see "Observations in Humans" above). Therefore, calculation of the levels of risk to humans who consume drinking water containing

130 DRINKING WATER AND HEALTH these compounds must rely on data that demonstrate the carcinogenicity of the individual THMs in experimental animals. Outlined above are studies that have described increased renal tumors in male Osborne-Mendel rats (Tables 4-1 and 4-2), liver tumors in female B6C3F~ mice (Tables 4-3 and 4-4), and renal tumors in male B6C3F~ (Table 4-5) and ICI mice (Table 4-6J. The subcommittee con- sidered that the data obtained in the study of Osborne-Mendel rats given chloroform in drinking water (Jorgenson et al., 1985) provided the most appropriate basis for estimating the carcinogenic risk to humans (Table 4-21. Although a relatively low response, the observation was repro- ducible in two different studies with varying experimental designs (Jor- gensen et al., 1985; NCI, 19761. Second, the mode of administration (drinking water) in this study was the most appropriate. Finally, the relatively large numbers of animals used provide a basis for estimates of risk with smaller confidence intervals. The increased incidence of liver tumors in B6C3F~ mice was ex- amined, and the extrapolation calculations were completed (Tables 4-3 and 4-41. The committee does not believe, however, that this study provides useful data for estimating risk to humans because the induction of these tumors by chloroform appears to be dependent on the use of a vehicle that contains large amounts of polyunsaturated fat (i.e., corn oil or olive oil). The liver tumors could not be reproduced in these animals at similar doses of chloroform in drinking water (Jorgenson et al., 19851. As discussed above, the use of such an oil vehicle appears to increase the toxicity of chloroform to the liver, as evidenced both by pathological examination and by indications of regenerative hyper- plasia. Data from Roe et al. (1979) suggest that renal tumors produced in male ICI mice may involve a similar interaction with the oil vehicle (Table 4-71. The estimation of carcinogenic risk calculated in Table 4-8 takes these factors into consideration. In the National Toxicology Program study of dibromochloromethane, results in B6C3F~ male mice show some or equivocal evidence, but not clear evidence, of carcinogenicity. Tumor incidence is shown in Table 4-9. The protocol for these studies was the same as that used for the NCI bioassay of chloroform and suffered from the same limitations as described for the chloroform study, i.e., extrapolation to exposure in water with consumption occurring over a period of time (rather than a single bolus) and the use of corn oil as the vehicle. For dibromo- chloromethane the problem is compounded by loss of the low-dose group of male mice (dose relatedness cannot be determined). Further studies are needed, and a protocol relevant to consumption in drinking water should be employed. The carcinogenic risk to humans of ingesting this compound is estimated in Table 4-10.

Selected Disinfectants and By-Products 131 TABLE 4-1 Tumor Incidence in Male Rats Fed Chloroform in Drinking Watera TumorbDose, Tumor Animal Sex Sitemg/kg low/day Rates Osborne-Mendel rats Male Kidney0 4/301 19 4/313 38 4/148 8 1 3/48 160 7/50 aBased on data from Jorgenson et al. (1985) bTubular cell adenomas and adenocarcinomas. TABLE 4-2 Carcinogenic Risk for Chloroforma Estimated for Humans with the Linearized Multistage Model Estimated HumanCancer Risk, Animal Sex Lifetime Riskb Upper 95% Confidence Levelb Osborne-Mendel rats Male 5.16 x 10-8 8.9 x 10-8 aBased on data from Jorgenson et al. (1985). bAssuming daily consumption of 1 liter of water containing the compound in a concentration of 1 ~g/liter. TABLE 4-3 Tumor Incidence in Female Mice Fed Chloroform in Corn Oil by Gavagea Tumor Dose, Tumor Animal Sex Site mg/kg low/day Rates B6C3F~ mice Female Liver 0 ~0/80 138 36/45 277 39/41 aBased on data from NCI (1976). TABLE 4-4 Carcinogenic Risk for Chloroforma Estimated for Humans with the Linearized Multistage Model Estimated Human Cancer Risk, Animal Sex Lifetime Riskb Upper 95% Confidence Levelb B6C3F~ mice Female 1.5 x 10-6 1.9 x 10-6 aBased on data from NCI (1976). bAssuming daily consumption of 1 liter of water containing the compound in a concentration of 1 ~g/liter.

132 DRINKING WATER AND HEALTH TABLE 4-5 by Gavagea Tumor Incidence in Male Mice Fed Chloroform in Corn Oil Animal Tumor Dose, Sex Site mg/kg low/day o 138 277 Tumor Rates 1/77 1/50 2/45 B6C3F~ mice Male Kidney aBased on data from NCI (1976). TABLE 4-6 Carcinogenic Risk for Chloroforma Estimated for Humans with the Linearized Multistage Model Estimated Human Cancer Risk, Animal Sex Lifetime Riskb Upper 95% Confidence Levelb B6C3F~ mice Male 1.7 x 10-8 4.7 x 10-8 aBased on data from NCI (1976). bAssuming daily consumption of 1 liter of water containing the compound in a concentration of 1 ~g/liter. TABLE 4-7 Tumor Incidence in Male Mice Fed Chloroform in Toothpaste Basea Tumor Dose, Tumor Animal Sex Site mg/kg low/day Rates ICI mice Male Kidney 0 0/72 17 0/37 60 8/38 aBased on data from Roe et al. (1979). TABLE 4-8 Carcinogenic Risk for Chloroforma Estimated for Humans with the Linearized Multistage Model Estimated Human Cancer Risk' Animal Sex Lifetime Riskb Upper 95% Confidence Levelb ICI mice Male 1.605 x 10-~° 3.6716 x 10-7 aBased on data from Roe et al. (1979). bAssuming daily consumption of 1 liter of water containing the compound in a concentration of 1 1lg/liter.

Selected Disinfectants and By-Products 133 TABLE 4-9 Tumor Incidence in Female Mice Fed Dibromochloromethane in Corn Oil by Gavagea Tumor Dose, Tumor Animal Sex Site mg/kg low/day Rates B6C3F~ Mice Female Liver O 50 00 6/50 10/49 19/50 aBased on data from NIP (1985). TABLE 4-10 Carcinogenic Risk for Dibromochloromethanea Estimated for Humans with the Linearized Multistage Model Estimated Human Cancer Risk, Animal Sex Lifetime Riskb Upper 95% Confidence Levelb B6C3F~ Mice Female 5.7 x 10-7 8.3 x 10-7 aBased on data from NIP (1985). bAssuming daily consumption of 1 liter of water containing the compound in a concentration of 1 ~g/liter. HALOACI DS DichIoroacetic Acid CAS No. 79-43-6 TrichIoroacetic Acid CAS No. 76-03-9 CHC12COOH CC13COOH Dichloroacetic acid (DCA), a colorless liquid with a pungent odor, is soluble in water. It boils at 193-194°C and has a density of 1.563. Its two crystalline forms melt at 9.7°C and-4°C (Windholz et al., 1983, p. 3,038~. Dichloroacetic acid is used as a chemical intermediate and in pharmaceuticals and medicine (Hawley, 1981, p. 3324. Trichloroacetic acid (TCA) takes the form of nonflammable, deliques- cent colorless crystals, also having a sharp pungent odor. The crystals melt at 57.5°C and boil at 197.5°C. At 25°C, 1.2 kg of TCA crystals is soluble in 1 liter of water. The compound is used in organic synthesis, as a reagent for detection of albumin, in medicine for the removal of warts and as an astringent, in pharmacy, and in herbicides.

134 DRINKING WATER AND HEALTH METABOLISM Harris and associates (1978) developed evidence that dichloroacetate (DCA) promotes leucine catabolism by the formation of glyoxylate, which acts as a substrate for the transamination of leucine by glyoxylate amino- transferase (GAT), an enzyme found in peroxisomes. Liver cells from male rats were incubated with t2-~4C]DCA, and labeled amino acids were assayed, as were such enzymes as cx-ketoisocaproate dehydrogenase and pyruvate dehydrogenase (PDH). The increased formation of t1-~4Cla-ketoisocaproate from cE1-~4C]leucine induced by DCA was not affected by valine, isoleucine, glutamate, lysine, tryptophan, praline, arginine, or threonine but was suppressed by serine, methionine, asparagine, glutamine, phenylalanine, and glycine. Aspara- gine was the most potent of the inhibitors, followed by phenylalanine and methionine. The effects of these amino acids were attributed largely to their influence on the accumulation of a-ketoisocaproate. This accumu- lation, in turn, was attributed to activation of the transamination of leucine rather than an effect of DCA on cx-ketoisocaproate dehydrogenase. It was concluded that DCA enhances leucine catabolism by the for- mation of glyoxylate, which acts as a substrate for the transamination of leucine by GAT. In support of this conclusion, it was pointed out that (a) the amino acids that act as substrates for GAT block the stimulatory effect of DCA on leucine oxidation; (b) glyoxylate stimulates this oxidation; (c) both DCA and glyoxylate increase the glycine content of isolated hepa- tocytes; (d) labeled DCA is converted by isolated hepatocytes to labeled glycine, oxalate, and CO2; (e) 2-chloropropionate (2CP) also activates PDH but is not converted to glyoxylate and does not stimulate the oxidation of leucine; and (f) ethylene glycol, which is also converted to glyoxylate by the liver, also stimulates leucine oxidation. Dierickx (1984) studied the interaction of acetic acid (AA), mono- chloroacetic acid (MCA), dichloroacetic acid (DCA), and trichloroacetic acid (TCA) with rat liver glutathione S-transferase (GST) in vitro. Initially, the reaction of these acids was studied in crude rat liver supernatants. The inhibition of GST activity, with glutathione (GS) and 1-chloro-2,4-dinitro- benzene (DNCB) as substrates, was dose-dependent but not linear. He speculated that the ability of TCA to precipitate proteins might account for the alterations in GST activity but concluded this was not the case under his experimental conditions after trying to induce such protein pre- cipitation at concentrations of 50 mM. These acids inhibited six of the seven GST isoenzymes from rat liver, but to greatly differing degrees; for example, GST A was much more inhibited by AA and MCA than was GST B. The seventh (GST D) was present in too small a quantity for the study. Kinetic experiments did not

Selected Disinfectants and By-Products 135 reveal competitive inhibition kinetics, indicating an attack by the acids at the same site as that of the substrates, tripeptide glutathione (GSH) and DCNB. Titration of remaining GSH in incubation mixtures did not indicate a catalytic function of GST in the conjugation of GSH with AA, MCA, DCA, or TCA, so it was concluded that the major mechanism of interaction involves the binding to GST. This suggested to the investigator that GST could have a protective function against these compounds. Blackshear et al. (1974) infused fasted male rats with aqueous solutions of sodium dichloroacetate (300 mg/kg bw per hour for 1 to 4 hours at a rate of 1.2 ml/hour). Other animals were similarly infused with NaCl solutions. The rats were anesthetized with sodium pentobarbitone for cath- eterization of the femoral artery and vein, the artery for withdrawal of arterial blood for analyses at various times during the infusion and the vein for infusions of the DCA or NaCl. There was a significant and rapid decrease in blood glucose, lactate, and pyruvate in rats treated with DCA in comparison with those treated with NaCl; there was also an unac- counted-for rise in blood glucose in the control (NaCl-treated) rats. Plasma insulin decreased. 3-Hydroxybutyrate and acetoacetate increased signifi- cantly also. Some animals were functionally hepatectomized after the infusion; that is, they were anesthetized with sodium pentobarbitone by vein, and lig- atures were placed around the celiac and superior mesenteric arteries and the hepatic portal vein, thereby preventing the pooling of blood in the viscera and producing the functional hepatectomy. The infusion of DCA caused significant decreases in hepatic glucose, glucose-6-phosphate, 2- phosphoglycerate, lactate/pyruvate ratio, citrate, malate, alanine, and glu- tamate and glutamine, indicative of a restriction in substrates for gluco- neogenesis. Rats so hepatectomized did not have a different disappearance of blood glucose following the 2-hour infusions, but they did have very significant decreases in the rate of accumulation of lactate, pyruvate, glycerol, and alanine compared with those of control animals. The in- creased accumulation of glutamine after functional hepatectomy seemed to compensate for the decreased accumulation of alanine. DCA caused a decrease in the clearance of ketone bodies. The authors concluded that DCA-induced hypoglycemia results from a decrease in the net release of extrahepatic precursors of gluconeogenesis, along with the inhibition of peripheral uptake of ketone bodies. Thus, their findings are consistent with previous indications that DCA activates PDH. DCA decreased glucose synthesis from lactate, pyruvate, and alanine in isolated rat hepatocytes, but it did not decrease such synthesis from substrates that do not involve pyruvate carboxylase, viz., propionate or glycerol (Demaugre et al., 19781. The DCA also inhibited pyruvate car- boxylation in isolated mitochondria, but only after a period of preincu

]36 DRINKING WATER AND HEALTH bation. Further, there was no effect on partially purified pyruvate carboxylase. DCA labeled with carbon-14 was incubated with hepatocytes or mito- chondria with or without exogenous substrate. The investigators inter- preted that the labeled DCA was biotransformed to labeled oxalate, which inhibited pyruvate carboxylase and mimicked the effects of DCA on mi- tochondrial pyruvate carboxylation. They concluded that DCA is trans- formed in isolated hepatocytes and mitochondria to oxalate, which, in turn, inhibits gluconeogenesis and pyruvate carboxylation. Crabb et al. (1981) have reviewed and summarized the effects of DCA on metabolic processes. DCA activates PDH in many tissues by inhibiting PDH kinase; similarly, DCA activates myocardial branched-chain cx-ke- toacid dehydrogenase by inhibiting branched-chain ~x-ketoacid dehydro- genase kinase (Paxton and Harris, 19841. Some of the effects of DCA in vitro are the result of biotransformation products of DCA, viz., oxalate and glyoxylate. DCA inhibited lactate gluconeogenesis by hepatocytes in an in vitro preparation because of the inhibition by oxalate of pyruvic carboxylase and stimulated the oxidation of leucine because of the trans- amination of leucine by glyoxalate. In vivo, DCA decreases blood glucose by restricting the supply in the liver of precursors of gluconeogenesis, an effect that is consequent to activation in peripheral tissues of pyruvic dehydrogenase. Crabb et al. pointed out that DCA lowers blood cholesterol in hyperlipidemic patients but did not offer a conclusion on the mechanism of this action. In their summary of toxic effects, they pointed out that DCA is neurotoxic, can cause cataracts, and may be mutagenic. Inasmuch as DCA was believed to activate PDH by inhibition of PDH kinase, O. B. Evans (1982) assayed PDH activity in muscle and liver tissues from rats that were administered DCA. The compound was given by intragastric intubation at 100 mg/kg bw either as a single dose or as a repeated dose given daily for 7 days. Three hours after a single dose of DCA, hepatic tissue concentrations of DCA increased to a maximum. Following the seven repeated doses of 100 mg/kg bw each, PDH activation was maximal 3 hours after and returned to basal activity 24 hours after the final dose. Hepatic tissue concentrations of DCA were maximal 3 hours after the last of the seven doses, and DCA was eliminated slowly over the next 3 days, with a half- life of 9.74 hours. DCA concentrations in liver and muscle were similar after this repeated administration. HEALTH EFFECTS Observations ~n Humans In a study of the metabolic effects of DCA in man (Stacpoole et al., 1978), daily oral doses of 3-4 g (43-57 mg/kg bw per day if body weights

Selected Disinfectants and By-Products 137 were 70 kg) of sodium dichloroacetate were administered for 6-7 days to patients ranging in age from 42 to 71 years and having diabetes mellitus or hyperlipoproteinemia or both. None had received any other treatment for 10 days before getting dichloroacetate. Seven women were studied for 7 additional days after treatment stopped, while three women and one man were studied in more detail for a 15-day period after treatment. DCA significantly reduced fasting blood glucose by an average of 24% (from hyperglycemic levels) and produced a marked, concomitant fall in plasma lactate (73%) and alanine (82%~. It also produced a significant decrease in plasma cholesterol (22%) and triglyceride (61%) and a significant in- crease in plasma ketone body levels (71%~. Urinary uric acid excretion and Orate excretion were reduced, with a concomitant elevation of serum uric acid. Maximum effects were seen at the end of the 6-7-day treatment period, and levels subsequently returned to pretreatment levels. Levels of plasma insulin, free fatty acid, and glycerol were unaffected. Plasma cholesterol levels were not affected by treatment in one patient, and, in the other patients, depressed cholesterol levels returned toward pretreat- ment levels after treatment stopped. Other laboratory tests showed that liver and kidney function were unaltered during or after DCA adminis- tration. A few patients felt mildly sedated, but no other laboratory or clinical evidence of adverse effects was noted during or soon after the treatments. Stacpoole and coworkers (1979) orally administered DCA at about 50 mg/kg bw to a 21 -year-old man with severe receptor-negative homozygous familial hypercholesterolemia, which had proved refractory to conven- tional dietary or pharmacological management. After 16 weeks of DCA, during which time his concentrations of total and low-density lipoprotein in the plasma fell markedly, he developed a polyneuropathy, characterized by weakness of facial, finger, and lower-extremity muscles, diminished deep tendon reflexes, and slowing of nerve conduction velocity. The neuropathy improved after the DCA therapy was stopped. No eye changes were found. There were no other adverse changes attributable to DCA therapy. Stacpoole and colleagues (1983) also administered sodium dichloro- acetate (NaDCA) to 13 patients hospitalized in an intensive-care unit. They ranged in age from 16 to 72 years. All had lactic acidosis, defined as a level of lactate in arterial blood of 5 mM/liter or higher, and their acidosis had been refractory to sodium bicarbonate treatment. Arterial blood pH was usually less than 7.30, and they were hypotensive. Two of the patients were inadvertently given sodium bicarbonate after initiation of DCA therapy, in violation of the intended protocol, so pharmacokinetic data from them were excluded in the analysis of the results; hence, con- clusions are based on the other 11 patients. DCA doses were 35-50 ma/

i38 DRINKING WATER AND HEALTH kg bw, given by vein for 30 minutes, with a second dose being given 2 or more hours after the first if plasma lactate levels exceeded 5 mM/liter and plasma DCA was less than 130 g/ml. Arterial plasma lactate decreased in all 11 patients (average decrease, 29%), and there was a slight rise in bicarbonate level and in pH. The DCA-induced reduction in lactate was significant (i.e., greater than a 20% reduction) in 7 individual patients. Arterial alanine levels were normal or elevated prior to treatment and fell an average of 54% in the 7 responding patients. Q-Hydroxybutyrate concentrations increased an average of 34% after DCA treatment. In all but 3 of the 13 originally selected patients, there was an increased arterial systolic blood pressure, usually occurring within a few minutes of infusion and lasting from several minutes to several hours. In 4 subjects whose cardiac output was monitored, the rise in systolic pressure was accompanied by a mean increase in cardiac output of 21%. The investigators concluded that their results were sufficiently encour- aging to warrant further studies of DCA in the treatment of lactic acidosis. They pointed out that, although the DCA therapy caused a marked im- provement in overall morbidity, only one of their patients survived to leave the hospital; previously, however, none of their patients with hy- perlactatemia of this magnitude had survived. Further, the patients whose hyperlactatemia was reversed by DCA died because of the failure to reverse their primary, life-threatening illnesses. It was thought that the DCA stimulated PDH activity, leading to accelerated oxidation of pyruvate, lactate, and alanine, as well as to increased bicarbonate formation and, consequently, to a rise in arterial pH. Lukas and coworkers (1980) administered sodium dichloroacetate to humans, rats, and dogs in a study of pharmacokinetic characteristics of this potential antidote for lactic acidosis. Labeled (TIC) or unlabeled so- dium dichloroacetate was given by vein to four human subjects (two at 10 and two at 20 mg/kg bw in 100 ml of saline). Maximal plasma con- centrations in the two subjects given 10 mg/kg bw were 19.9-24.7 ~g/ml and in those given 20 mg/kg bw, 57.3-74.9 ~g/ml. Half-lives for the four subjects varied between 20 and 36 minutes. Observations in Other Species Acute Elects Lukas and coworkers (1980) administered labeled (14C) or unlabeled sodium dichloroacetate by vein to three rats (100 mg/kg bw as 10% aqueous solution) and to two beagle hounds (100 mg/kg bw as 20% aqueous solution) in addition to the humans mentioned above. In the rat, maximal plasma DCA concentrations of 120 and 164 ~g/ml were

Selected Disinfectants and By-Products 139 obtained; subsequent declines were with half-lives of 2.1 and 4.4 hours. The same dose in dogs yielded maximal plasma concentrations of 447 to 508 ~g/ml, and the decline was slower, with half-lives of 17.1 to 24.6 hours. As mentioned earlier, the half-lives for the four human subjects varied between 20 and 36 minutes. Subchronic Effects Katz et al. (1981) administered sodium dichlo- roacetate daily for 3 months to CD rats and beagle hounds. Ten rats of each sex (131 to 156 g bw) were intubated with an aqueous solution at 125 and 500 mg/kg bw per day, and 15 of each sex were intubated with 0 or 2,000 mg/kg bw per day. Dogs received 0, 50, 75, and 100 mg/kg bw orally as a solid contained in gelatin capsules (4 of each sex at 0 and 100 mg/kg bw, 3 of each sex at 50 and 75 mg/kg bw). In rats, 2,000 ma/ kg bw proved lethal and 500 mg/kg bw nonlethal; in dogs, the lethal dosage was 75 mg/kg bw with 50 mg/kg bw nonlethal. Both species experienced reduced food consumption and body weight gain; hind limb weakness; frequent urination; progressive reduction in erythrocyte counts, hematocrit, and hemoglobin levels; reduced blood levels of glucose, lac- tate, and pyruvate; vacuolation of myelinated white tracts in the cerebrum and, to a lesser extent, in the cerebellum; and degeneration of germinal epithelium of the testes, with syncytial giant cell formation. Additionally, rats often had aspermatogenesis, whereas dogs had atrophy of the prostate gland, cystic mucosal hyperplasia in the gall bladder, hemosiderin-laden Kupffer cells, and eye lesions consisting of bilateral lenticular opacities, injected bulbar conjunctivae, and superficial corneal vascularization, with a tendency toward keratoconjunctivitis sicca. Some animals of each species were not killed at the end of the experiment, but were allowed a 1-month recovery period before being necropsied. During this period, there was some recovery; however, the lenticular opacities and gall bladder anom- alies in dogs, the brain lesions in both species, and the aspermatogenesis and loss of testicular germinal epithelium in rats persisted or improved only minimally. Organs were weighed at the time of necropsy. Livers of female rats at 500 and 2,000 mg/kg bw were significantly heavier than those of controls. Relative liver weights (not defined, but probably ratios of liver weights to body weights) were significantly increased at all doses among both sexes, in a dose-dependent manner. There were also significant increases in relative weights of kidneys (females at all doses) and adrenals (males at 500 mg/kg bw and females at 2,000 mg/kg bw). Absolute and relative organ weights of intoxicated rats approached those of controls at the end of the month's recovery period. The brain lesions in dogs were thought to be characteristic of edema, but their persistence in hounds permitted to recover for 5 weeks indicated

140 DR! N K! NO WATER AN D H EALTH to the investigators that more than simple edema was involved in the development of the lesions. Moreover, the lesions were not seen in the . . . optic or scam nerves. The high dose of 2,000 mg/kg bw resulted in a 13~o mortality rate in both male and female rats. Piloerection, tactile-induced vocalization, low body position, and unthriftiness occurred prior to death in male rats. Female rats that died first exhibited cachexia and unthriftiness. Dogs were dramatically more sensitive: one of three females died at a dosage level of 75 mg/kg bw and one of four males died at 100 mg/kg bw. In contrast to the changes observed after oral administration of DCA, the adminis- tration by vein to dogs (up to 100 mg/kg bw for 30 days) was without evident testicular, prostate, or central nervous system effects. Yount and coworkers (1982) compared the metabolic and toxic effects of DCA and 2-chloropropionate (2-CP) as activators of PDH in rats. In suckling rats, both compounds were effective in causing a lowered blood lactate and glucose and increased blood ketone bodies. The effects were similar except that DCA caused a greater increase in blood ketone bodies. In a longer-term study, young male rats were given 2-CP or DCA in their feed (0.04 M of either compound/kg of feed) for 12 weeks in a study of the effects produced by 2-CP or DCA. Rats given 2-CP and DCA ate less than controls. Initially, the reduction in feed consumption by rats and the consequent weight changes were more marked with 2-CP than with DCA; however, at 9 weeks, there was little difference in the mean body weights of the 2-CP and the DCA groups. Within 2-4 weeks, hind-limb weakness and abnormal gait were seen in both intoxicated groups. DCA-treated animals showed significantly reduced organ weights of spleen, lungs, heart, testes plus epididymides, and brain. In the 2-CP-treated group, the weights of the liver, kidneys, spleen, heart, testes plus epididymides, and brain were significantly less than in control animals. The DCA-treated rats had increased organ-to-body-weight ratios for adrenal glands, brain, and kidneys. There were consistent anatomic abnormalities seen in the testes of 2- CP-treated rats, and similar abnormalities were seen in DCA-treated rats. There was an arrest of testicular maturation and degeneration of germ cells, some of which contained enlarged or multiple nuclei. Ketone body levels were markedly increased in DCA-treated rats, and glucose levels were practically unchanged. Triacylglycerol levels were unchanged in DCA-intoxicated rats (although significantly decreased in 2-CP rats). Free glycerol was not affected. There were also changes in function and mor- phology of nerves; these are discussed in the later section on neurotoxic effects. Davis (1986), in a study of renal toxicity of chloroacetic acids, reported that DCA and TCA at lower doses did not affect food consumption,

Selected Disinfectants and By-Products 141 growth, or urine output, whereas higher doses caused decreased food consumption and lowered body weight gain. These effects were observed in male rats from daily gavage of TCA at 300 mg/kg bw or administration through drinking water at 3 g/liter, equivalent to about 250 mg/kg bw per day. No effects were observed at 0.3 g/liter, equivalent to about 25 ma/ kg bw per day. Impaired growth in male rats from DCA occurred only at the highest concentration in drinking water, 7.5 g/liter; however, in female rats, there was weight loss after exposure to drinking water con- taining 1.875 g/liter. No effects were observed in either males or females exposed to DCA in drinking water at 0.5 g/liter, equivalent to about 60 mg/kg bw per day. Exposures at the higher concentrations caused lower outputs of urine and elevated urine osmolalities. Hippurate excretion was not increased in the groups given water containing DCA; results from groups given TCA were not described. Mutagenicity Crabb et al. (1981) reviewed the evidence of mutagen- icity of DCA in bacterial systems, showing that there is some reason to believe that such mutagenicity stems from an impurity in the DCA, not from DCA itself. One impurity suspected of causing mutagenicity is di- chloroacetaldehyde. In an investigation of several compounds produced by chlorination, Rapson and associates (1980) found no evidence of a mutagenic effect by TCA at levels of 100 ng, 1 fig, 10 fig, 100 fig, and 1 mg per plate in S. typhimurium strain TA100. Waskell (1978) tested 0.45 mg of TCA in the TA100 strain of S. typhimurium. No evidence of mutagenicity was found at that level in the presence or absence of S9 liver homogenate. Carcinogenicity In a study by Herren-Freund et al. (Sydna Herren- Freund, Health Effects Research Laboratory, U.S. Environmental Pro- tection Agency, Cincinnati, Ohio, personal communication, 1986), both DCA and TCA administered to male B6C3F~ mice intraperitoneally at a dose of 5 g/liter of drinking water produced hepatocellular carcinomas in 61 weeks. Although these data were derived from an initiation-promotion study, the effect was seen in 81% of the animals receiving DCA and in 32% of those receiving TCA without prior initiation with ethylnitrosourea. These results should be confirmed by other studies, as no background incidence of hepatocarcinomas occurred in controls. Neurotoxic Effects Reversible polyneuropathy developed in a 21-year- old man who was administered DCA at about 50 mg/kg bw for 16 weeks (Stacpoole et al., 19791. Katz et al. (1981) described brain lesions and reversible hind-limb

]42 DRINKING WATER AND HEALTH weakness or paralysis in rats dosed orally with DCA for 90 consecutive days to rats at 125, 500, and 2,000 mg/kg bw per day. In dogs paralysis occurred at 100 mg/kg bw, and even dogs treated at 50 mg/kg bw suffered vacuolation of white myelinated tracts in the brain. White tract changes of the brain were still present in animals of both species after a 4-week recovery period. This study is described in greater detail earlier in this section. In the subchronic study of Yount et al. (1982) described earlier, young male rats were administered 2-CP or DCA in their feed at 0.04 M of either compound/kg feed for 12 weeks. In both groups of rats, there were sig- nificant reductions in nerve conduction velocities. At least some of the hind-limb weakness, abnormal gait, and smaller morphometric measure- ments of tibial nerves seen in rats intoxicated with DCA or 2-CP in this study were thought to have stemmed from impairment of maturation of nerves, rather than from a direct neurotoxic effect, although both of these effects could have occurred. In DCA-treated and in 2-CP-treated rats, there were significant reductions in nerve conduction velocities. Sciatic nerve lipid and protein contents (myelin phospholipid, myelin cholesterol, myelin protein, or total protein) were equivalent in both treated groups and controls. Reproductive Effects Katz et al. (1981) reported that the brain and testes appeared to be the principal target organs of DCA intoxication in the rat. All high-dose male rats had small testes, even after a 4-week recovery period. At 500 mg/kg bw, 40% suffered testicular germinal epithelial degeneration; at 2,000 mg/kg bw all animals were affected. Male dogs also experienced dose-dependent testicular changes and prostate glandular atrophy. More details of this study were described earlier in this section. After administering 0.04 M of DCA or 2-CP per kg of feed to rats for 12 weeks, Yount et al. (1982) consistently found arrested testicular ma- turation and degeneration of germ cells, some of which contained enlarged or multiple nuclei, in the testes of these animals. Testicular changes noted by Katz et al. (1981) were still present in some animals during a 4-week recovery period (see above). Other Elects Dogs receiving oral DCA at 50, 75, or 100 mg/kg bw per day for 13 weeks showed eye changes consisting of bilateral lenticular opacities, bulbar conjunctivitis, superficial corneal vascularization, and a tendency toward keratoconjunctivitis sicca; of these, the lenticular opaci- ties were found to be irreversible (Katz et al., 1981) On the other hand, after 16 weeks of administering 50 mg/kg bw of DCA to a 21-year-old man, Stacpoole et al. (1979) found no eye changes.

Selected Disinfectants and By-Products 143 CONCLUSIONS AND RECOMMENDATIONS DCA produces neurological, reproductive, and ocular effects. The neu- rological effects are seen in both the central and peripheral nervous sys- tems. Reproductive effects are seen in the testes, and ocular effects are mainly changes in the lenticular tissue. Less is known concerning the toxic effects of TCA. Davis (1986) has examined the effects of both DCA and TCA on renal function. The principal changes seen were decreased food consumption and lower body weight gain. No-observed-effect levels (NOELs) were 60 mg/kg bw per day and 25 mg/kg bw per day for DCA and TCA, respectively. Using these NOELs, suggested no-adverse-effect levels (SNARLs) may be calculated for DCA and TCA assuming that a 70-kg human consumes 2 liters of water daily, which contributes 20% of total intake. The committee chose uncertainty factors of 1,000 for these two substances because of the severity of the neurological and ocular effects found for DCA and because of the general lack of toxicity infor- mation, especially for TCA. The committee estimated SNARLs as: 60 mg/kg/day x 70 kg x 0.2 1,000 x 2 liters 25 mg/kg/day x 70 kg x 0.2 1,000 x 2 liters 0.420 mg DCA/liter, or 420 fig DCA/liter; 0.175 mg TCA/liter, or 175 fig TCA/liter. SNARLS may also be estimated for a 10-kg child consuming 1 liter of water daily, which contributes 20% of total intake: 60 mg/kg/day x 10 kg x 0.2 1,000 x 1 liter 25 mg/kg/day x 10 kg x 0.2 1,000 x 1 liter HALOALDEHYDES ChIoroacetaldehyde CAS No. 107-20-0 CH2ClCHO 0.120 mg DCA/liter, or 120 fig DCA/liter; 0.050 mg TCA/liter, or 50 fig TCA/liter.

~ 44 OR ~ N K'NG WATER AN D H EALTH Dich Ioroaceta ~ dehyde CAS No. 79-02-7 Trich loroaceta Idehyde, ChIoral CAS No. 75-87-6 CHCl2CHO CC13CHO Tnchloroacetaldehyde, one of the best known of the haloaldehydes, has a long history of use as a hypnotic agent in human medicine in its mono- hydrate form (chloral hydrate) but is no longer widely used clinically because it is habit forming with prolonged use. The oral hypnotic dose of chloral hydrate for humans is 0.3 to 1.0 g. A summary of the physical properties and solubility for three haloaldehydes is shown in Table 4-11. METABOLISM No specific data on the absorption rates of the haloaldehydes were found. Acute animal toxicity data available on chloroacetaldehyde, dichloro- acetaldehyde, and tnchloroacetaldehyde suggested absorption of these compounds from the intestinal tract since effects were seen after oral administration. Chloral hydrate produces rapid anesthesia in man, and thus its reaction product, trichloroethanol, must be rapidly absorbed from the gastrointestinal tract (Gilman et al., 1985, pp. 360-3621. Because the haloaldehydes are intermediate metabolites of the halo- ethylenes, the majority of metabolism data for the chloroacetaldehydes comes from the haloethylene literature. Vinyl chloride is metabolized to chloroethylene oxide and 2-chloroacetaldehyde by micro so mat cyto- chrome P450 (Guengerich et al., 19791. The 2-chloroacetaldehyde was thought initially to result from rearrangement of the epoxide. However, Miller and Guengerich (1982) found that the epoxide is not an obligate intermediate in the oxidative metabolism of trichloroethylene but is pro- duced along with chloroacetaldehydes and chloroacyl chlondes by various TABLE 4-1 1 Physical Properties and Solubility of Haloaldehydes . Compound Molecular Weight 78.5 1 12.9 147.4 Boiling Melting Specific Point Point Gravity Solubilitya Chloroacetaldehyde Dichloroacetaldehyde Trichloroacetaldehyde 85 90 97.8 3 2 - 57.5 1.51 1, 2, 3 aIn medium 1, water; 2, alcohol; 3, ether.

Selected Disinfectants and By-Products 145 transformations of a common intermediate product. Similarly, Liebler et al. (1985) found that vinylidene chloride is oxidized by microsomal cy- tochrome P450 to 2,2-dichloroacetaldehyde, 2-chloracetyl chloride, 2- chloroacetic acid, and 1,1-dichloroethylene oxide. These compounds were isolated from rat liver microsomes and hepatocytes. Because both vinyl chloride and vinylidene chloride are considered carcinogenic or potentially carcinogenic, respectively, the oxides were considered initially to be the most reactive metabolic products from their parent compounds and were postulated to react with deoxyribonucleic acid, whereas the haloaldehydes reacted with protein. In fact, the haloaldehydes were found bound to glutathione and microsomal protein in these in vitro experiments. Covalent binding of ti4C]vinylidene chloride to microsomes was inhibited by glu- tathione but not by lysine, suggesting that thiol groups rather than amino groups are the major targets for the reactive metabolic products (Liebler et al., 19851. Chloral is an intermediate in the catabolism of trichloroethylene (TCE), together with trichloroethylene oxide and other active intermediates. This was found in experiments in vitro with hepatocytes and P450 prepared from rat liver. In the studies, formation of the chloral appeared to be in a separate metabolic pathway than the epoxide intermediates (Miller and Guengerich, 1982, 19831. Chloral is metabolized further to trichloro- ethanol (TCEA) and trichloroacetic acid (TCAA); however, other inter- mediates may yield the same products. Humans exposed to 170 ppm TCE for 7 hours excreted 44% of the dose as TCEA and 18% of the dose as TCAA (Ogata et al., 197 11. The metabolism of trichloroethylene and monochlorobenzene was stud- ied in mouse, rat, rabbit, and man in vitro and in vivo. TCE was metab- olized to chloral by the hepatic microsomal drug-metabolizing enzyme system and then to trichloroethanol by nicotinamide adenine dinucleotide phosphate (NADPH)-alcohol dehydrogenase and NADPH-dependent un- specified enzyme in vitro. The enzyme metabolizing chloral to trichlo- roacetic acid was localized in mitochondria and supernatant fractions. TCE was metabolized mainly to trichloroethanol in rabbits and to trichloroe- thanol and trichloroacetic acid in rats, mice, and man in vivo. Within 5 days total excretion was 13-22% of the dose administered in all species. The ratio of mercapturic acid conjugates to 4-chlorocatechol conjugates after oral and subcutaneous administration of monochlorobenzene was 3- 9 in rats and after oral administration in man was >0.4 (Ogata, 19829. Butler (1948) injected dogs intravenously with chloral hydrate and found that the animals excreted TCEA and TCAA in urine. Dogs given doses of either chloral hydrate or TCE excreted most of the TCEA produced within their bodies as TCEA glucuronide in urine. The con- centration of chloral in venous blood plasma fell as that of TCEA rose

146 DRINKING WATER AND HEALTH to its peak at about 30 minutes after intravenous injection of chloral hydrate. The concentration of TCAA in venous blood plasma rose more slowly and to only about one-third the value to which the TCEA rose; it remained at much more of a plateau, however, than that of TCEA. Consequently, 120 minutes after the injections of chloral hydrate, the concentration of TCEA in the plasma was slightly lower than that of TCAA at the same time. Ikeda and Ohtsuji (1972) administered chloral hydrate in concentrations of 2.78 Mlkg bw to rats by intraperitoneal injection and quantified the metabolites 48 hours later. In the urine, they found 198 mg/kg bw of TCEA, 29.2 mg/kg bw of TCAA, and 236.3 mg/kg bw total trichloro- compounds. Quantification of these metabolites was performed using the Fujiwara reaction, a calorimetric reaction for TCAA. The values for tri- chloroethanol were obtained by difference after oxidizing all metabolites to TCAA. Sellers et al. (1972) concluded that reduction of chloral to TCEA by reductases of erythrocytes is more important than that by enzymes of the hepatocytes. Muller et al. (1974) gave three human subjects oral doses of chloral hydrate of 15 mg/kg bw and found that the concentrations of TCEA in blood during 100 hours thereafter were practically identical with those established by oral doses of TCEA of 10 mg/kg bw. The concentrations of TCAA in plasma after the dose of chloral hydrate were about 24.6% above those established by the dose of TCEA, but the time relations of these two curves were similar, both reaching peaks at 25 to 35 minutes after ingestion of the compounds and declining to about one-half the peak value by 60 to 70 hours after the doses. Ikeda et al. (1980) stated that the conversion of chloral hydrate to chloroethanol was apparently carried out in rat liver cytosol by at least two NADPH-dependent enzymes other than alcohol dehydrogenase. The formation of TCAA from chloral hydrate was not so well characterized. Earlier studies reported that chloral hydrate is not a substrate for human aldehyde dehydrogenase; however, Cooper and Friedman (1958) found that an aldehyde dehydrogenase prepared from rabbit liver mitochondria has converted chloral hydrate to TCAA. Observations in Humans Toxic doses of chloral hydrate (approximately 10 g in humans) produce severe respiratory depression and hypotension. Liver and kidney damage may also be induced. Observations in Other Species Acute Effects The oral LD50 values for chloroacetaldehyde in rats and mice range from 0.069 to 0.087 ml/kg bw (approximately 82 and 104

Selected Disinfectants and By-Products 147 mg/kg bw) (Lawrence et al., 19721. These authors indicated also that chloroacetaldehyde was highly irritating to the eye (3 on a scale of 0 to 3 at a concentration of a 0.25%) and less so to the skin (3 on a scale of O to 3 at a concentration of 7.5%~. Others have reported similar findings; that chloroacetaldehyde is corrosive and destructive to lipids and mem- brane structures, causing irritation of the eyes, mucous membranes, re- spiratory tract, and skin was reported by the U.S. Department of Health and Human Services (NIOSH/OSHA, 19811. These investigators (Lawrence et al., 1972) also found chloroacetal- dehyde to cause a dose-related increase in pentobarbital sleeping time, apparently indicating a decrease in liver enzyme activity. Mice were given chloroacetaldehyde by intraperitoneal injection or by inhalation for 3 days prior to pentabarbital administration. Doses or concentrations were 0.1, 0.2, or 0.5 times the LDso or the LTso (median survival time). The in- creases in sleeping times were statistically significant in all cases except at the shortest inhalation exposure, 15 seconds. Acute oral LDso values for trichloroacetaldehyde monohydrate in rats and mice have been reported to be 285 and 1,100 mg/kg bw, respectively (NIOSH, 1974, p. 1941. In two abstracts, Saunders and Harper (1980, 1981) reported that tri- chloroethylene and trichloroacetaldehyde produced damage to lungs, but not to livers, of rats given daily injections intraperitoneally of 0.2 to 2.0 g/kg bw for 5 days. This procedure reduced the mixed function oxidase activity of pulmonary microsomes by 25% to 50%. Trichloroacetaldehyde reduced pulmonary mixed function oxidase activity but did not cause any histological changes. In a behavorial study that will be described in more detail later, chloral given by oral gavage to mice in doses less than 0.1 LD50 was found to cause acute motor incoordination 5 minutes to 2 hours after dosing. The EDso for this effect was 84.2 mg/kg bw. Subchronic E;ffects Lawrence et al. (1972) administered chloroacet- aldehyde for 30 consecutive days to male Sprague-Dawley rats by intra- peritoneal injection. Dose levels were 1.88 and 3.76 ~1 (approximately 2.2 and 4.5 mg/kg bw) of chloroacetaldehyde/kg bw (0.3 and 0.6 times the intraperitoneal LDso). Mortality at the end of the study was 25% and 67%, respectively. Weight gains were significantly depressed (p < 0.05) in both dose groups. At the end of the 30-day period, hematocrit, he- moglobin, and erythrocyte counts were significantly depressed (p < 0.05) in rats receiving the high dose. The low-dose group had an increased count of segmented neutrophils and monocytes and a decreased lymphocyte count. Results from sulfobromophthalein (BSP) tests of liver function in both test groups were similar to those of controls. Significant differences in ratios of organ to body weights were observed in the case of several

]48 DRINKING WATER AND HEALTH organs. The weight ratios of brain and lungs were significantly increased in both low- and high-dose groups, and gonad, heart, liver, kidney, and spleen organ weight ratios were significantly increased in the high-dose group. The body weight gains of these two groups of animals decreased significantly, such that their weights were approximately half (at the high dose) to three-quarters (at the low dose) of control weights. These investigators (Lawrence et al., 1972) also performed a 12-week study of chloroacetaldehyde, in which male Sprague-Dawley rats were administered 0.32, 0.8, 1.6, or 3.2 1/kg bw (approximately 0.4, 1.0, 1.9, or 3.8 mg/kg bw) of chloroacetaldehyde. The compound was injected as a 0.5% aqueous solution into the peritoneal cavity three times a week for 12 weeks. Controls were injected with saline solution. Each group consisted of 12 rats, except for the highest-dose group, which consisted of 8 rats. In the highest-dose group, mortality was 5/8, while only two deaths occurred in other groups, one in the lowest-dose group and another in a mid-dose group. Weight gain was reduced significantly in the two highest-dose groups, and the reduction in weight gain was dose-related. Significant decreases in erythrocyte count and elevations of segmented neutrophils occurred in the two highest-dose groups, together with sig- nificantly increased hemoglobin at the highest dose. Clotting time was also prolonged in rats administered the highest dose. Organ-to-body-weight ratios of adrenals, brain, gonads, heart, kidneys, liver, lungs, and spleen were determined at the end of the experiment. At the highest dose, brain and liver weight ratios were markedly elevated, which might have reflected the marked reductions in weight gains. There were other statistically significant changes, seemingly scattered and not dose-related. Lungs of animals receiving the two highest doses had focal, chronic bronchopneumonia and changes in the respiratory tract suggested to be premalignant; however, these changes were not described. Similar changes in lower-dose groups were said to be of a lesser degree and not accom- panied by evidence of atypia of the respiratory epithelium. Sanders and coworkers (1982) investigated acute toxicity characteristics of chloral hydrate in male and female CD-1 mice. Mice were given the chloral hydrate by intragastric catheter, after which they were observed for 14 days and then necropsied. Urine and blood samples were taken at that time. LDsoS were 1.44 g/kg bw for males and 1.26 g/kg bw for females. Within 10 minutes of administration of low doses, mice became sedated, whereas at intermediate and high doses the mice became lethargic and lost the righting reflex. At the high doses, respiration was markedly depressed, and this depression was apparently responsible for the ensuing deaths. Gastric hyperemia was the only significant pathological change. Sanders and associates (1982) administered chloral hydrate to CD-1 mice in the drinking water for 90 days. A 14-day range-finding study had

Selected Disinfectants and By-Products 149 been conducted in male mice at doses of 14.4 and 144 mg/kg bw, which were 0.01 and 0.1 times the LDso. Blood counts and coagulation values were normal. Serum glutamic pyruvate transaminase (SGPT) and blood urea nitrogen (BUN) were normal, but lactic dehydrogenase (LDH) was depressed in mice given 144 mg/kg bw. The concentrations of chloral hydrate in the drinking water (0.07 and 0.7 mg of chloral hydrate/ml) in the 90-day study were estimated to yield the same doses as those used in the 14-day, range-finding study. The control group consisted of 260 mice/ sex, and each test group had 140/sex/group. Forty-eight mice of each sex of the controls and 32 mice of each sex in each treatment group were randomly selected for measurement of body weights and fluid consump- tion, recorded twice weekly. Average doses were calculated to be 18 and 173 mg/kg bw per day for female mice and 16 and 160 mg/kg bw per day for males. Males gained weight in a dose-related manner over the 90-day period; no effect on weights of females was apparent. Macroscopic lesions related to exposure were not seen. However, ratios of liver weights to body weights and absolute liver weights were increased in male mice. Lung weights of males were decreased, but the effect was significant only when expressed as a ratio to body weight and was not dose-related. Decreased RBCs in high-dose females, increased fibrinogen levels in high-dose males and low-dose females, and slightly increased activated partial thrombo- plastin times in both sexes were seen but were not statistically significant. Microsomal protein was increased 10% over control in females exposed to 0.7 mg/ml, and cytochrome b5 levels were increased 26% in males given 0.07 mg/ml and 40% in males at 0.7 mg/ml, contrasted with a 12% decrease in the cytochrome b5 levels in females on the higher dose. At least one dose induced an increase in aminopyrine N-demethylase and an increase in aniline hydroxylase in one or the other sex. Males receiving 0.7 mg/ml had decreased serum calcium and phosphorus and increased LDH and serum glutamic oxaloacetic transaminase (SGOT); BUN de- creased in both groups of male mice. The pattern was different in females: At 0.7 mg/ml there were increases in potassium, glucose, cholesterol, and phosphorus, with a decrease in nonprotein sulfhydryl. Only a few of these changes (potassium, cholesterol, and phosphorus in females and calcium, phosphorus, LDH, and SGOT in males) were significantly different from those of corresponding controls (p < 0.051. To the extent that a pattern is evident from these results, it seems that the liver was most affected by this repeated exposure to chloral hydrate, with males being more affected than females. These mice were also examined with respect to humoral and cell-me- diated immunity in an investigation described by Kauffmann and co- workers (19821. In the study of humoral immunity, IgM response to sheep erythrocytes was estimated. In the study of cell-mediated immunity, de

~ 50 DRINKING WATER AND HEALTH rayed-type hypersensitivity to sheep erythrocytes, lymphocyte respon- siveness, bone marrow DNA synthesis, and functional ability of the reticuloendothelial system were evaluated. Neither sex exhibited any change in cell-mediated immune status, and males did not demonstrate any al- teration in humoral immune function. However, the humoral immune function of female mice exposed to chloral hydrate for 90 days was de- pressed. There were fewer splenic antibody-forming cells produced against sheep erythrocytes, 36% fewer in females given 0.07 mg/ml and 40% fewer in those given 0.7 mg/ml. The effects of repeated doses of trichloroacetaldehyde on behavior were measured in a study by Kallman et al. (19841. Adult male CD-1 mice were administered doses of 14.4 or 144 mg/kg bw orally for 14 days. No effects on body weight, behavioral performance, locomotor activity, motor coordination, or swimming endurance were found 24-48 hours after the exposure was terminated. Concentrations of chloral of 0.70 or 0.07 mg/ml in drinking water (equivalent to 15.7 and 159.8 mg/kg bw per day) resulted in a slower weight gain in mice during 90 days as compared with that of controls (Kallman et al., 19841. Animals treated with the higher dose showed a decrease in body temperature at 45 days, and both groups showed a decrease at 91 days. No changes were observed in the behavioral tests used. Chronic Effects See "Carcinogenicity" below. Mutagenicity All the chloroacetaldehydes were found to be mutagenic in one or more strains of Salmonella typhimurium (Bignami et al., 1980a,b; Bruce and Heddle, 19791. In addition, all chloroacetaldehydes showed activity in a forward and back mutation system in S. coelicolor and in two forward mutation systems in Aspergillus nidulans (Bignami et al., 1980b). The mutagenic activity of the metabolites of vinyl chloride was assessed in the Ames TA 1535 test. Chloroacetaldehyde was found to be 1/190 as potent as chloroethylene oxide when the contact time for the bacteria with chloroethylene oxide was 6 minutes and that with chloro- acetaldehyde was 60 minutes (Perrard, 19851. Carcinogenicity No recent carcinogenicity data were found in the lit- erature. Prior references tested chloroacetaldehyde for carcinogenicity by repeated skin applications in mouse skin initiation-promotion studies, by repeated subcutaneous injections, and after intragastric feeding in mice and found it inactive (Van Duuren et al., 19791. The mouse skin bioassays were performed on 30 female Ha:ICR Swiss mice and used a dose of 1.0 mg/application/mouse three times weekly for 581 days. Benign lung pap

Selected Disinfectants and By-Products 151 illomas were observed in 14 mice, and a squamous cell carcinoma was found in 1 mouse. Subcutaneous injections of 0.25 mg/injection/mouse in 30 mice showed 1 mouse with local sarcomas after 630 days on test. The same dose given to 30 males and 30 females resulted in 3 males and 1 female with forestomach tumors after 630 to 636 days on test. The incidence of tumors was not considered significantly different than con- trols. A related compound, 2-chloropropanal, induced the formation of in- jection-site sarcomas in female Swiss mice after subcutaneous injection and of tumors of the stomach in both male and female mice after intra- gastric delivery (Van Duuren et al., 19791. Four of 30 female mice injected with vehicle (trioctanoin) and none of 30 injected with water on similar schedules developed such tumors. Three of 30 male mice, one of which had a squamous carcinoma of the forestomach, and 6 of 30 female mice had tumors of the stomach after Savage once per week of 1 mg of 2- chloropropanal for about 82 weeks; whereas none of the 30 male or 30 female mice Savaged on similar schedules with the vehicle (trioctanoin) developed tumors of the stomach. Because chloroacetaldehyde is a metabolite of the carcinogen vinyl chloride, it was tested for carcinogenic activity along with chloroethylene oxide in an initiation-promotion mouse skin experiment (Zajdela et al., 1980~. Groups of 20 to 28 mice were given a single dose on the skin of 0.05, 0.1, 1.0, or 2.5 mg of chloroacetaldehyde dissolved in 80 Al of acetone followed by applications three times a week of 12-O-n-tetra- decanoylphorbol- 13-acetate for 42 weeks. Chloroethylene oxide produced skin tumors, but chloroacetaldehyde did not. In vitro experiments have shown that chloroacetaldehyde can react with calf-thymus DNA to form products that are found in liver DNA isolated from rats exposed to 250 ppm vinyl chloride in their drinking water for 2 years (Green and Hathway, 19781. Etheno-deoxyadenosine and etheno- deoxycytidine were identified in the enzyme hydrolysates from these two experiments. In addition, chloroacetaldehyde has been identified as an important alkylating agent when purified rat P450 systems were examined (Guengerich et al., 1979~. Whether chloroacetaldehyde reaches the sen- sitive tissues to react with them is unknown. Teratogenici?;y In the study by Kallman et al. (1984), mice were peri- natally exposed to 21.3 and 204.8 mg/kg bw per day of trichloroacetal- dehyde in the drinking water. The pups had normal body weight, normal development of the neurobehavioral reflexes, and normal motor coordi- nation. Retention of passive avoidance learning was reduced at 1 and 24 hours among mice exposed to the high dose, but the significance of this

~ 52 DRINKING WATER AND HEALTH finding in the absence of other findings was unclear. No reproductive or teratogenic studies with other haloaldehydes were found. Other Related Substances Small amounts of several aldehydes have been identified in drinking water but will not be reviewed in this report. One of the most common is acrolein, known to many as an irritant com- pound found in many kitchens, especially during the heating of lipids as in the frying of food. Investigations of the toxic features of acrolein have been critically reviewed by Beauchamp and associates (1985) and by the NRC (1981, pp. 234-2414. Beauchamp et al. (1985) pointed out the strongly irritant properties and reviewed the absorption of the inhaled acrolein and its metabolism, as well as many experimental investigations, mostly of its toxic effects on the respiratory tract. They commented on a bioassay of possible carcin- ogenicity, then under way, involving administration of acrolein in drinking water (625 ppm) to Fischer 344 rats; there had been a seemingly significant increase in adenomas and of neoplastic nodules in the adrenal cortex of female rats. Conclusions will undoubtedly await completion of the study and a thorough evaluation of the data. In a 52-week inhalation study of Syrian golden hamsters, acrolein was not found to be carcinogenic (Feron and Kruysse, 19771. A group of halogenated acroleins (propenals) are formed in the chlo- rination of water containing humic acids and wood pulp. Some of these chemicals are potent mutagens (Meter et al., 19851. There are little other data with respect to their potential health hazards. They are potentially important disinfection by-products that require further toxicological char- acterization. Haloaldehydes are derivatives of acetaldehyde, a material that has been extensively studied because it is thought to be the toxic metabolite of ethanol. Liver damage, central nervous system (CNS) toxicity, and fetal alcohol syndrome are well-recognized effects of alcohol abuse. One hypothesis to explain liver toxicity is that hepatic metabolism forms the active intermediate, acetaldehyde, which covalently binds to hepato- cellular macromolecules (especially proteins) and alters hepatocellular structure and function. This binding results in injury and eventually to irreversible disease (Tuma and Sorrell, 1985~. The electrophilic nature of the carbonyl carbon of acetaldehyde is susceptible to attack by a variety of nucleophilic agents including amino acids, peptides, proteins, lipids, nucleic acids, and other endogenous materials. The reaction product be tween proteins and acetaldehyde, a Schiff base, is chemically unstable, reforming the protein and aldehyde. The Schiff base is probably converted to more stable products by reduction with ascorbic acid. These products have been found in vitro with liver homogenates and liver slices, and the

Selected Disinfectants and By-Products 153 amount of adduct found is reduced by increased concentrations of cysteine and glutathione. Adducts with acetaldehyde have also been found with phospholipids in rat liver microsome preparations (Kenney, 1984) and with tetrahydrofolate in liver in vitro incubations (LaBaume and Guynn, 1985). A recent paper described an alteration in acetaldehyde metabolism after carbon tetrachloride treatment and suggested that toxicant-toxicant inter- actions should also be considered in a description of aldehyde toxicity (Yuki et al., 19844. Specifically, rats were treated with carbon tetrachloride at 2.08 M/kg bw intraperitoneally twice a week for 8 to 12 weeks. The changes in acetaldehyde metabolism were examined in a perfused liver system. The researchers found that the activity of aldehyde dehydrogenase in hepatic mitochondria was decreased by about 75%, that the mitochon- drial nicotinamide adenine dinucleotide (NADH) oxidation was reduced by approximately 35% of the control level, that the basal level of hepatic oxygen uptake was reduced by 50% and that this was decreased further by acetaldehyde, and that hepatic acetaldehyde metabolism was decreased. The findings described for acetaldehyde should be included in consid- erations of the toxicity of haloacetaldehydes because the reactions de- scribed should be the same. This is particularly true of adduct formation, in which the presence of the halogen groups should make the initial nucleophilic attack even more favorable and rapid. Whether successive reactions will proceed with haloacetaldehydes must be examined on an individual basis and should probably be more carefully researched. SUMMARY AND CONCLUSIONS Haloaldehydes are rapidly metabolized to alcohols and acids, and they probably react with proteins, lipids, and nucleic acids like acetaldehyde. If sites of unsaturation are present as in chloropropenals, there will be additional reactions with sulfhydryl-containing proteins like glutathione in a manner analogous to the metabolism of acrolein. The haloaldehydes are one of the intermediates found in the metabolism of the haloethylenes, compounds that have been identified as carcinogens. Toxicity data are lacking for these compounds. The acute oral LDso values for the chloroacetaldehydes in rats are approximately 100 to 300 mg/kg bw; and effects on weight gain, organ-to-body-weight ratios, and lung pathology were found after 12 weeks of study with doses in the approximate range of 2-5 mg/kg bw of monochloroacetaldehyde. Tri- chloroacetaldehyde showed liver effects and body weight changes in mice after 90 days of 16-18 mg/kg bw. These compounds are mutagens, but very little work has been done to determine their carcinogenic activity. Teratology data are also lacking.

1 54 OR ~ N K! NG WATER AN D H EALTH The relative importance of the haloaldehydes as a hazard in drinking water cannot be evaluated. There are insufficient data to assess their toxicity, and there are no good measures of their concentration in drinking water because they are difficult to analyze. They are direct-acting mutagens and are postulated intermediates in the metabolism of chloroethylenes, compounds that are known or suspected carcinogens. The compounds are important for future study. There are not sufficient data available for the calculation of suggested no-adverse-effect levels (SNARLS) for the haloaldehydes described in this section. HALOKETON ES I, I, I-TrichIoroacetone CAS No. 918-00-3 I, i,3,3-TetrachIoroacelone CASNo.632-21-3 HexachIoroacetone CAS No. I 16-16-5 C13CCOCH3 Cl2CHCOCHCl2 Cl3CCOCCl3 The haloketones are principally found as chemical intermediates in industrial processes. A summary of their physical and chemical properties is shown in Table 4-12. METABOLISM No data are available. HEALTH ASPECTS Observations in Humans No data are available.

Selected Disinfectants and By-Products 155 TABLE 4-12 Physical and Chemical Properties of Selected Haloketonesa Boiling Melting Molecular Point, Point, Specific Compound Weight °C °C Gravity Solubilityb 1,1-Dichloro- 127 120 NA 1.31 2, 3 acetone 1,1,1-Tr~chloro- 161 149 NA 1.44 2, 3 acetone 1,1,3,3-Tetra- 196 180 NA NA 2, 3, 4, 5 chloroacetone Pentachloro- 230 192 2.1 1.69 NA 2-acetone Hexachloro- 265 202 - 2 1.44 5 2-acetone 2-butanone 141 166 NA 1.31 2, 3, 4, 5 Modified from Weast (1983), pp. C-74, C-75, and C-193. bin medium 1, water; 2, alcohol; 3, ether; 4, acetone; 5, benzene; NA, not available. Observations in Other Species Acute Elects Borzelleca and Lester (1965) determined the oral LDso in male Wistar rats for hexachloroacetone to be 1,550 mg/kg bw. They also determined the dermal LD50 in male albino rabbits to be 2,980 ma/ kg bw and the LCso in albino rats to be 660 ppm for a 3-hour and 360 ppm for a 6-hour inhalation exposure. Examination of the lungs imme- diately after death showed extensive hemorrhage. Those animals that sur- vived showed edema, hemorrhagic changes, and congestion in the lungs 15 days after exposure. Chronic Effects No data are available. Mutagenicity A comprehensive investigation of the mutagenicity in Salmonella bacteria induced by four haloketones (1,1,3-trichloroacetones, 1,1,3,3-tetrachloroacetones, pentachloroacetones, and hexachloroace- tones) was performed by Nestmann et al. (1985) in a study that confirmed the initial report by Zochlinski and Mower (1981) concerning solvent effects in tests with hexachloroacetone. Trichloroacetone-induced dose- relatedincreases of mutants were found in strains TA1535, TA97, TA98, and TA100, but not in strains TA1537 or TA1538; addition of an S9 activation mixture had no activating effect. Tetrachloroacetone was mu- tagenic in strains TA98 and TA100, with S9 enhancing the effects in TA98 only; negative results were found with strains TA1535, TA1537, TA1538,

156 DRINKING WATER AND HEALTH and TA97. Pentachloroacetone was mutagenic in strains (TA97, TA98, and TA100) containing plasmid pKM101 but was nonmutagenic in the strains without the plasmid (TA1535, TA1537, and TA15381; S9 slightly increased the effects in strains TA97 and TA98 but not in strain TA100. Hexachloroacetone was mutagenic in strains TA98 and TA100 and was shown previously to be nonmutagenic in strains TA1535, TA1537, and TA1538 (Zochlinski and Mower, 19811. Nestmann et al. (1985) dissolved the haloketones in either acetone or dimethyl sulfoxide (DMSO), and all compounds except for hexachloroacetone, which was negative when dis- solved in acetone, were mutagenic in either solvent. Hexachloroacetone and DMSO interacted, resulting in a chemical reaction and the formation of a mutagen more potent than was hexachloroacetone, a liquid, by itself. This and certain previous studies on solvent effects and interactions led Nestmann et al. (1985) to propose recommendations for choosing solvents and performing tests to prevent artificial results from appearing in the literature. Carcinogenicity No data are available. Reproductive E~ects No data are available. CONCLUSIONS AND RECOMMENDATIONS There are few toxicological data available for the haloketones. Never- theless, trichloroacetone, tetrachloroacetone, pentachloroacetone, and hexachloroacetone were shown to be mutagenic in Salmonella bacteria. There are insufficient data available for the committee to estimate sug- gested no-adverse-effect levels (SNARLs) or to perform other r~sk as sessments. HALOACETON ITRI LES Dich loroaceton itri le CAS No. 3018-12-0 Trich loroacet~n itri le. Trich loromethyl n itri le. Tritox, TrichIoromethyl cyanide CAS No. 545-06-2 CHC12CN CC13CN

Selected Disinfectants and By-Products 157 Bromoch loroaceton itri le CAS No. 83463-62-] D i b r o m o a c e t o n i ~ r i ~ e CAS No. 3253-43-5 CHBrClCN CHBr2CN The haloacetonitnles are colorless to yellow, volatile liquids. The chlo- nnated acetonitnles are used as insecticides and fungicides. A summary of their chemical and physical properties is shown in Table 4-13. METABOLISM Data from Pereira et al. (1984) suggest that absorption of the halo- acetonitriles (HANs) from the gastrointestinal tract is limited. No other absorption data were found. Roby et al. (1986) studied the pharmacokinetics of dichloroacetonitnle in male Fischer 344 rats and male B6C3F~ mice. Aqueous solutions of t2-~4C]dichloroacetoni~ile (DCAN) were administered at 0.2, 2.0, arid 15 mung bw; He highest dose level represented 5-10% of He oral LDso of 2 rnM/kg bw determined for rats. The rats eliminated 84% of He administered dose in 48 hours; total recoveries from tissue and excrete were greater Han 95%. Tissues contained about 15% of the amount administered after 48 hours, win He liver (5%), muscle (4%), and blood (3%) retaining most of He dose. In another experiment, rats Hat received 0.1% of the LDso of TABLE 4-13 Chemical and Physical Properties of Haloacetonitriles Dibromo Dichloro- Trichloro- Bromochloro- acetonitrile acetonitrile acetonitrile acetonitrile (DBAN) CHC12CN CC13CN CHBrClCN CHBr2CN Molecular weight 109.94 144.39 154.4 198.9 Appearance Liquid Colorless, Liquid Liquid volatile liquid Density (g/ml) 1.37 1.44 1.68 2.30 Melting point (°C) a -42 a a Boiling point (°C) 112.3 85.7 125-130 67-69 aNot available.

]58 DRINKING WATER AND HEALTH METABOLIC FATE OF HALOACETONITRILES 1. X H-C-CN H \MFO 2. ~-~-we 1 H \MFO X \ X-C-CN X OH H-C-C19 H OH X-C-CN- - H O 1 OH 1 X-C-CN 1 X O 11 -X-C-CN - o 11 H-C-H + CN o 11 X-C-H + CN 11 C-CN - CO + CN o 11 X-C-X + CN o 11 HO-C-CN - CO2 + CN FIGURE 4-1 Proposed metabolic pathways for cyanide release from haloacetonitr~les. MFO is mixed function oxidase. X is ClBr. From Pereira et al. (1984) with permission. t1-14C] DCAN by gavage took 6 days to eliminate 70% of the dose. Total recoveries in tissues and excrete were more than 89%. Urinary excretion (32~o) was similar to that found in the 2-position-labeled compound, but He amount expired as CO2 was much lower (9%~. Tissues contained about 19% of the dose after 6 days, with blood (7%), muscle (4%), skin (3%), and liver (2%) containing most of the retained dose. These findings support that the C-C bond of DCAN seems to be split during biotransformation. Pereira et al. (1984) reported that 7.7% (DB AN) to 12.8% (B CAN) of orally administered single doses of DCAN (82 mg/kg bw), DBAN (149 mg/kg bw), B CAN (115.8 mg/kg bw), and TCAN (108 mg/kg bw) to male Sprague-Dawley rats was converted to thiocyanate and excreted in  the urine and that some HANs inhibited hepatic manganese demethylase activity. Levels of HANs or metabolites in feces and exhaled air were not reported. They concluded that haloacetonitriles are converted to highly toxic metabolites. A proposed metabolic pathway appears in Figure 4-1. Daniel et al. (1983) reported in an abstract that haloacetonitriles are direct- acting alkylating agents, and this could be the mechanism for organ toxicity and carcinogenicity. Bull (1982), Meter et al. (1983), and Simmon et al ~ 1977) all suggested that the haloacetonitriles may be biologically reactive acting directly or by conversion to toxic products. HEALTH EFFECTS Observations in Humans Data are not available.

Selected Disinfectants and By-Products 159 Observations in Other Species Acute Elects Hayes et al. (1986) reported the following oral LDsoS for DCAN in male CD-1 mice, 270 mg/kg bw; in CD-1 females, 279 mg/kg bw; in male Charles River CD rats, 339 mg/kg bw; and in Charles River CD females, 330 mg/kg bw. Death was preceded by slowed res- piration, decreased activity, ataxia, and coma. No consistent compound- related gross pathological effects were noted at necropsy. Hayes et al. (1986) also reported the following oral LDsoS for DBAN: in male CD-1 mice, 289 mg/kg bw; in CD-1 females, 303 mg/kg bw; in male Charles River CD rats, 245 mg/kg bw; and in Charles River CD females, 361 mg/kg bw. Death was preceded by slowed respiration, ataxia, depressed activity, and coma. No consistent compound-related gross pathological effects were noted at necropsy. Meter et al. (1985) reported an oral LDso of 295 mg/kg bw for rats. Subchronic Effects Hayes and associates (1986) administered a so lution of DCAN in corn oil to male and female CD rats by gavage at doses of 12, 23, 45, or 90 mg/kg bw per day for 14 days or at 8, 33, or 65 mg/kg bw per day for 90 days. The compound-related mortality at 65 mg/kg bw was 50% of males and 25% of females; at 33 mg/kg bw it was 10% of males and 5% of females at 8 mg/kg bw. Two percent of males had died. The apparent NOEL was determined to be 8 mg/kg bw per day. There were no deaths in either study. Administration of 65 mg/kg bw per day caused a depression of weight gain; reduction of weights and ratios of liver, spleen, thymus, lungs, kidneys, and gonads in the male rats; and a depression of weight gain and reduction in weight and ratio of the spleen in female rats. Similar findings were observed at 90 mg/kg bw per day after 14 days of dosing. Cholesterol levels were significantly lower at 65 and 90 mg/kg bw per day. There were no significant or consistent com- pound-related effects on formed elements in blood; on serum calcium, chloride, phosphorous, glutamic pyruvic transaminase (SGPT), glutamic oxaloacetic transaminase (SGOT), alkaline phosphatase, 5'-nucleotidase, protein (albumin, globulin), or creatinine; on blood urea nitrogen; on urinary pH, protein, glucose, ketone, or blood; or on macroscopic anat omy. Hayes and associates (1986) also administered DBAN in corn oil to male and female CD rats by gavage at doses of 23, 45, 90, or 180 ma/ kg bw per day for 14 days or 6, 23, or 45 mg/kg bw per day for 90 days. All the animals that received 180 mg/kg bw per day died within 7 days. At 90 mg/kg bw per day, 40% of males and 20% of females were dead by day 14. The following were reported for the 14-day repeated dosing study: a dose-dependent depression of weight gain; decreases in serum

}60 DRINKING WATER AND HEALTH proteins albumin and globulin, alkaline phosphatase (ALP), 5'-nucleoti- dase, glucose, and phosphorous; and increases in cholesterol and blood urea nitrogen (BUN)/creatinine ratio at 90 mg/kg bw per day only; decrease in weight and ratio of spleen and thymus in males at 90 mg/kg bw per day and increase in liver and decrease in lung and thymus weight and ratios in females at the 90 mg/kg bw per day dose only. The reported no- observed-affect level was 23 mg/kg bw per day in the 14-day study. The following were noted after 90 days of dosing: dose-dependent depression of body weight gain in males; decreased thymus weight and ratios in both males and females at 45 mg/kg bw per day. Compound-related mortality was 5% for males at 45 mg/kg bw and 0% for males and 10% for females at 23 mg/kg bw. There were no significant and consistent compound- related effects on formed elements in blood; on serum calcium chloride, phosphorus, SGPT, SOOT, alkaline phosphatase, 5'-nucleotidase, protein (albumin, globulin), or creatinine; on blood urea nitrogen; on urinary pH, protein, glucose, ketone, or blood; or on macroscopic anatomy. Reproductive Elects Smith and coworkers (1986) studied reproduc- tive effects of HANs. Pregnant female Long-Evans rats were administered by gavage a single high dose, the maximum tolerated dose (MID, 50- 55 mg/kg bw) during gestation, and litters were born normally. Evaluation of offspring for weight and number was performed on days 1 and 4. Maternally toxic doses of HANs (55 mg/kg bw) were fetotoxic, as indi- cated by reduced weights of pups at birth and by postnatal growth. This toxicity study compared chloroacetonitrile (CAN), DCAN, and TCAN as well as other HANs, and it was concluded that the toxicity was increased with increasing substitution of the alpha carbon. The investigators cited the findings of Pereira et al. (1984) on the reduction in urinary excretion of cyanide with increasing chloride substitution of acetonitrile. Mutagenicity Bull et al. (1985) evaluated the mutagenic and carcin- ogenic properties of CAN, DCAN, TCAN, BCAN, and DBAN. They reported that DCAN and B CAN were direct-acting mutagens in Salmo- nella; all five HANs-induced sister chromatic exchanges in Chinese ham- ster ovary cells in vitro; all five compounds were without activity in the mouse micronucleus test; DCAN, DBAN, and CAN applied topically initiated tumors in mouse skin. DCAN was reported by Valencia et al. (1985) to induce sex-linked recessive lethals in Drosophila. The activity of DCAN, DBAN, TCAN, and BCAN in the mouse sperm-head abnor- mality assay was evaluated by Meier et al. (19851. All the compounds were negative in this test. Meier et al. (1983) and Bull (1982) reported

Selected Disinfectants and By-Products 161 that DCAN was positive in Salmonella tester strains TA98, TA100, and TA1535. Kraybill (1980, 1983) identified DCAN as a mutagen or sus- pected mutagen in chlorinated drinking water in the United States. Simmon et al. (1977) reported that DCAN was positive in the Salmonella TA100 assay but was negative in Saccharomyces cerevisiae. These data suggest that DCAN is mutagenic and possibly carcinogenic. Carcinogenicity Bull and Robinson (1985) evaluated the oncogenic effect (lung tumors) of DCAN, DBAN, TCAN, and B CAN in female A/J mice. The HANs were administered orally, three times weekly for 8 weeks and the animals sacrificed at 9 months. TCAN and B CAN signif- icantly increased the incidence of lung tumors; DCAN and DBAN elicited marginal increases in incidence (not statistically significant). The authors assessed the ability of DCAN, DBAN, TCAN, or B CAN to act as tumor initiators in Sencar mice. The animals received six topical applications over a 2-week period at total doses of 1,200, 2,400, or 4,800 mg/kg bw. 12-O-Tetra-decanoylphorbol-13-acetate (TPA) was then applied three times per week for 20 weeks. Mice were observed for 1 year and sacrificed. Squamous-cell tumors of skin were increased significantly only in the B CAN group that received 4,800 mg/kg bw and in the groups given DBAN at 1,200 and 2,400 mg/kg bw. CONCLUSIONS AND RECOMMENDATIONS There are few toxicological data available in the HANs. However, it has been shown that DCAN and DBAN are absorbed from the gastroin- testinal tract and are biotransformed to thiocyanate and excreted in the urine, and small percentages of administered DCAN and DBAN are stored in tissues. The committee noted that available evidence suggests that DCAN is mutagenic and potentially carcinogenic. The conservative no- observed-effect level (NOEL) in rats exposed by Savage for 90 consecutive days is 8 mg/kg bw per day. Available evidence does not support the mutagenicity or carcinogenicity of DBAN, although DBAN was positive in a Sencar mouse-skin initiation model. The NOEL in rats exposed (by Savage) for 90 consecutive days is 23 mg/kg bw per day. A suggested no-adverse-effect level (SNARL) for DCAN may be cal- culated based on the NOEL of 8 mg/kg bw per day assuming that a 70- kg human consumes 2 liters of water daily, which contributes 20% of the total intake: 8 mg/kg low/day x 70 kg x 0.2 1,000 x 2 liters _ 0.056 mg/liter, or 56 1lg/liter.

i62 DRINKING WATER AND HEALTH Even using a conservative 1,000 uncertainty factor, the committee does not recommend this SNARL because of the concern that DCAN may be carcinogenic. The committee recommends that further data be developed. A SNARL for DBAN may also be calculated based on the NOEL of 23 mg/kg bw per day using the same assumptions as above and a con- servative uncertainty factor of 1,000 because of the general lack of tox- icological data and the positive finding in the Sencar mouse-skin initiation model for this substance: 23 mg/kg low/day x 70 kg x 0.2 = 1,000 x 2 liters 0. 161 mg/liter, or 161 1lg/liter. A SNARL may be estimated for a 10-kg child consuming 1 liter of water daily: 23 mg/kg low/day x 10 kg x 0.2 0.046 mg/liter, or = 1,000 x 1 liter 46 ~g/liter. The committee recommends these two SNARLs for DBAN but also notes that further toxicological data should be developed. CHLOROPICRIN N itrotrich Ioromethane, TrichIoron itromethane, N itrochIoroform CAS No. 76-06-2 CC13NO2 Chloropicrin is a slightly oily, colorless, refractive liquid that is rela- tively stable, nonflammable, and is not decomposed by water or mineral acids. It has a boiling point of 112°C and a freezing point of-69.2°C; its specific gravity is 1.692 at 0°C; and it is soluble in alcohol, benzene, ether, and carbon disulfide and slightly soluble in water (0.17 g/100 g water at 18°CJ. A strong irritant that is toxic when ingested or inhaled, chloropicrin is used in organic synthesis, dyestuffs (crystal violet), fu- migants, fungicides, insecticides, rat extermination, and tear gas. The American Conference of Governmental Industrial Hygienists has set 0.1 ppm (~0.7 mg/m3) as the threshold limit value (TLV) and 0.3 ppm (~2.0 mg/m3) as the short-term exposure limit (STEL; 15 minutes) for chloro- picrin in air (ACGIH, 19861.

Selected Disinfectants and By-Products 163 CHEMISTRY AND ENVIRONMENTAL FATE In 1980, de Greef et al. reported that tap water from 13 major water supplies of The Netherlands contained chloropicrin in concentrations of about 0.01 ~g/liter, whereas tap water from 7 water supplies that had undergone chlorination contained up to 3.0 ~g/liter. Sayato et al. (1982) in Japan and Bruchet et al. (1985) in France also reported chloropicrin in surface waters that had been chlorinated. Duguet et al. (1985), using waters from several French water systems, found that the presence of nitrate in water increases only slightly (3.4%) the production of chloropicrin during chlorination. In contrast, the addition of nitrite to the water increased the production of chloropicrin by more than 3,300 times during chlorination. The production of chloropicrin during chlorination passed through a max- imum at a concentration of 100 mg of nitrite/liter in the water; with this concentration of nitrite the production of chloropicrin was nearly 3,800 times that measured in the original water. A concentration of 200 mg of nitrite/liter added to the water, however, resulted in the production of less chloropicrin following chlorination than in the original water. Duguet et al. (1985) reported that the initial reaction between nitrite and chlorine in water loaded with natural organic matter is very rapid, with nearly 51% of the chloropicrin produced during a contact time of 120 hours being formed during the first 46 minutes. The presence of organic matter in the water had an important modifying action on the production of chloropicrin. When nitrite in a dilution of 20 mg/liter was added to a sample of water with a low content of organic matter and to another sample of the same water to which fulvic acid in a dilution of 50 mg/liter had also been added, and both samples were chlorinated for 120 hours, the latter sample contained 15.6 times as much chloropicrin as the former. When ammonia was present in water, the production of chloro- picrin during chlorination was found to depend on the ratio of chlorine to nitrogen, the ratio for maximal production of chloropicrin in such water being 8. Chloropicrin in water hydrolyzes very slowly: after 10 days in darkness no detectable loss of chloropicrin had occurred (Castro and Belser, 19811. These investigators found that exposure of the contaminated water to sunlight or to light from an incandescent lamp for 72 hours resulted in loss of nearly one-half of the chloropicrin. However, exposure of the mixture to ultraviolet light for about 9 hours resulted in loss of more than 95% of the chloropicrin. The stoichiometry was represented as follows: he C13CNO2 H O 3Cl- + NO3- + 4H+ + CO2.

]64 DRINKING WATER AND HEALTH In the vapor phase, Moilanen et al. (1978) found that chloropicrin is quite stable in the dark. During 70 days of storage, 7.9% of the chloropicrin was lost. When the vapor was exposed to radiation above 290 ~m, 75.2% of the chloropicrin was lost during 70 days. The products of degradation both in the dark and in the light were COC12 (phosgene), NOC1, NOX, and C12. The formation of phosgene, a potent lung irritant that may induce pulmonary edema, provides a secondary hazard of exposure to vaporized chloropicrin that has existed in a closed space for a protracted period of time. Castro et al. (1983) reported the isolation from soil of four species of Pseudomonas that were capable of degrading chloropicrin: C13CNO2 ~ C12CHNO2 ~ ClCH2NO2 ~ CH3NO2. In the degradation, about 4% of the chloropicrin is degraded completely to CO2 METABOLISM Castro et al. (1985) found that a preparation of cytochrome P450 prepared from Pseudomonas putida reacted with chloropicrin in solution approximately as well as did the intact organism and with essentially the same stoichiometry. Each step in the sequence of dechlorination reactions required two hemes. Ghosal et al. (1985) pointed out that the genes for many degradative enzymes active against halogenated compounds are associated with bacterial plasmids and are clustered. This allows cloning of the gene clusters in broad-host- range vectors for dissemination to a large number of gram-negative bacteria, as has been done for 2,4-dichlorophenoxyacetic acid (2,4-D) and 2,4,5- trichlorophenoxyacetic acid (2,4,5-T). The researchers raised the possibility of cloning the gene clusters responsible for the organism's degradative abil- ities. They suggested that the naturally existing population of organisms capable of degrading chloropicrin could be increased by creating new or . . . . gamsms Wlt ~ t MS CapaClty. Although no actual study of metabolism of chloropicrin by mammals has been found, the report by Castro et al. (1985) summarized above suggests that the direct metabolic products, produced primarily in the liver, probably would be the same as those produced by P. putida. Modification of some of these direct products may occur through other enzyme systems. Also, the direct products and any secondary derivations may modify nor- mal enzymatic and other functions in such a way as to alter the simple degradative chain of reactions that is now known. Therefore, study of the catabolism of chloropicrin in mammalian organisms seems to be of con- siderable importance to understanding and accurately assessing the hazard of exposure to this substance.

Selected Disinfectants and By-Products 165 HEALTH EFFECTS Observations in Humans Little information on human health effects has been reported. Flury and Zernik (1931), using data accumulated during World War I, when chlo- ropicrin was used as a toxic antipersonnel agent, reported that a concen- tration of 0.3 to 3.7 ppm of chloropicrin in air induced closing of the eyelids, and a concentration of 4 ppm incapacitated men sufficiently to render them unfit for combat. Sufficient irritation of nasal and tracheo- bronchial mucosa to be unbearable for longer than 1 minute resulted from exposure to 15 ppm, and that concentration could induce damage within the respiratory system. In addition, Prentiss (1937) reported that 0.3 ppm induced lachrimation and that exposures to concentrations of 119 ppm for 30 minutes or of 298 ppm for 10 minutes were lethal. Kvetensky et al. (1979) described a mass poisoning by chloropicrin without documenting the number of persons involved. Irritation of the trachea and bronchi resulted in frequent coughing (up to 35 times per minute) and feelings of nausea. Blood pressure initially remained reason- ably steady at 110/80, but the pulse rate rose to 110 per minute. Later, as pulmonary edema developed, the blood pressure fell to shock levels in some patients. For treatment of severe intoxication by chloropicrin, these investigators recommended cardiotonics, forced inhalation of 5 to 7 liters per minute of 50% oxygen that had been bubbled through 95% ethanol, infusion of hypertonic glucose (40%) that may contain calcium gluconate, codeine, vitamin C, aminophyllin-like drugs, furosemide, and corticoids. Exposed individuals may become sensitized to chloropicrin, so that sub- sequent exposures may induce more serious changes than the initial one. Pitt (1982) has devised a "vapor hazard index" to attempt to express the hazard implicit in working with volatile chemicals that may be released accidentally into a workplace. The index was defined as the quotient of the concentration of a saturated vapor divided by 1,000 times the TLV for that chemical. Chloropicrin has an index of 260. For comparison, methylhydrazine has an index of 240, and the index for 1,1-dimethyl- hydrazine is 270. Obviously, a change in the value of the TLV for a chemical will result in a change in the vapor hazard index. The value of this index for safeguarding human health remains to be established. Observations in Other Species Acute Elects From data provided by the Dow Chemical Company in 1972, Tatken and Lewis (1983, p. 687) determined the oral LDso of

~ 66 OR ~ N K! NG WATER AN D H EALTH chloropicrin to be 250 mg/kg bw in the rat. Stokinger (1982, pp. 4,164- 4,166) reported the acute lethality of chloropicrin in mice, cats, and dogs from inhalation exposures. In mice, 25 ppm for 15 minutes was not lethal, while 50 ppm for 15 minutes resulted in death within 10 days. Mice exposed to 125 ppm for 15 minutes died within 1 day. In cats, doses of 38-76 ppm for 20-25 minutes were lethal doses in cats. Death occurred in 1-12 days. Dogs exposed to 48 ppm for 15 minutes tolerated the exposure, 43% of those exposed to 117-140 ppm for 30 minutes died, and those exposed to 155 ppm for 12 minutes became ill, but the mortality experience was not reported. Kane et al. (1979) exposed rats to chloropicrin at various concentrations to determine the concentration that increased significantly the rate of breathing of one-half the group of animals. This was designated the RDso and was found to be 7.98 ppm with 95% confidence limits of 6.22 to 10.6 ppm. The effect of various concentrations of chloropicrin on the rate of breathing was found to be represented by the expression y = 9.54 + 44.87 log x, where y is the percentage of decrease in the rate of breathing and x is the concentration of chloropicrin in air expressed in parts per million. Subchronic Effects Buckley et al. (1984) reported exposing mice for 6 hours/day during 5 days to chloropicrin at 7.98 ppm. This exposure resulted in exfoliation, erosion, ulceration, and necrosis of the respiratory epithelium, in ulceration and necrosis of the olfactory epithelium, espe- cially in the dorsal meatus of the nostrils (of marked severity), and in inflammation and squamous metaplasia of the respiratory epithelium (of moderate severity). In the lung, chloropicrin induced serious exudation and severe fibrosing peribronchitis and peribronchiolitis. In a study un- dertaken for the National Cancer Institute (NCI, 1978), the mean daily dose of chloropicrin that resulted in no significant decrease in growth of male and female mice or in their probability of survival was 32 mg/kg bw. Mice were given chloropicrin by Savage 5 days/week at daily doses of 25 mg/kg bw for 13 weeks and 35 mg/kg bw for 65 weeks. Chronic Elects Male and female B6C3F~ mice and Osborne-Mendel rats were exposed to mean daily doses of chloropicrin at 24.5 and 25.7 mg/kg bw (male rats), 20 . 3 and 2 1 . 5 mg/kg bw (female rats), and 3 3 . 3 and 66.7 mg/kg bw (male and female mice) gavage for 5 days/week during 78 weeks. The exposures resulted in dose-related loss of weight by both sexes of both species and in significant and dose-related mortality among rats. The mice given the larger mean daily dose had significantly decreased probabilities of survival (NCI, 19781.

Selected Disinfectants and By-Products 167 Carcinogenicity In the study performed for the National Cancer In- stitute (NCI, 1978) and outlined above under the heading "Chronic Ef- fects," it was found that the incidence of tumors in rats given chloropicrin was lower than that in the control rats. When the summed incidences of tumors among the control animals and among those given chloropicrin were adjusted to the same-sized populations, the control rats had 16, whereas the treated rats had only 5, of the same types of tumors: chromo- phobe adenoma and mammary gland adenocarcinoma or fibroadenoma. In mice, after the same sort of adjustment as for the rats, the control animals had 10.2, versus 7 in the treated mice, of these types of tumors: alveolar or bronchiolar adenoma and hepatocellular carcinoma. Of possible significance are the findings that 1 of 48 female mice given the low dose of chloropicrin had a squamous-cell papilloma of the stomach and that 2 of 48 male mice given the high dose of chloropicrin had squamous-cell carcinomas of the stomach. No such lesions of the stomach were found in control mice. Despite that fact, the report concluded that the occurrence of these tumors of the stomach was not significant under the Bonferroni criterion. In 1980, 2 years after publication of the results of the study performed for the National Cancer Institute, Griesemer and Cueto (1980) placed chloropicrin in a group of "chemicals with equivocal evidence for car- cinogenicity in animals" on the basis of the tumors of the stomach found in mice given chloropicrin in that study. Mutagenicity Morlya et al. (1983) tested chloropicrin as a mutagen in six bacteria: one strain of Escherichia cold (WPZ her) and five strains of Salmonella typhimurium (TA98, TA100, TA1535, TA1537, TA15381. Chloropicrin was without effect in TA1535, TA1537, and TA1538. Chlo- ropicrin induced dose-related reversions in WPZ her and TABS, but they were not numerous enough to satisfy the investigators' criterion for positive mutagenic action of an excess of 100 beyond the number of spontaneous revertants. Chloropicrin was positively mutagenic for TA100 when it was applied with S9 activation. Shirasu et al. (1982, 1984) reported that chlo- ropicrin is an indirect mutagen on TA100 with S9 activation when applied along with the S9 mix. Valencia et al. (1985) examined the mutagenic potential of chloropicrin on Drosophila, applying the chemical either by ingestion or by intracoe- lomic injection. The investigators concluded that chloropicrin given by feeding is questionably inductive of sex-linked recessive lethal mutations in the fruit fly but is not mutagenic at all by injection. This finding suggests that any mutational activity possessed by chloropicrin in this organism arises from a derivative formed in the digestive tract of the fly rather than from the compound itself.

68 OR ~ N KING WATER AN D H EALTH No information on the mutagenic activity of Chloropicrin in vertebrate species was found. Reproductive Effects No studies of effects on reproduction have been found. CONCLUSIONS AND RECOMMENDATIONS Chloropicrin, formed during the chlorination of nitrite-containing and heavily organically contaminated waters, is acutely lethal and toxic by either inhalation or ingestion in mammals. Ingestion of the compound may lead to formation within the body of phosgene, dechlorinated deriv- atives of the parent compound, and nitrogen oxides. However, no me- tabolism studies have been conducted in mammals to see if these metabolites are actually produced by metabolism. Some of the apparent toxicity of Chloropicrin may be due to these derivatives instead of to the original material. Chloropicrin has lacrimatory activity, but its major effects are on the olfactory and respiratory epithelia. Inhalation exposure to chloro- picrin particularly injures medium and small bronchi, inducing extensive coughing. Injury to the alveoli is followed by pulmonary edema, which may appear relatively soon after an exposure. Death due to broncho- pneumonia, bronchiolitis obliterans, or secondary infections may occur days or even weeks after an acute exposure. An exposed individual may become sensitized to subsequent exposures to chloropicrin. Chloropicrin had questionable carcinogenic activity in the mouse and none in the rat in the one available study. Chloropicrin had indirect mu- tagenic activity in the TA100 strain of Salmonella typhimurium and pos- sibly had weak direct mutagenic activity in the WPZ her strain of Escherichia cold and the TA98 strain of S. t~yphimurium. It had no mutational activity in the TA1535, TA1537, and TA1538 strains of S. typhimurium. Chlo- ropicrin had mild mutagenic activity in the fruit fly when ingested. No information on mutagenic and carcinogenic activities of Chloropicrin in vertebrate or mammalian species seems to exist other than one assay of its carcinogenic activity in the mouse and the rat performed for NCI. No research on its effects on reproductive processes was found. No infor- mation on its catabolism within the mammalian body has been located. Additional toxicity and carcinogenicity data in mammals are required to assess adequately potential risks posed by the Chloropicrin concentrations present in finished waters. For this reason, the committee decided not to calculate a suggested no-adverse-effect level (SNARL).

Selected Disinfectants and By-Products 169 CHLOROPHENOLS MONOCH LOROPH ENOLS 2-Chlorophenol, o-ChIoropheno' CAS No. 95-57-S DICH LOROPH ENOLS 2,4-Dichlorophenol CAS No. 120-83-2 TRICHLOROPHENOLS 2,4, 6-Trichlorophenol, Dowicide 2S CAS No. B8-06-2 C6HsC10 C6H4C12O C6H3C13O 2-Chlorophenol is a light amber liquid with an unpleasant penetrating odor. It is very soluble in water with a melting point of 9. 3°C and a boiling point of 174.9°C; its density is 1.2634. 2-Chlorophenol (2-CP) is a soil sterilant and is also used for organic synthesis (dyes) and as a chemical intermediate, e.g., for higher chlorinated phenols. 2,4-Dichlorophenol (2,4-DCP) is a white, low-melting-point solid that is slightly soluble in water with a melting point of 45°C and a boiling point of 210°C. Its density is 1.383, and the vapor pressure is 1 mm of mercury at 53.0°C. 2,4-pop is used in organic synthesis, in synthesis of the anthelmintic bithionol sulfoxide, and as a chemical intermediate, e.g., for 2,4-di- chlorophenoxyacetate, Bifenox, Dichlorprop, and 4-~2,4-dichlorophen- oxy~butyrate herbicides. 2,4-pop was reviewed in Volume 1 of Drinking Water and Health (NRC, 1977, pp. 725-7261. Principally, studies com- pleted since that review are described here.

170 DRINKING WATER AND HEALTH 2,4,6-Trichlorophenol (2,4,6-TCP), Dowicide 2S, is in a yellow flake or needle form with a strong phenolic odor. Its solubility is <0.1 g/100 g of water with a boiling point of 246°C and a melting point of 69°C. Its density is 1.4901. 2,4,6-TCP is used as a fungicide, bactericide, preser- vative, and disinfectant; as an isomeric mixture, herbicide, and defoliant; and as a sanitizer. 2,4,6-TCP was reviewed in Volume 4 of Drinking Water and Health (NRC, 1982, pp. 264-2681. METABOLISM Data are not available. HEALTH EFFECTS Observations in Humans Data are not available. Observations in Other Species Acute Effects Borzelleca et al. (1985a) reported acute oral toxicity data in CD-1 ICR mice from a series of chlorinated phenols. The data are summarized in Table 4-14. Acute oral toxicity data for 2,4-pop are summarized in Table 4-15. Subchronic Effects Borzelleca et al. (1985a) exposed male and female CD-1 mice to 2,4-pop dissolved in drinking solutions (10% Emulphor) for 90 days. The concentrations of solutions and actual doses are shown in Table 4-16. There were no consistently significant compound-related adverse effects on any of the parameters evaluated (behavior, body weight, organ weights and ratios, gross pathology, hematological parameters, serum chemistries, urinalyses). Kobayashi et al. (1972) exposed male mice to 2,4-pop as a dietary mixture for 6 months at doses of 45, 100, and 230 mg/kg bw per day. The authors reported a slight abnormality in liver histopathology at the highest dose. Chronic Effects Data are not available. Reproductive Effects Seyler and associates ( 1984) evaluated the effects of a series of dichlorophenols on a number of in vitro parameters designed to assess potential reproductive effects and reported that none of the pops

171 ~ ~ U~ _ _ __ _ oo . _ , ., ~ o ~s~ _ tA.A~ W Ct~At - ~W~ ~: C~ Ct  ~_ ·_ ~_ - o ~: S:  ~5 ~: ._ o - o V, W ~At    =^ V) W o - ~3 O D w O e~ ~V~ C~ ~ ~ ~ _4 £1 oo __ ._._ W~AOO ~;>WW -`-._._ WO__ ~ V: O O  O  3 O D W -.y ~ _ ~ £ W W W . _ Ct Ct - ~ O ~ £= _ O W o m ~_ 2 oo _` _ _ ~ O ~ ~_ 1 1 0 o0C ~ vA ~(~(~1_ ~A~AvA~)~A~ _ _ _ ~  2 _` ~0 U~ 1 ~ + 0 ~ o o 3 ~o ~o o . . . X oo ~ _ _ _ O Z 2 - C~ 1 o0 ~ _ ~ C~ 2 _` ~A~ 00 - 1 - ~ ~ oo - ~ - c~ 2 ~_ ~0 C~ ~A~ O ~ ~ - 2 o - 1 ~A <) O ~ - 2 ~_ ~D - 1 C~ O O0 0  a~ ~ 2 oo ~A - 1 ~O ~A ~(~ ~ - r~ o C~ 1 00 tA~- ~(5~ - C~ 2 2 o0 ~ oo M0 ~ ~O ~ _ ~ _ ~_ a~ C~ 1 U~ _ _ 2 ~_ o C~ - ~A~ 00 ~ _ ~ o0 - C~ - 2 _ - oo - 1 ~;tA ~A~ O ~ _ ~ 2 o 1 r~ C~ - 00 ~00 ~ ~0 ~cr, _ _ _ ._ ._ ._ O O O O O O ~ O c) 2 ~_ - C~ 2 ~_  C~ U~ C~ 1 U~ - _ _ ._ ._ O O O O V ~ OO OO .. tA~tA~ __ O O O ~ O O O \~ ~A ~ ~D ~C~ _ _ _ ~ ~ o - o _ _ _ _ _ _ O O O O O O_ ~A ~ -5 ~W W W W W~ s ,s , 2 ~ 2 ' - 4 2 ~>o o o o ,_ s O ~O ~=: ~00 5; ~£A ~\ ~_ ~ o oo o ~ ·o ~ ~ o .O ~ .o ~ .Q ~ .o _ ~ °° t_ O t_ O ~ ~ ~ ~ ~ 00 ~ ~ ~ ~ ~ ~ ~e C: ~ \~ ~ ~ ~ ~ ~ W ~d- ( ~( ~( ~A ~vA~ _` D U~ -  - Ct - W ~At C) W - W O ~ ~ ._ - A £ ~ . _ ~ _ O W ~: ~ W W =, o <: ~ ~ 2

i72 DRINKING WATER AND HEALTH TABLE 4-15 Acute Oral Toxicity Data for 2,4-Dichlorophenol Acute Oral LDso, Species Strain/Sexmg/kg bw Reference Rat 1,600 Kobayashi et al., 1972 580 Deichmann, 1943 2,830 Vernot et al., 1977 Mouse 1,630 Vernot et al., 1977 CD-1/M1,276 Borzelleca et al., 1985a F1,352 Borzelleca et al., 1985a affected sperm motility; 2,5-pop, 3,4-pop, and 3,5-pop depressed sperm penetration of mouse ova; and 3,4-pop and 3,5-pop disrupted the sperm acrosome. The levels that the animals received were not provided. Mutagenicity Rapson et al. (1980) investigated a number of com- pounds believed to be produced by chlorination, including 2-CP, 2,4 DCP, and 2,3,6-TCP, in an Ames assay (TA1001. None of the compounds evaluated was found to be mutagenic in this test system. Probst et al. (1981) reported that 2,4-pop was negative in the hepatocyte unscheduled DNA synthesis and in a modified Ames assay. Carcinogenicity No data are available except for 2,4,6-TCP, which was reviewed in Volume 4 of this series (NRC, 19821. Technical-grade 2,4,6-TCP was shown to be positive in animal carcinogenicity bioassays. The role of dioxins or other impurities in the technical material in pro- ducing this effect was not established. SUMMARY AND CONCLUSIONS There are relatively few toxicological data available for the chloro- phenols, with the exception of 2,4-pop. Nonetheless data have yet to be TABLE 4-16 Dosing of CD-1 Mice with 2,4-pop in Drinking Solutions Concentrations Actual Doses of 2,4-pop Theoretical mg/kg low/day in Drinking Solution, Doses, mg/ml mg/kg low/day Females Males 0.0 0 0 0 0.2 50 50 40 0.6 150 143 114 2.0 1,500 491 383

Selected Disinfectants and By-Products 173 developed assessing the chronic effects of 2,4-pop. Kobayashi et al. (1972) had shown a possible no-observed-effect level (NOEL) of 100 ma/ kg bw per day during a 6-month dietary mouse study for 2,4-pop. More recently, Borzelleca et al. (l98Sa) were able to show no effects at levels of 383 and 491 mg/kg bw per day in male and female mice, respectively, exposed to 2,4-pop in drinking water for 90 days. To be conservative, the committee chose to use 100 mg/kg bw per day as the NOEL for estimating a suggested no-adverse-effect level (SNARL). As- suming that a 70-kg human consumes 2 liters of water daily, which con- ~ibutes 20% of total intake, a SNARL may be calculated using an uncertainty factor of 100 as: 100 mg/kg low/day x 70 kg x 0.2 = 7 mg/liter. A SNARL may also be estimated for a 10-kg child consuming 1 liter of water daily, which contributes 20% of total intake: 100 mg/kg low/day x 10 kg x 0.2 ~ . mg/l~ter. 100 x 1 liter However, these levels are above the odor threshold for 2,4-pop, so that esthetic factors would make human consumption of this amount highly unlikely. The committee recommends that further toxicological data be developed for this class of compounds. REFERENCES Chlorine ACGIH (American Conference of Governmental Industrial Hygienists, Inc.). 1986. Chlor- ine. P. 117 in Documentation of the Threshold Limit Values and Biological Exposure Indices, 5th ed. American Conference of Governmental Industrial Hygienists, Inc., Cincinnati, Ohio. Chang, J. C. F., and C. S. Barrow. 1984. Sensory irritation tolerance and cross-tolerance in F-344 rats exposed to chlorine or formaldehyde gas. Toxicol. Appl. Pharmacol. 76:319-327. Dennis, W. H., Jr., V. P. Olivieri, and C. W. Kruse. 1978. Reaction of uracil with hypochlorous acid. Biochem. Biophys. Res. Commun. 83:168-171. Hoyano, Y., V. Bacon, R. E. Summons, W. E. Pereira, B. Halpern, and A. M. Duffield. 1973. Chlorination studies. IV. The reaction of aqueous hypochlorous acid with pyrim- idine and purine bases. Biochem. Biophys. Res. Commun. 53:1195-1199. Meier, J. R., R. J. Bull, J. A. Stober, and M. C. Cimino. 1985. Evaluation of chemicals used for drinking water disinfection for production of chromosomal damage and sperm- head abnormalities in mice. Environ. Mutagen. 7:201-211.

]74 DRINKING WATER AND HEALTH NRC (National Research Council). 1980. Drinking Water and Health, Vol. 2. National Academy Press, Washington, D.C. 393 pp. Olivieri, V. P., W. H. Dennis, M. C. Snead, D. T. Richfield, and C. W. Kruse. 1980. Reaction of chlorine and chloramines with nucleic acids under disinfection conditions. Pp. 651-663 in R. L. Jolley, W. A. Brungs, R. B. Cumming, and V. A. Jacobs, eds. Water Chlorination: Environmental Impact and Health Effects, Vol. 3. Ann Arbor Sci- ence, Ann Arbor, Mich. Philipp, R., C. Shepherd, F. Fawthrop, and B. Poulsom. 1985. Domestic chlorine poi- soning. Lancet 2:495. Revis, N. W., P. McCauley, R. Bull, and G. Holdsworth. 1986. Relationship of drinking water disinfectants to plasma cholesterol and thyroid hormone levels In experimental studies. Proc. Natl. Acad. Sci. USA 83:1485-1489. Trehy, M. L., and T. I. Bieber. 1981. Detection, identification and quantitative analysis of dihaloacetonitriles in chlorinated natural waters. Pp. 941-975 in L. H. Keith, ed. Advances in the Identification & Analysis of Organic Pollutants in Water, Vol. 2. Ann Arbor Science, Ann Arbor, Mich. Vogt, C. R., J. C. Liao, G. Y. Sun, and A. Y. Sun. 1979. In vivo and in vitro formation of chloroform in rats with acute dosage of chlorinated water and the effect of membrane function. Pp. 453-460 in D. D. Hemphill, ed. Trace Substances in Environmental Health XIII. Proceedings of University of Missouri's 13th Annual Conference on Trace Substances in Environmental Health. University of Missouri, Columbia, Mo. White, G. C. 1972. Handbook of Chlorination. Van Nostrand Reinhold, New York. 744 PP ChIorine Dioxide Abdel-Rahman, M. S., D. Couri, and J. D. Jones. 1980a. Chlorine dioxide metabolism in rat. J. Environ. Pathol. Toxicol. 3(1):421-430. Abdel-Rahman, M. S., D. Couri, and R. J. Bull. 1980b. Kinetics of ClO2 and effects of ClO', ClO~-, and ClO3- in drinking water on blood glutathione and hemolysis in rat and chicken. J. Environ. Pathol. Toxicol. 3(1):431-449. Abdel-Rahman, M. S., D. Couri, and R. J. Bull. 1982. Metabolism and pharmacokinetics of alternate drinking water disinfectants. Environ. Health Perspect. 46:19-23. Abdel-Rahman, M. S., D. Couri, and R. J. Bull. 1984. Toxicity of chlorine dioxide in drinking water. J. Am. Coll. Toxicol. 3(4):277-284. Bercz, J. P., L. Jones, L. Garner, D. Murray, D. A. Ludwig, and J. Boston 1982. Subchronic toxicity of chlorine dioxide and related compounds in drinking water in the nonhuman primate. Environ. Health Perspect. 46:47-55. Bercz, J. P., L. L. Jones, R. M. Harrington, R. Bawa, and L. Condie. 1986. Mechanistic aspects of ingested chlorine dioxide on thyroid function: Impact of oxidants on iodide metabolism. Environ. Health Perspect. 69:249-254. Couri, D., and M. S. Abdel-Rahman. 1980. Effect of chlorine dioxide and metabolites on glutathione dependent system in rat, mouse and chicken blood. J. Environ. Pathol. Toxicol. 3(1):451-460. Elkins, H. B. 1959. Inorganic compounds. Pp. 79-97 in H. B. Elkins, The Chemistry of Industrial Toxicology, 2nd ed. John Wiley, New York. Ferris, B. G., Jr., W. A. Burgess, and J. Worcester. 1967. Prevalence of chronic respiratory disease in a pulp mill and a paper mill in the United States. Br. J. Ind. Med. 24:26-37.

Selected Disinfectants and By-Products 175 Gloemme, J., and K.-D. Lundgren. 1957. Health hazards from chlorine dioxide. A.M.A. Arch. Ind. Health 16:169-176. Lubbers, J. R., and J. R. Bianchine. 1984. Effects of the acute rising dose administration of chlorine dioxide, chlorate and chlorite to normal healthy adult male volunteers. J. Environ. Pathol. Toxicol. Oncol. 5(4/5):215-228. Lubbers, J. R., S. Chauhan, and J. R. Bianchine. 1981. Controlled clinical evaluations of chlorine dioxide, chlorite and chlorate in man. Fund. Appl. Toxicol. 1:33~338. Lubbers, J. R., S. Chauan, and J. R. Bianchine. 1982. Controlled clinical evaluations of chlorine dioxide, chlorite and chlorate in man. Environ. Health Perspect. 46:57-62. Lubbers, J. R., J. R. Bianchine, and R. J. Bull. 1983. Safety of oral chlorine dioxide, chlorite, and chlorate ingestion in man. Pp. 1335-1341 in R. L. Jolley, W. A. Brungs, J. A. Cotruvo, R. B. Fuming, J. S. Mattice, and V. A. Jacobs, eds. Water Chlorination: Environmental Impact and Health Effects, Vol. 4. Book 2: Environment, Health, and Risk. Ann Arbor Science, Ann Arbor, Mich. Lubbers, J. R., S. Chauhan, J. K. Miller, and J. R. Bianchine. 1984. The effects of chronic administration of chlorine dioxide, chlorite and chlorate to normal healthy adult male volunteers. J. Environ. Pathol. Toxicol. Oncol. 5(4/5):229-238. Meter, J. R., R. J. Bull, J. A. Stober, and M. C. Cimino. 1985. Evaluation of chemicals used for drinking water disinfection for production of chromosomal damage and sperm- head abnormalities in mice. Environ. Mutagen. 7:201-211. Michael, G. E., R. K. Miday, J. P. Bercz, R. G. Miller, D. G. Greathouse, D. F. Kraemer, and J. B. Lucas. 1981. Chlorine dioxide water disinfection: A prospective epidemiology study. Arch. Environ. Health 36:20-27. Moore, G. S., and E. J. Calabrese. 1980. The effects of chlorine dioxide and sodium chlorite on erythrocytes of A/J and C56L/J mice. J. Environ. Pathol. Toxicol. 4(2,3):513- 524. NRC (National Research Council). 1982. Drinking Water and Health, Vol. 4. National Academy Press, Washington, D.C. 299 pp. Orme, J., D. H. Taylor, R. D. Laurie, and R. J. Bull. 1985. Effects of chlorine dioxide on thyroid function in neonatal rats. J. Toxicol. Environ. Health 15:315-322. Revis, N. W., P. McCauley, R. Bull, and G. Holdsworth. 1986. Relationship of drinking water disinfectants to plasma cholesterol and thyroid hormone levels in experimental studies. Proc. Natl. Acad. Sci. USA 83:1485-1489. Sub, D. H., M. S. Abdel-Rahman, and R. J. Bull. 1983. Effect of chlorine dioxide and its metabolites in drinking water on fetal development in rats. J. Appl. Toxicol. 3:75- 79. Taylor, D. H., and R. J. Pfohl. 1985. Effects of chlorine dioxide on neurobehavioral development of rats. Pp. 355-364 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environ- mental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Tuthill, R. W., R. A. Giusti, G. S. Moore, and E. J. Calabrese. 1982. Health effects among newborns after prenatal exposure to ClO2-disinfected drinking water. Environ. Health Perspect. 46:39-45. ChIoramines Abdel-Rahman, M. S., M. R. Berardi, and R. J. Bull. 1982. Effect of chlorine and monochloramine in drinking water on the developing rat fetus. J. Appl. Toxicol. 2:156- 159.

176 DRINKING WATER AND HEALTH Abdel-Rahman, M. S., D. M. Waldron, and R. J. Bull. 1983. A comparative kinetics study of monochloramine and hypochlorous acid in rat. J. Appl. Toxicol. 3:175-179. Abdel-Rahman, M. S., D. H. Sub, and R. J. Bull. 1984. Toxicity of monochloramine in rat: An alternative drinking water disinfectant. J. Toxicol. Environ. Health 13:825-834. Beck, H.-I. 1983. Type I reaction to chloramine. Contact Dermatitis 9:155-156. Bercz, J. P., and R. Bawa. 1986. Iodination of nutrients in the presence of chlorine based disinfectants used in drinking water treatment. Toxicol. Lett. 34:141-147. Bercz, J. P., L. Jones, L. Garner, D. Murray, D. A. Ludwig, and J. Boston. 1982. Subchronic toxicity of chlorine dioxide and related compounds in drinking water in the nonhuman primate. Environ. Health Perspect. 46:47-55. Berthelot, P.-E.-M. 1859. Correspondance: Violet d'aniline. Repert. Chim. Pure Appl. 1(8):284. Bull, R. J. 1980. Health effects of alternate disinfectants and their reaction products. J. Am. Water Works Assoc. 72:299-303. Bull, R. J., M. Robinson, J. R. Meter, and J. Stober. 1982. Use of biological assay systems to assess the relative carcinogenic hazards of disinfection by-products. Environ. Health Perspect. 46:215-227. Burttschell, R. H., A. A. Rosen, F. M. Middleton, and M. B. Ettinger. 1959. Chlorine derivatives of phenol causing taste and odor. J. Am. Water Works Assoc. 51:205-214. Carlson, R. M., and S. Lin. 1985. Characterization of the products from the reaction of hydroxybenzoic and hydroxycinnamic acids with aqueous solutions of chlorine, chlorine dioxide, and chloramine. Pp. 835-841 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Carlton, B. D., P. Barlett, A. Basaran, K. Colling, I. Osis, and M. K. Smith. 1986. Reproductive effects of alternative disinfectants. Environ. Health Perspect. 69:237-241. Cheh, A. M., J. Skochdopole, C. Heilig, P. M. Koski, and L. Cole. 1980a. Destruction of direct-acting mutagens in drinking water by nucleophiles: Implications for mutagen identification and mutagen elimination from drinking water. Pp. 803-815 in R. L. Jolley, W. A. Brungs, R. B. Cumming, and V. A. Jacobs, eds. Water Chlorination: Environ- mental Impact and Health Effects, Vol. 3. Ann Arbor Science, Ann Arbor, Mich. Cheh, A. M., J. Skochdopole, P. Koski, and L. Cole. 1980b. Nonvolatile mutagens in drinking water: Production by chlorination and destruction by sulfite. Science 207:90- 92. Colton, E., and M. M. Jones. 1955. Monochloramine. J. Chem. Ed. 32:485-487. Eaton, J. W., C. F. Kolpin, H. S. Swofford, C.-M. Kjellstrand, and H. S. Jacob. 1973. Chlorinated urban w-ater: A cause of dialysis-induced hemolytic anemia. Science 181:463- 464. EPA (U.S. Environmental Protection Agency). 1983. National interim primary drinking water regulations; trihalomethanes. Fed. Regist. 48:8406-8414. Fetner, R. H. 1962. Chromosome breakage in Vicia faba by monochloramine. Nature 196:1 122-1 123. Grisham, M. B., M. M. Jefferson, D. F. Melton, and E. L. Thomas. 1984. Chlorination of endogenous amines by isolated neutrophils. Ammonia-dependent bactericidal, cyto- toxic, and cytolytic activities of the chloramines. J. Biol. Chem. 259:1040~10413. GSRI (Gulf South Research Institute). 1981. A subchronic study of chloramine generated in-situ in the drinking water of F344 rats and B6C3F1 mice. Draft report prepared for Tracor-Jitco, Inc., Rockville, Md. (276 pp.)

Selected Disinfectants and By-Products ~ 77 Herren-Freund, S. L., and M. A. Pereira. 1986. Carcinogenicity of by-products of dis- infection in mouse and rat liver. Environ. Health Perspect. 69:59-65. Isaac, R. A., and J. C. Morris. 1980. Rates of transfer of active chlorine between nitro- genous substrates. Pp. 183-191 in R. L. Jolley, W. A. Brungs, R. B. Cumming, and V. A. Jacobs, eds. Water Chlorination: Environmental Impact and Health Effects, Vol. 3. Ann Arbor Science, Ann Arbor, Mich. Isaac, R. A., and J. C. Morris. 1983. Transfer of active chlorine from chloramine to nitrogenous organic compounds. 1. Kinetics. Environ. Sci. Technol. 17:738-742. Isaac, R. A., and J. C. Morris. 1985. Transfer of active chlorine from chloramine to nitrogenous organic compounds. 2. Mechanism. Environ. Sci. Technol. 19:810-814. Jacangelo, J. G., and V. P. Olivieri. 1985. Aspects of the mode of action of monochlor- amine. Pp. 575-586 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Kjellstrand, C. M., J. W. Eaton, Y. Yawata, H. Swofford, C. F. Kolpin, T. J. Buselmeier, B. von Hartitzsch, and H. S. Jacob. 1974. Hemolysis in dialized patients caused by chloramines. Nephron 13:427-433. Kovacic, P., M. K. Lowery, and K. W. Field. 1970. Chemistry of N-bromamines and N- chloramines. Chem. Rev. 70:639-665. Laakso, M., I. Arvala, S. Tervonen, and M. Sotarauta. 1982. Chloramine-induced pneu- monitis from mixing household cleaning agents. Br. Med. J. 285:1086. Le Cloirec, C., and G. Martin. 1985. Evolution of amino acids in water treatment plants and the effect of chlorination on amino acids. Pp. 821-834 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Lubbers, J. R., S. Chauhan, and J. R. Bianchine. 1981. Controlled clinical evaluations of chlorine dioxide, chlorite and chlorate in man. Fund. Appl. Toxicol. 1:33i338. Maziarka, S., I. Kongiel-Chablo, M. Rybak, S. Szulinski, and J. Wojcik. 1976. Investi- gations of effects of chlorine and chloramines in drinking water on animal organisms. Rocz. Panstw. Zakl. Hig. 27:123-131. (in Polish; English summary) Meier, J. R., R. J. Bull, J. A. Stober, and M. C. Cimino. 1985. Evaluation of chemicals used for drinking water disinfection for production of chromosomal damage and sperm- head abnormalities in mice. Environ. Mutagen. 7:201-211. Moore, G. S., and E. J. Calabrese. 1980. The health effects of chloramines in potable water supplies: A literature review. J. Environ. Pathol. Toxicol. 4(1):257-263. Moore, G. S., E. J. Calabrese, and M. McGee. 1980. Health effects of monochloramines in drinking water. J. Environ. Sci. Health A15:239-258. Revis, N. W., P. McCauley, R. Bull, and G. Holdsworth. 1986. Relationship of drinking water disinfectants to plasma cholesterol and thyroid hormone levels in experimental studies. Proc. Natl. Acad. Sci. USA 83:1485-1489. Rosenblatt, D. H. 1975. Chlorine and oxychlorine species reactivity with organic sub- stances. Pp. 249-276 in J. D. Johnson, ed. Disinfection: Water and Wastewater. Ann Arbor Science, Ann Arbor, Mich. Scully, F. E., Jr., K. E. Mazina, D. E. Sonenshine, and F. B. Daniel. 1985. Reactions of hypochlorite and organic N-chloramines in stomach fluid. Pp. 175-184 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich.

178 DRINKING WATER AND HEALTH Scully, F. E., Jr., K. Mazina, D. Sonenshine, and F. Kopfler. 1986. Quantitation and identification of organic N-chloramines formed in stomach fluid on ingestion of aqueous hypochlorite. Environ. Health Perspect. 69:259-265. ShTh, K. L., and J. Lederberg. 1976. Chloramine mutagenesis in Bacillus subtilis. Science 192:1141-1143. Snyder, M. P., and D. W. Margerum. 1982. Kinetics of chlorine transfer from chloramine to amines, amino acids, and peptides. Inorg. Chem. 21:2545-2550. Sub, D. H., and M. S. Abdel-Rahman. 1983. Kinetics study of chloride in rat. J. Toxicol. Environ. Health 12:467-473. Wilcox, P., and S. Denny. 1985. Effect of dechlorinating agents on the mutagenic activity of chlorinated water samples. Pp. 1341-1353 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Chiorite and Chlorate Abdel-Rahman, M. S. 1985. Pharmacokinetics of chlorine obtained from chlorine dioxide, chlorine, chloramine and chloride. Pp. 281-293 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Abdel-Rahman, M. S., D. Couri, and J. D. Jones. 1980. Chlorine dioxide metabolism in rat. J. Environ. Pathol. Toxicol. 3(1):421-430. Abdel-Rahman, M. S., D. Couri, and R. J. Bull. 1982. Metabolism and pharmacokinetics of alternate drinking water disinfectants. Environ. Health Perspect. 46:19-23. Abdel-Rahman, M. S., D. Couri, and R. J. Bull. 1984a. The kinetics of chlorite and chlorate in the rat. J. Am. Colt. Toxicol. 3(4):261-267. Abdel-Rahman, M. S., D. Couri, and R. J. Bull. 1984b. Toxicity of chlorine dioxide in drinking water. J. Am. Coil. Toxicol. 3(4):277-284. Bercz, J. P., L. Jones, L. Garner, D. Murray, D. A. Ludwig, and J. Boston. 1982. Subchronic toxicity of chlorine dioxide and related compounds in drinking water in the nonhuman primate. Environ. Health Perspect. 46:47-55. Bloxham, C. A., N. Wright, and J. G. Hoult. 1979. Self-poisoning by sodium chlorate- some unusual features . Clin. Toxicol . 15: 185- 188. Carlton, B. D., and M. K. Smith. 1985. Reproductive effects of alternate disinfectants and their by-products. Pp. 295-305 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environ- mental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Couri, D., and M. S. Abdel-Rahman. 1980. Effect of chlorine dioxide and metabolites on glutathione dependent system in rat, mouse and chicken blood. J. Environ. Pathol. Toxicol. 3(1):451-460. Couri, D., M. S. Abdel-Rahman, and R. J. Bull. 1982. Toxicological effects of chlorine dioxide, chlorite and chlorate. Environ. Health Perspect. 46:13-17. Eckhardt, K., E. Gocke, M.-T. King, and D. Wild. 1982. Mutagenic activity of chlorate, bromate, and iodate. Mutat. Res. 97:85. (Abstract 42) Haag, H. B. 1949. The Effect on Rats of Chronic Administration of Sodium Chlorite and Chlorine Dioxide in the Drinking Water. Report submitted to The Mathieson Alkali Works, Niagara Falls, N.Y. (Prepared by the Department of Physiology and Pharma- cology, Medical College of Virginia, Richmond, Va.) 18 pp. Heffernan, W. P., C. Guion, and R. J. Bull. 1979a. Oxidative damage to the erythrocyte induced by sodium chlorite, in vitro. J. Environ. Pathol. Toxicol. 2:1501-1510.

Selected Disinfectants and By-Products 179 Heffernan, W. P., C. Guion, and R. J. Bull. 1979b. Oxidative damage to the erythrocyte induced by sodium chlorite, in vivo. J. Environ. Pathol. Toxicol. 2:1487-1499. Helliwell, M., and J. Nunn. 1979. Mortality in sodium chlorate poisoning. Br. Med. J. 1:1119. Kurokowa, Y., N. Takamura, Y. Matsushima, T. Imazawa, and Y. Hayashi. 1984. Studies on the promoting and complete carcinogenic activities of some oxidizing chemicals in skin carcinogenesis. Cancer Lett. 24:299-304. Lubbers, J. R., and J. R. Bianchine. 1984. Effects of the acute rising dose administration of chlorine dioxide, chlorate and chlorite to normal healthy adult male volunteers. J. Environ. Pathol. Toxicol. Oncol. 5(4/5):215-228. Lubbers, J. R., S. Chauhan, and J. R. Bianchine. 1981. Controlled clinical evaluations of chlorine dioxide, chlorite and chlorate in man. Fund. Appl. Toxicol. 1:334-338. Lubbers, J. R., S. Chauan, and J. R. Bianchine. 1982. Controlled clinical evaluations of chlorine dioxide, chlorite and chlorate in man. Environ. Health Perspect. 46:57-62. Lubbers, J. R., J. R. Bianchine, and R. J. Bull. 1983. Safety of oral chlorine dioxide, chlorite, and chlorate ingestion in man. Pp. 1335-1341 in R. L. Jolley, W. A. Brungs, J. A. Cctruvo, R. B. Cumming, J. S. Mattice, and V. A. Jacobs, eds. Water Chlorination: Environmental Impact and Health Effects, Vol. 4. Book 2: Environment, Health, and Risk. Ann Arbor Science, Ann Arbor, Mich. Lubbers, J. R. S. Chauhan, J. K. Miller, and J. R. Bianchine. 1984. The effects of chronic administration of chlorine dioxide, chlorite and chlorate to normal healthy adult male volunteers. J. Environ. Pathol. Toxicol. Oncol. 5(4/5):229-238. Meter, J. R., R. J. Bull, J. A. Stober, and M. C. Cimino. 1985. Evaluation of chemicals used for drinking water disinfection for production of chromosomal damage and sperm- head abnormalities in mice. Environ. Mutagen. 7:201-21 1 . Moore, G. S., and E. J. Calabrese. 1980. The effects of chlorine dioxide and sodium chlorite onerythrocytesofA/J and C56L/J mice. J. Environ. Pathol. Toxicol. 4(2,3):513- 524. Moore, G. S., and E. J. Calabrese. 1982. Toxicological effects of chlorite in the mouse. Environ. Health Perspect. 46:31-37. NRC (National Research Council). 1980. Drinking Water and Health, Vol. 3. National Academy Press, Washington, D.C. 415 pp. NRC (National Research Council). 1982. Drinking Water and Health, Vol. 4. National Academy Press, Washington, D.C. 299 pp. Stavrou, A., R. Butcher, and A. Sakula. 1978. Accidental self-poisoning by sodium chlorate weed-killer. The Practitioner 221:397-399. Steffen, C., and R. Seitz. 1981. Severe chlorate poisoning: Report of a case. Arch. Toxicol. 48:281-288. Sub, D. H., M. S. Abdel-Rahman, and R. J. Bull. 1983. Effect of chlorine dioxide and its metabolites in drinking water on fetal development in rats. J. Appl. Toxicol. 3:75- 79. Sub, D. H., M. S. Abdel-Rahman, and R. J. Bull. 1984. Biochemical interactions of chlorine dioxide and its metabolites in rats. Arch. Environ. Contam. Toxicol. 13:163- 169. Chloroform and DibromochIoromethane Ahmadizadeh, M., C.-H. Kuo, and J. B. Hook. 1981. Nephrotox~city and hepatotoxicity of chloroform in mice: Effect of deuterium substitution. J. Toxicol. Environ. Health 8:105-111.

]80 DRINKING WATER AND HEALTH Bomski, H., A. Sobolewska, and A. Strakowski. 1967. Toxische Schadigung der Leber durch Chloroform bet Chemiebetriebswerkern. Int. Arch. Gewerbepathol. Gewerbehyg. 24:127-134. Branchflower, R. V., D. S. Nunn, R. J. Highet, J. H. Smith, J. B. Hook, and L. R. Pohl. 1984. Nephrotoxicity of chloroform: Metabolism to phosgene by the mouse kidney. Tox~col. Appl. Pharmacol. 72:159-168. Bull, R. J., J. M. Brown, E. A. Meierhenry, T. A. Jorgenson, M. Robinson, and J. A. Stober. 1986. Enhancement of the hepatotoxicity of chloroform in female B6C3F~ mice by corn oil: Implications for chloroform carcinogenesis. Environ. Health Perspect. 69:49- 58. Callen, D. F., C. R. Wolf, and R. M. Philpot. 1980. Cytochrome P-450 mediated genetic activity and cytotoxicity of seven halogenated aliphatic hydrocarbons in Saccharomyces cerevisiae. Mutat. Res. 77:55-63. Cantor, K. P., R. Hoover, T. Mason, and L. J. McCabe. 1977. Association of cancer mortality rates and trihalomethane levels in municipal drinking water supplies. Am. J. Epidemiol. 106:230-231. (Abstract) Cantor, K. P., R. Hoover, T. J. Mason, and L. J. McCabe. 1978. Associations of cancer mortality with halomethanes in drinking water. J. Natl. Cancer Inst. 61:979-985. Cantor, K. P., R. Hoover, P. Hartge, T. J. Mason, D. T. Silverman, and L. I. Levin. 1985. Drinking water source and risk of bladder cancer: A case-control study. Pp. 145- 152 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Challen, P. J. R., D. E. Hickish, and J. Bedford. 1958. Chronic chloroform intoxication. Br. J. Ind. Med. 15:243-249. Chu, I., D. C. Villeneuve, V. E. Secours, and G. C. Becking. 1982a. Toxicity of trihalo- methanes: I. The acute and subacute toxicity of chloroform, bromodichloromethane, chlorodibromomethane and bromoform in rats. J. Environ. Sci. Health B17:205-224. Chu, I., D. C. Villeneuve, V. E. Secours, and G. C. Becking. 1982b. Trihalomethanes: II. Reversibility of toxicological changes produced by chloroform, bromodichloro- methane, chlorodibromomethane end bromoformin rats. J. Environ. Sci. HealthB17:225- 240. Cragle, D. L., C. M. Shy, R. J. Struba, and E. J. Siff. 1985. A case-control study of colon cancer and water chlorination in North Carolina. Pp. 153-159 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Pub- lishers, Chelsea, Mich. Cresteil, T., P. Beaune, J. P. Leroux, M. Lange, and D. Mansuy. 1979. Biotransformation of chloroform by rat and human liver microsomes; in vitro effect on some enzyme activities and mechanism of irreversible binding to macromolecules. Chem.-Biol. Inter- act. 24:153-165. Dunnick, J. K., J. K. Haseman, H. S. Lilja, and S. Wyand. 1985. Toxicity and carcin- ogenicity of chlorodibromomethane in Fischer 344/N rats and B6C3F~ mice. Fund. Appl. Toxicol. 5:1128-1136. Eschenbrenner, A. B., and E. Miller. 1945. Induction of hepatomas in mice by repeated oral administration of chloroform, with observations on sex differences. J. Natl. Cancer Inst. 5:251-255. Fry, B. J., T. Taylor, and D. E. Hathway. 1972. Pulmonary elimination of chloroform and its metabolite in man. Arch. Int. Pharmacodyn. 196:98-111.

Selected Disinfectants and By-Products 181 Herren-Freund, S. L., and M. A. Pereira. 1986. Carcinogenicity of by-products of dis- infection in mouse and rat liver. Environ. Health Perspect. 69:59-65. Hewitt, W. R., H. Miyajima, M. G. Ct. and G. L. Plaa. 1980. Acute alteration of chloroform-induced hepato- and nephrotoxicity by n-hexane, methyl n-butyl ketone, and 2,5-hexanedione. Toxicol. Appl. Pharmacol. 53:230-248. Heywood, R., R. J. Sortwell, P. R. B. Noel, E. A. Street, D. E. Prentice, F. J. C. Roe, P. F. Wadsworth, A. N. Worden, and N. J. Van Abbe. 1979. Safety evaluation of toothpaste containing chloroform. III. Long-term study in beagle dogs. J. Environ. Pathol. Toxicol. 2(3):835-851. Hill, R. N. 1977. Differential toxicity of chloroform in the mouse. Ann. N.Y. Acad. Sci. 298: 170-176. Jorgenson, T. A., and C. J. Rushbrook. 1980. Effects of Chloroform in the Drinking Water of Rats and Mice: Ninety-Day Subacute Toxicity Study. Publication no. EPA-600/1-80- 030. Health Effects Research Laboratory, U.S. Environmental Protection Agency, Cin- cinnati. 44 pp. (Available from the National Technical Information Service, Springfield, Va., as publication no. PB-80-219108) Jorgenson, T. A., C. J. Rushbrook, and D. C. L. Jones. 1982. Dose-response study of chloroform carcinogenesis in the mouse and rat: Status report. Environ. Health Perspect. 46:141-149. Jorgenson, T. A., E. F. Meierhenry, C. J. Rushbrook, R. J. Bull, and M. Robinson. 1985. Carcinogenicity of chloroform in drinking water to male Osborne-Mendel rats and female B6C3F1 mice. Fund. Appl. Toxicol. 5:760-769. Klaunig, J. E., R. J. Ruch, and M. A. Periera. 1986. Carcinogenicity of chlorinated methane and ethane compounds administered in drinking water to mice. Environ. Health Perspect. 69:89-95. Kluwe, W. M., and J. B. Hook. 1981. Potentiation of acute chloroform nephrotoxicity by the glutathione depletor diethyl maleate and protection by the microsomal enzyme in- hibitor piperonyl butoxide. Toxicol. Appl. Pharmacol. 59:457-466. Kuzma, R. J., C. M. Kuzma, and C. R. Buncher. 1977. Ohio drinking water source and cancer rates. Am. J. Public Health 67:725-729. Moore, D. H., L. F. Chasseaud, S. K. Majfed, D. E. Prentice, F. J. C. Roe, and N. J. Van Abbe. 1982. The effect of dose and vehicle on early tissue damage and regenerative activity after chloroform administration to mice. Food Chem. Toxicol. 20:951-954. Munson, A. E., L. E. Sain, V. M. Sanders, B. M. Kauffmann, K. L. White, Jr., D. G. Page, D. W. Barnes, and J. F. Borzelleca. 1982. Toxicology of organic drinking water contaminants: Trichloromethane, bromodichloromethane, dibromochloromethane and tri- bromomethane. Environ. Health Perspect. 46:117-126. NCI (National Cancer Institute). 1976. Report on the Carcinogenesis Bioassay of Chlo- roform. National Cancer Institute, Bethesda, Md. (64 pp.) (Available from the National Technical Information Service, Springfield, Va., as publication no. PB-264-018/AS) NRC (National Research Council). 1977. Drinking Water and Health. National Academy of Sciences, Washington, D.C. 939 pp. NRC (National Research Council). 1980. Drinking Water and Health, Vol. 3. National Academy Press, Washington, D.C. 415 pp. NRC (National Research Council). 1982. Drinking Water and Health, Vol. 4. Nationai Academy Press, Washington, D.C. 299 pp. NTP (National Toxicology Program). 1985. Toxicology and Carcinogenesis Studies of Chlorodibromomethane (CAS No. 124-48-1) in F344/N Rats and B6C3F1 Mice (Gavage Studies). NTP Technical Report Series No. 282. National Toxicology Program, U.S. Department of Health and Human Services, Research Triangle Park, N.C. 174 pp.

182 DRINKING WATER AND HEALTH Page, T., R. H. Harris, and S. S. Epstein. 1976. Drinking water and cancer mortality in Louisiana. Science 193:55-57. Palmer, A. K., A. E. Street, F. J. C. Roe, A. N. Worden, and N. J. Van Abbe. 1979. Safety evaluation of toothpaste containing chloroform. II. Long term studies in rats. J. Environ. Pathol. Toxicol. 2(3):821-833. Pereira, M. A., G. L. Knutsen, and S. L. Herren-Freund. 1985. Effect of subsequent treatment of chloroform or phenobarbital on the incidence of liver and lung tumors initiated by ethylnitrosourea in 15 day old mice. Carcinogenesis 6:203-207. Pohl, L. R., and G. Krishna. 1978. Deuterium isotope effect in bioactivation and hepa- totoxicity of chloroform. Life Sci. 23: 1067- 1072. Pohl, L. R., B. Bhooshan, N. F. Whittaker, and G. Krishna. 1977. Phosgene: A metabolite of chloroform. Biochem. Biophys. Res. Commun. 79:684-691. Rapson, W. H., M. A. Nazar, and V. V. Butsky. 1980. Mutagenicity produced by aqueous chlorination of organic compounds. Bull. Environ. Contam. Toxicol. 24:590-596. Roe, F. J. C., A. K. Palmer, A. N. Worden, and N. J. Van Abbe. 1979. Safety evaluation of toothpaste containing chloroform. I. Long-term studies in mice. J. Environ. Pathol. Toxicol. 2(3):799-819. Ruch, R. J., J. E. Klaunig, N. E. Schultz, A. B. Askari, D. A. Lacher, M. A. Pereira, and P. Goldblatt. 1986. Mechanisms of chloroform and carbon tetrachloride toxicity in primary cultured mouse hepatocytes. Environ. Health Perspect. 69:301-305. Smith, J. H., and J. B. Hook. 1984. Mechanism of chloroform nephrotoxicity. III. Renal and hepatic microsomal metabolism of chloroform in mice. Toxicol. Appl. Pharmacol. 73:511-524. Stevens, J. L., and M. W. Anders. 1981. Effect of cysteine, diethyl maleate, and pheno- barbital treatments on the hepatotoxicity of [iH]- and [2H]chloroform. Chem.-Biol. Interact. 37:207-217. Withey, J. R., B. T. Collins, and P. G. Collins. 1983. Effect of vehicle on the pharma- cokinetics and uptake of four halogenated hydrocarbons from the gastrointestinal tract of the rat. J. Appl. Toxicol. 3:249-253. Young, T. B., D. A. Wolf, and M. S. Kanarek. In press. Case-control study of colon cancer and drinking water trihalomethanes in Wisconsin. Int. J. Epidemiol. Haloacids Blackshear, P. J., P. A. H. Holloway, and K. G. M. M. Alberti. 1974. The metabolic effects of sodium dichloroacetate in the starved rat. Biochem. J. 142:279-286. Crabb, D. W., E. A. Yount, and R. A. Harris. 1981. The metabolic effects of dichlo- roacetate. Metab. Clin. Exp. 30:1024-1039. Davis, M. E. 1986. Effect of chloroacetic acids on the kidneys. Environ. Health Perspect. 69:209-214. Demaugre, F., C. Cepanec, and J.-P. Leroux. 1978. Characterization of oxalate as a catabolite of dichloroacetate responsible for the inhibition of gluconeogenesis and py- ruvate carboxylation in rat liver cells. Biochem. Biophys. Res. Commun. 85:1180-1185. Dierickx, P. J. 1984. In vitro binding of acetic acid and its chlorinated derivatives by the soluble glutathione S-transferases from rat liver. Res. Commun. Chem. Pathol. Phar- macol. 44:327-330. Evans, O. B. 1982. Dichloroacetate tissue concentrat~ons and its relationship to hypolac- tatemia and pyruvate dehydrogenase activation. Biochem. Pharmacol. 31:3124-3126.

Selected Disinfectants and By-Products If33 Harris, R. A., D. W. Crabb, and R. M. Sans. 1978. Studies on the regulation of leucine catabolism. II. Mechanism responsible for dichloroacetate stimulation of leucine oxi- dation by the liver. Arch. Biochem. Biophys. 190:8-16. Hawley, G. G. 1981. The Condensed Chemical Dictionary, 10th ed. Van Nostrand Rein- hold, New York. 1,135 pp. Katz, R., C. N. Tai, R. M. Diener, R. F. McConnell, and D. E. Semonick. 1981. Dichloroacetate, sodium: 3-month oral toxicity studies in rats and dogs. Toxicol. Appl. Pharmacol. 57:273-287. Lukas, G., K. H. Vyas, S. D. Brindle, A. R. Le Sher, and W. E. Wagner, Jr. 1980. Biological disposition of sodium dichloroacetate in animals and humans after intravenous administration. J. Pharm. Sci. 69:419-421. Paxton, R., and R. A. Harris. 1984. Clofibric acid, phenylpyruvate, and dichloroacetate inhibition of branched-chain cx-ketoacid dehydrogenase kinase in vitro and in perfused rat heart. Arch. Biochem. Biophys. 231:58-66. Rapson, W. H., M. A. Nazar, and V. V. Butsky. 1980. Mutagenicity produced by aqueous chlorination of organic compounds. Bull. Environ. Contam. Toxicol. 24:590-596. Stacpoole, P. W., G. W. Moore, and D. M. Kornhauser. 1978. Metabolic effects of dichloroacetate in patients with diabetes mellitus and hyperlipoproteinemia. N. Engl. J. Med. 298:526-530. Stacpoole, P. W., G. W. Moore, and D. M. Kornhauser. 1979. Toxicity of chronic dichloroacetate. N. Engl. J. Med. 300:372. Stacpoole, P. W., E. M. Harman, S. H. Curry, T. G. Baumgartner, and R. I. Misbin. 1983. Treatment of lactic acidosis with dichloroacetate. N. Engl. J. Med. 309:390-396. Waskell, L. 1978. A study of the mutagenicity of anesthetics and their metabolites. Mutat. Res. 57:141-153. Windholz, M., S. Budavart, R. F. Blumetti, and E. S. Otterbein. 1983. The Merck Index, 10th ed. Merck & Co., Rahway, N.J. 1,463 pp. ( + supplementary material) Yount, E. A., S. Y. Felten, B. L. O'Connor, R. G. Peterson, R. S. Powell, M. N. Yum, and R. A. Harris. 1982. Comparison of the metabolic and toxic effects of 2-chloropro- pionate and dichloroacetate. J. Pharmacol. Exp. Ther. 222:501-508. Haloaldehydes Beauchamp, R. O., Jr., D. A. Andjelkovich, A. D. Kligerman, K. T. Morgan, and H. d'A. Heck. 1985. A critical review of the literature on acrolein toxicity. CRC Crit. Rev. Toxicol. 14:309-380. Bignami, M., R. Crebelli, G. Conti, L. Conti, F. Misuraca, A. Puglia, R. Randazzo, G. Sciandrello, and A. Carere. 1980a. In vitro mutagenenicity studies with halogenated aliphatic hydrocarbons. Mutat. Res. 74:200. (Abstract 62) Bignami, M., G. Conti, L. Conti, R. Crebelli, F. Misuraca, A. M. Puglia, R. Randazzo, G. Sciandrello, and A. Carere. 1980b. Mutagenicity of halogenated aliphatic hydrocar- bons in Salmonella typhimurium, Streptomyces coelicolor and Aspergillus nidulans. Chem.- Biol. Interact. 30:9-23. Bruce, W. R., and J. A. Heddle. 1979. The mutagen~c activity of 61 agents as determined by the micronucleus, Salmonella, and sperm abnormality assays. Can. J. Genet. Cytol. 21:319-333. Butler, T. C. 1948. The metabolic fate of chloral hydrate. J. Pharmacol. Exp. Ther. 92:49- 58.

184 DRINKING WATER AND HEALTH Cooper, J. R., and P. J. Friedman. 1958. The enzymic oxidation of chloral hydrate to trichloroacetic acid. Biochem. Pharmacol. 1 :76-82. Feron, V. J., and A. Kraysse. 1977. Effects of exposure to acrolein vapor in hamsters simultaneously treated with benzo[a]pyrene or diethylnitrosamine. J. Toxicol. Environ. Health 3:379-394. Gilman, A. G., L. S. Goodman, T. W. Rail, and F. Murad. 1985. Goodman and Gilman's The Pharmacological Basis of Therapeutics, 7th ed. Macmillan, New York. 1,839 pp. Green, T., and D. E. Hathway. 1978. Interactions of vinyl chloride with rat-liver DNA in vivo. Chem.-Biol. Interact. 22:211-224. Guengerich, F. P., W. M. Crawford, Jr., and P. G. Watanabe. 1979. Activation of vinyl chloride to covalently bound metabolites: Roles of 2-chloroethylene oxide and 2-chlo- roacetaldehyde. Biochemistry 18:5177-5182. Ikeda, M., and H. Ohtsuji. 1972. A comparative study of the excretion of Fujiwara reaction- positive substances in urine of humans and rodents given trichloro- or tetrachloro-deri- vatives of ethane and ethylene. Br. J. Ind. Med. 29:99-104. Ikeda, M., Y. Miyake, M. Ogata, and S. Ohmori. 1980. Metabolism of trichloroethylene. Biochem. Pharmacol. 29:2983-2992. Kallman, M. J., G. L. Kaempf, and R. L. Balster. 1984. Behavioral toxicity of chloral in mice: An approach to evaluation. Neurobehav. Toxicol. Teratol. 6:137-146. Kauffmann, B. M., K. L. White, Jr., V. M. Sanders, K. A. Douglas, L. E. Sain, J. F. Borzelleca, and A. E. Munson. 1982. Humoral and cell-mediated immune status in mice exposed to chloral hydrate. Environ. Health Perspect. 44:147-151. Kenney, W. C. 1984. Formation of Schiff base adduct between acetaldehyde and rat liver microsomal phosphatidylethanolamine. Alcohol. Clin. Exp. Res. 8:551-555. LaBaume, L. B., and R. W. Guynn. 1985. Investigation into the substrate capacity of the acetaldehyde-tetrathydrofolate condensation product. Pp. 189-200 in M. A. Collins, ed. Aldehyde Adducts in Alcoholism. Progress in Clinical and Biological Research, Vol. 183. Alan Liss, New York. Lawrence, W. H., E. O. Dillingham, J. E. Turner, and J. Autian. 1972. Toxicity profile of chloacetaldehyde. J. Pharm. Sci. 61:19-25. Liebler, D. C., M. J. Meredith, and F. P. Guengerich. 1985. Formation of glutathione conjugates by reactive metabolites of vinylidene chloride in microsomes and isolated hepatocytes. Cancer Res . 45: 186- 193. Meier, J. R., H. P. Ringhand, W. E. Coleman, J. W. Munch, R. P. Streicher, W. H. Kaylor, and K. M. Schenck. 1985. Identification of mutagenic compounds formed during chlorination of humic acid. Mutat. Res. 157:111-122. Miller, R. E., and F. P. Guengerich. 1982. Oxidation of trichloroethylene by liver micro- somal cytochrome P-450: Evidence for chlorine migration in a transition state not in- volving trichloroethylene oxide. Biochemistry 21:1090-1097. Miller, R. E., and F. P. Guengerich. 1983. Metabolism of trichloroethylene in isolated hepatocytes, microsomes, and reconstituted enzyme systems containing cytochrome P- 450. Cancer Res. 43:1145-1152. Muller, G., M. Spassovski, and D. Henschler. 1974. Metabolism of trichloroethylene in man. II. Pharmacokinetics of metabolites. Arch. Toxicol. 32:283-295. NIOSH (National Institute for Occupational Safety and Health). 1974. The Toxic Substance List. HEW Publication No. (NIOSH) 74-134. National Institute for Occupational Safety and Health, U.S. Department of Health, Education, and Welfare, Rockville, Md. NIOSH/OSHA (National Institute for Occupational Safety and Health/Occupational Safety and Health Administration). 1981. Occupational health guidelines for chloroacetaldehyde

Selected Disinfectants and By-Products 185 (September 1978). In F. W. Mackison, R. S. Stricoff, and L. J. Partridge, Jr., eds. Occupational Health Guidelines for Chemical Hazards. DHHS (NIOSH) Publication No. 81-123. U.S. Department of Health and Human Services/U.S. Department of Labor, Washington, D.C. 5 pp. (Available from the National Technical Information Service, Springfield, Va., as Publication No. PB 83-154-609) NRC (National Research Council). 1981. Formaldehyde and Other Aldehydes. National Academy Press, Washington, D.C. 340 pp. Ogata, M. 1982. Studies of the metabolism and the toxicity of organochlorine compounds. I. Species difference of the metabol~tes. Saishin Igaku 37:2404-2408. (in Japanese; English abstract in Chem. Abstr. 98:66721r) Ogata, M., Y. Takatsuka, and K. Tomokuni. 1971. Excretion of organic chlorine com- pounds in the urine of persons exposed to vapours of trichloroethylene and tetrachloro- ethylene. Br. J. Ind. Med. 28:386-391. Perrard, M. H. 1985. Mutagenicity and toxicity of chloroethylene oxide and chloroacet- aldehyde. Experientia 41:676-677. Sanders, V. M., B. M. Kauffmann, K. L. White, Jr., K. A. Douglas, D. W. Barnes, L. E. Sain, T. J. Bradshaw, J. F. Borzelleca, and A. E. Munson. 1982. Toxicology of chloral hydrate in the mouse. Environ. Health Perspect. 44:137-146. Saunders, D. S., Jr., and C. Harper. 1980. Evidence for pulmonary injury by trichloro- ethylene. Fed. Proc. 39:545. (Abstract 1474) Saunders, D. S., and C. Harper. 1981. Pulmonary toxicity of chloral hydrate and trichloro- acetaldehyde. Fed. Proc. 40:698. (Abstract 2681) Sellers, E. M., M. Lang, J. Koch-Weser, E. LeBlanc, and H. Kalant. 1972. Interaction of chloral hydrate and ethanol in man. I. Metabolism. Clin. Pharmacol. Ther. 13:37- 49. Tuma, D. J., and M. F. Sorrell. 1985. Covalent binding of acetaldehyde to hepatic proteins: Role in alcoholic liver injury. Pp. 3-17 in M. A. Collins, ed. Aldehyde Adducts in Alcoholism. Progress in Clinical and Biological Research, Vol. 183. Alan Liss, New York. Van Duuren, B. L., B. M. Goldschmidt, G. Loewengart, A. C. Smith, S. Melchionne, I. Seldman, and D. Roth. 1979. Carcinogenicity of halogenated olefinic and aliphatic hydrocarbons in mice. J. Natl. Cancer Inst. 63:1433-1439. Yuki, T., T. Hashimoto, S. Ohkuma, J.-I. Tamura, and K. Kurlyama. 1984. Alteration of acetaldehyde metabolism in carbon tetrachloride-intoxicated rat liver: Analysis using liver perfusion system. Alcohol Alcohol . 19: 101 - 107. Zajdela, F., A. Croisy, A. Barbin, C. Malaveille, L. Tomatis, and H. Bartsch. 1980. Carcinogenicity of chloroethylene oxide, an ultimate reactive metabolite of vinyl chloride, and bis(chloromethyl) ether after subcutaneous administration and in initiation-promotion experiments in mice. Cancer Res. 40:352-356. Haloketones Borzelleca, J. F., and D. Lester. 1965. Acute toxicity of some perhalogenated acetones. Toxicol. Appl. Pharmacol. 7:592-597. Nestmann, E. R., G. R. Douglas, D. J. Kowbel, and T. R. Harrington. 1985. Solvent interactions with test compounds and recommendations for testing to avoid artifacts. Environ. Mutagen. 7: 163-170. Weast, R. C., ed. 1983. CRC Handbook of Chemistry and Physics, 64th ed. CRC Press, Boca Raton, Fla.

186 DRINKING WATER AND HEALTH Zochlinski, H., and H. Mower. 1981. The mutagenic properties of hexachloroacetone in short-term bacterial mutagen assay systems. Mutat. Res. 89:137-144. Haloacetonitriles Bull, R. J. 1982. Experimental methods for evaluating the health risks associated with organic chemicals in drinking water. Toxicol. Environ. Chem. 6:1-17. Bull, R. J., and M. Robinson. 1984. Carcinogenic activity of haloacetonitrile and haloac- etone derivatives in the mouse skin and lung. Pp. 221-227 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Bull, R. J., J. R. Meier, M. Robinson, H. P. Ringhand, R. D. Laurie, and J. A. Stober. 1985. Evaluation of mutagenic and carcinogenic properties of brominated and chlorinated acetonitriles: By-products of chlorination. Fund. Appl. Toxicol. 5:1065-1074. Daniel, F. B., M. A. Pereira, E. H. C. Lin, D. L. Haas, and J. K. Mattox. 1983. Genotoxicity of haloacetonitriles. Toxicologist 3:37. (Abstract 148) Hayes, J. R., L. W. Condie, Jr., and J. F. Borzelleca. 1986. Toxicology of haloacetonitriles. Environ. Health Perspect. 69:183-202. Kraybill, H. F. 1980. Evaluation of public health aspects of carcinogenic/mutagenic bio- refractories in drinking water. Prev. Med. 9:212-218. Kraybill, H. F. 1983. Assessment of human exposure and health risk to environmental contaminants in the atmosphere and water with special reference to cancer. J. Environ. Sci. Health C1(2):175-232. Meier, J. R., K. L. Baker, R. J. Bull, and M. R. Robinson. 1983. Mutagenicity and tumor initiating activity of halogenated acetonitriles. Environ. Mutagen. 5:447-448. (Abstract Ea- 12) Meier, J. R., R. J. Bull, J. A. Stober, and M. C. Cimino. 1985. Evaluation of chemicals used for drinking water disinfection for product~on of chromosomal damage and sperm- head abnormalities in mice. Environ. Mutagen. 7:201-211. Pereira, M. A., L.-H. C. Lin, and J. K. Mattox. 1984. Haloacetonitrile excretion as thiocyanam and inh~bition of dimethylnitrosamine demethylase: A proposed metabolic scheme. J. Toxicol. Environ. Health 13:633-641. Roby, M. R., S. Carle, M. A. Pereira, and D. E. Carter. 1986. Excretion and tissue disposition of dichloroacetonitrile in rats and mice. Environ. Health Perspect. 69:215- 220. Simmon, V. F., K. Kauhanen, and R. G. Tardiff. 1977. Mutagenic activity of chemicals identified in drinking water. Pp. 249-258 in D. Scott, B. A. Bridges, and F. H. Sobels, eds. Progress in Genetic Toxicology. Elsevier/North-Holland, New York. Smith, M. K., H. Zenick, and E. L. George. 1986. Reproductive toxicology of disinfection by-products. Environ. Health Perspect. 69:177-182. Valencia, R., J. M. Mason, R. C. Woodruff, and S. Zimmering. 1985. Chemical muta- genesis testing in Drosophila. III. Results of 48 coded compounds tested for the National Toxicology Program. Environ. Mutagen. 7:325-348. ChIoropicrin ACGIH (American Conference of Governmental Industrial Hygienists, Inc.). 1986. Chlo- ropicrin. P. 134 in Documentation of the Threshold Limit Values and Biological Exposure

Selected Disinfectants and By-Products 187 Indices, 5th ed. American Conference of Governmental Industrial Hygienists, Inc., Cincinnati, Ohio. Bruchet, A., Y. Tsutsumi, J. P. Duguet, and J. Mallevialle. 1985. Characterization of total halogenated compounds during various water treatment processes. Pp. 1,165-1,184 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Buckley, L. A., X. Z. Jiang, R. A. James, K. T. Morgan, and C. S. Barrow. 1984. Respiratory tract lesions induced by sensory irritants at the RD50 concentration. Toxicol. Appl. Pharmacol. 74:417-429. Castro, C. E., and N. O. Belser. 1981. Photohydrolysis of methyl bromide and chloropicrin. J. Agric. Food Chem. 29: 1005-1008. Castro, C. E., R. S. Wade, and N. O. Belser. 1983. Biodehalogenation. The metabolism of chloropicrin by Pseudomonas sp. J. Agric. Food Chem. 31: 1184- 1187. Castro, C. E., R. S. Wade, and N. O. Belser. 1985. Biodehalogenation: Reactions of cytochrome P-450 with polyhalomethanes. Biochemistry 24:204-210. de Greef, E., J. C. Morris, C. F. van Kreijl, and C. F. H. Morra. 1980. Health effects in the chemical oxidation of polluted waters. Pp. 913-924 in R. L. Jolley, W. A. Brungs, R. B. Cumming, and V. A. Jacobs, eds. Water Chlorination: Environmental Impact and Health Effects, Vol. 3. Ann Arbor Science, Ann Arbor, Mich. Duguet, J. P., Y. Tsutsumi, A. Bruchet, and J. Mallevialle. 1985. Chloropicrin in potable water: Conditions of formation and production during treatment processes. Pp. 1201- 1213 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Flury, F., and F. Zernik. 1931. Schadliche Gase: Dampfe, Nebel, Rauch- und Staubarten. Julius Spnnger, Berlin. 637 pp. Ghosal, D., I.-S. You, D. K. Chatterjee, and A. M. Chakrabarty. 1985. Microbial deg- radation of halogenated compounds. Science 228:135-142. Griesemer, R. A., and C. Cueto, Jr. 1980. Toward a classification scheme for degrees of experimental evidence for the carcinogenicity of chemicals for animals. Pp. 259-281 in R. Montesano, H. Bartsch, and L. Tomatis, eds. Molecular and Cellular Aspects of Carcinogen Screening Tests. IARC Scientific Publications No. 27. International Agency for Research on Cancer, Lyon, France. Kane, L. E., C. S. Barrow, and Y. Alarie. 1979. A short-term test to predict acceptable levels of exposure to airborne sensory irritants. Am. Ind. Hyg. Assoc. J. 40:207-229. Kvetensky, J., S. Cyprich, E. Hanzen, and V. Hybl. 1979. Zkusenosti z hronadnych otrav chloreni a chlorpikrinem. Voj. Zdravotn. Listy 48: 115- 117. Moilanen, K. W., D. G. Crosby, J. R. Humphrey, and J. W. Giles. 1978. Vapor-phase photodecomposition of chlorop~tcrin (trichloronitromethane). Tetrahedron 34:3345-3349. Moriya, M., T. Ohta, K. Watanabe, T. Miyazawa, K. Kato, and Y. Shirasu. 1983. Further mutagenicity studies on pesticides in bacterial reversion assay systems. Mutat. Res. 116:185-216. NCI (National Cancer Institute). 1978. Bioassay of Chloropicrin for Possible Carcinogen- icity. National Cancer Institute Carcinogenesis Technical Report Series No. 65. DHEW Publication No. (NIH) 78-1315. National Cancer Institute, U.S. Department of Health, Education, and Welfare, Bethesda, Md. (82 pp.) Pitt, M. J. 1982. A vapour hazard index for volatile chemicals. Chem. Ind. 1982(20):804- 806.

188 DRINKING WATER AND HEALTH Prentiss, A. M. 1937. Chemicals in War: A Treatise on Chemical Warfare. McGraw-Hill, New York. 739 pp. Sayato, Y., K. Nakamuro, and S. Matsui. 1982. Studies on mechanism of volatile chlo- rinated organic compound formation (III) Mechanism of formation of chloroform and chloropicrin by chlorination of humic acid. Suishitsu Odaku Kenkyu 5:127-134. (in Japanese; English abstract) Shirasu, Y., M. Moriya, H. Tezuka, S. Teramoto, T. Ohta, and T. Inoue. 1982. Muta- genicity screening studies on pesticides. Pp. 331-335 in T. Sugimura, S. Kondo, and H. Takebe, eds. Environmental Mutagens and Carcinogens. Proceedings of the 3rd International Conference on Environmental Mutagens, Tokyo, Mishima and Kyoto, Sep- tember 21-27, 1981. Alan Liss, New York. Shirasu, Y., M. Moriya, H. Tezuka, S. Teramoto, T. Ohta, and T. Inoue. 1984. Muta- genicity of pesticides. Pp. 617-624 in E. H. Y. Chu and W. M. Generoso, eds. Mutation, Cancer, and Malformation. Plenum, New York. Stokinger, H. E. 1982. Aliphatic nitro compounds, nitrates, nitrites. Pp. 4141-4208 in G. D. Clayton and F. E. Clayton, eds. Patty's Industrial Hygiene and Toxicology, 3rd rev. ea., Vol. 2C. John Wiley, New York. Tatken, R. L., and R. J. Lewis, Sr., eds. 1983. Registry of Toxic Effects of Chemical Substances. 1981-82 ea., Vol. 2. National Institute for Occupational Safety and Health, U.S. Department of Health and Human Services, Cincinnati, Ohio. Valencia, R., J. M. Mason, R. C. Woodruff, and S. Zimmering. 1985. Chemical muta- genesis testing in Drosophila. III. Results of 48 coded compounds tested for the National Toxicology Program. Environ. Mutagen. 7:325-348. ChIorophenols Borzelleca, J. F., J. R. Hayes, L. W. Condie, and J. L. Egle, Jr. 1985a. Acute and subchronic toxicity of 2,4-dichlorophenol in CD-1 mice. Fund. Appl. Toxicol. 5:478- 486. Borzelleca, J. F., L. W. Condie, and J. R. Hayes. 1985. Toxicological evaluation of selected chlorinated phenols. Pp. 331-343 in R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr., and V. A. Jacobs, eds. Water Chlorination: Chemistry, Environmental Impact and Health Effects, Vol. 5. Lewis Publishers, Chelsea, Mich. Deichmann, W. 1943. The toxicity of chlorophenols for rats. Fed. Proc. 2:76-77. Kobayashi, S., S. Toida, H. Kawamura, H. S. Chang, T. Fukuda, and K. Kawaguchi. 1972. Chronic toxicity of 2,4-dichlorophenol in mice: A simple design for the toxicity of residual metabolites of pesticides. Toho Igakkai Zasshi (J. Med. Soc. Toho, Japan) 19:356-362. (in Japanese; English summary) NRC (National Research Council). 1977. Drinking Water and Health. National Academy of Sciences, Washington, D.C. 939 pp. NRC (National Research Council). 1982. Drinking Water and Health. Vol. 4. National Academy Press, Washington, D.C. 299 pp. Probst, G. S., R. E. McMahon, L. E. Hill, C. Z. Thompson, J. K. Epp, and S. B. Neal. 1981. Chemically-induced unscheduled DNA synthesis in primary rat hepatocyte cultures: A comparison with bacterial mutagenicity using 218 compounds. Environ. Mutagen. 3:11-32. Rapson, W. H., M. A. Nazar, and V. V. Butsky. 1980. Mutagenicity produced by aqueous chlorination of organic compounds. Bull. Environ. Contam. Toxicol. 24:590-596.

Selected Disinfectants and By-Products 189 Seyler, D. E., J. M. East, L. W. Condie, and J. F. Borzelleca. 1984. The use of in vitro methods for assessing reproductive toxicity, dichlorophenols. Toxicol. Lett. 20:309- 315. Vernot, E. H., J. D. MacEwen, C. C. Haun, and E. R. Kinkead. 1977. Acute toxicity and skin corrosion data for some organic and inorganic compounds and aquaeous so- lutions. Toxicol. Appl. Pharmacol. 42:417-423.