(a) Formation of halogenated disinfection by-products in drinking-water

The drinking-water disinfectant chlorine reacts with natural organic matter to produce halogenated disinfection by-products, and trihalomethanes and haloacetic acids are the two most prevalent groups of known specific by-products formed during disinfection of natural waters with chlorine-containing oxidizing compounds (Hua & Reckhow, 2007). These compounds are formed when drinking-water supplies containing natural organic matter (e.g. humic or fulvic acids) are disinfected with compounds such as chlorine gas, hypochlorous acid and hypochlorite (Huang et al., 2004). When bromide is present in the source water, it may be oxidized to hypobromous acid–hypobromite ion, which can react with organic matter to form brominated organic compounds. The reaction of brominated and/or chlorinated oxidizing agents with natural organic matter produces mixed brominated and chlorinated compounds. The relative amount of brominated haloacetates produced in chlorinated drinking-water is a function of the concentration of bromide in the source water and of the initial bromine/chlorine ratio. The relative amounts of disinfection by-products produced in drinking-water supplies are affected by the nature and concentration of the organic precursor materials, water temperature, pH, the type of disinfectant, the disinfectant dose and contact time (Liang & Singer, 2003; Huang et al., 2004). Treatment of natural waters with chloramine or chlorine dioxide produces haloacetic acids, but at levels substantially lower than those formed by free chlorine (Richardson et al., 2000; Hua & Reckhow, 2007). Because commonly used alternative disinfectants (ozone, chloramines and chlorine dioxide) produce lower levels of most haloacetic acids, many water utilities have switched from chlorination to these alternatives to meet the regulation limits in terms of disinfection by-products (Krasner et al., 2006; Richardson et al., 2007).

Data from the USA revealed that water-treatment systems that used chlorine dioxide had higher levels of nine haloacetic acids than those that used chlorine or chloramine only (McGuire et al., 2002). This is because the water-treatment systems that used chlorine dioxide also used chlorine or chloramines (mostly as post-disinfectants). Similarly to chloramines and chlorine dioxide, ozone used in water treatment is well known to lower the levels of haloacetic acids formed relative to chlorination (Richardson et al., 2007). However, when source waters contain elevated levels of natural bromide, the levels of brominated compounds were shown to increase when pre-ozone treatment was performed before chlorination (IPCS, 2000; Richardson et al., 2007).

According to IPCS (2000) and WHO (2008), the optimized use of combinations of disinfectants that function as primary and secondary disinfectants should allow further control of disinfection by-products. There is a trend towards combination/sequential use of disinfectants: ozone is used exclusively as a primary disinfectant; chloramines are used exclusively as a secondary disinfectant; and both chlorine and chlorine dioxide are used in either role.

According to WHO (2004), bromide ions occur naturally in surface water and groundwater; their levels exhibit seasonal fluctuations, and can also increase due to saltwater intrusion resulting from drought conditions or pollution (IPCS, 2000).

(b) Concentrations in drinking-water

A nationwide study of the occurrence of disinfection by-products in different geographical regions of the USA was conducted between October 2000 and April 2002 (Weinberg et al., 2002), in which samples were taken from 12 water-treatment plants that had different levels of source water quality and bromide and used the major disinfectants (chlorine, chloramines, ozone and chlorine dioxide). Concentrations of bromochloroacetic acid in finished water samples ranged from 1.3 to 18 μg/L.

Data from drinking-water supplies in the USA (EPA, 2000 cited in WHO 2004) indicated that bromochloroacetic acid was detected in groundwater and surface water distribution systems at mean concentrations of 1.47 and 3.61 μg/L, respectively.

In a survey of 20 drinking-waters prepared from different source waters in the Netherlands (Peters et al., 1991), haloacetic acids were found in all drinking-waters prepared from surface water, whereas they could not be detected in drinking-waters prepared from groundwater. The total concentrations of haloacetic acids were in the range of 0.5–14.7 μg/L (surface water only) with levels of bromochloroacetic acid ranging from 0.2 to 2.5 μg/L. The limit of detection of this study was 0.1 μg/L, and brominated acetic acids accounted for 65% of the total haloacetic acid concentration.

Bromochloroacetic acid was measured in water samples taken from a water-treatment plant in Barcelona (Spain) between November 1997 and March 1998 (Cancho et al., 1999). Haloacetic acids were rapidly formed during the pre-chlorination step, but their concentration did not increase during either sand filtration or the ozonation step. At these two stages, the concentration of total haloacetic acids represented 60% of the total trihalomethane levels. A significant decrease in total concentration of haloacetic acids was observed when ozonated water was passed through granular activated carbon filters, but the acids were formed again during post-chlorination, although at concentrations lower than those during the previous stages. The average total level of haloacetic acids was around 22 μg/L in tap-water (range, 11–32 μg/L). Bromochloroacetic acid was not detected in raw water, but was detected in pre-chlorinated water (mean, 8.8 μg/L; range, 6.4–15 μg/L), sand-filtered water (mean, 8.2 μg/L; range, 7.1–11 μg/L), ozonated water (mean, 9.1 μg/L; range, 8–11.4 μg/L), granulated activated carbon-filtered water (mean, 0.8 μg/L; range, not detected−3.2 μg/L) and post-chlorinated water (mean, 2.5 μg/L; range, 1–3.9 μg/L), i.e. water that was ready for consumption. [The limit of detection was not reported.]

Water samples were collected from 35 Finnish waterworks between January and October in 1994 and from three waterworks and distribution systems during different seasons in 1995 (Nissinen et al., 2002). Bromochloroacetic acid was detected at 32 of the 35 Finnish waterworks sampled in 1995 with concentrations between 0.3 and 19 μg/L. Levels at the other facilities were below the limit of quantitation of 0.2 μg/L. The concentration of six haloacetic acids, including bromochloracetic acid, exceeded that of trihalomethanes. Chlorinated drinking-waters originating from surface waters contained the highest concentration of haloacetic acids (108 μg/L). The lowest concentrations of disinfection by-products were measured from ozonated and/or activated carbon-filtered and chloraminated drinking-waters (20 μg/L). Higher concentrations of the six haloacetic acids were measured in summer than in winter [data not reported].

In the USA, finished waters from the Philadelphia (PA) Suburban Water Co., the Metropolitan Water District of Southern California, and utilities at the cities of Houston (TX) and Corpus Christi (TX) were collected at the point of entry to the water distribution system and analysed for the nine haloacetic acids (Cowman & Singer, 1996). These samples included waters with relatively low (Philadelphia), moderate (Houston), and high (Southern California, Corpus Christi) bromide concentrations. Several of the utilities (Houston, Southern California, Corpus Christi) added ammonia to their waters after chlorination to control disinfection by-product formation. Levels of bromochloroacetic acid were below the limit of detection [not reported] in the Philadelphia utility, where the bromide ion concentration was 50.6 μg/L. For the others utilities, where levels of bromide ion ranged from 72 to 412 μg/L, those of bromochloroacetic acid ranged from 4.68 to 10.8 μg/L.

Drinking-water was studied in Israel because its source water (the Sea of Galilee, a freshwater lake, also called Lake Kinereth) has among the highest natural levels of bromide in the world for surface water (2000 μg/L) and chlorine dioxide is used for disinfection at full-scale treatment plants (Richardson et al., 2003). Chlorine-containing disinfection by-products that are usually dominant under conditions of low levels of bromide (for chlorination and chloramine disinfection) — chloroform and dichloroacetic acid — were found at very low concentrations or not at all in these samples, with a shift to bromoform and dibromoacetic acid occurring under these conditions of high levels of bromide. Thus, the high bromide content in the source water had a major impact on the speciation of the disinfection by-products. Bromochloroacetic acid was detected at levels between 1 and 3.9 μg/L.

Between October 1994 and April 1996, a mean concentration of 0.6 μg/L bromochloroacetic acid was measured in the Santa Ana River (USA) downstream of a discharge point for highly treated municipal wastewater effluent (Ding et al., 1999).

A study was conducted in nine distribution systems of the greater area of Québec City (Province of Québec, Canada) (Legay et al., 2010). Nine individual haloacetic acids, including bromochloroacetic acid, were analysed during 2006–08, and concentrations were: mean, 25.3–115.2 μg/L; 25th percentile, 16.7–73.5 μg/L; 50th percentile, 23.0–113.5 μg/L; and 75th percentile, 31.6–145.1 μg/L.

In a study based on data from several European countries (Belgium, France, Germany, Italy, the Netherlands and Spain) and covering two decades (from 1980 to 2000; Palacios et al., 2000), the levels of organohalogenated compounds in surface and groundwaters after chlorination were evaluated. A mean concentration of 3.53 μg/L bromochloroacetic acid was measured in post-treatment surface water (range, not detected−13.7 μg/L), but was not detected in post-treatment groundwater from disinfection utilities [limit of detection not reported].

(c) Dietary exposure from drinking-water

To assess exposure to disinfection by-products through drinking-water, WHO uses a default consumption value of 2 L drinking-water per capita per day and a typical body weight (bw) of 60 kg (WHO, 2008). The underlying assumption is that of a total water consumption of 3 L per capita per day, including water present in food (WHO, 2003).

The mean concentrations and ranges of bromochloroacetic acid from all references available were used by the Working Group to assess dietary exposure in adults and infants (weighing 60 kg and 5 kg, respectively), assuming a consumption of 2 L and 0.75 L drinking-water, respectively, i.e. 33 mL/kg bw and 150 mL/kg bw, respectively (Table 1.1). The infant scenario (expressed in mL/kg bw) would correspond to the consumption of 9 L drinking-water per day in a 60-kg adult and therefore cover any possible scenario of physically active persons and increased temperature.

Table 1.1

Dietary exposure to bromochloroacetic acid from drinking-water.

Based on the available data on average concentrations of bromochloroacetic acid, dietary exposure through drinking-water in a standard 60-kg adult ranges from 0.02 to 0.08 μg/kg bw per day, and high observed concentration values would lead to a dietary exposure of 0.1–0.6 μg/kg bw per day. Similarly, dietary exposure through drinking-water in a 5-kg infant ranges from 0.1 to 0.4 μg/kg bw per day, and high observed concentration values would lead to a dietary exposure of 0.4–2.8 μg/kg bw per day (Table 1.1).

(d) Other dietary sources

No data on the levels of haloacetic acids in foods (other than drinking-water) could be identified. Extrapolations from values in drinking-water to values in food are difficult to achieve because the conditions of the chemical interactions, dosages, temperatures, contact times and especially the precursors differ considerably (FAO/WHO, 2009).

(a) Absorption, distribution and excretion

Dihaloacetates are rapidly absorbed from the gastrointestinal tract of rats after oral exposure (James et al., 1998; Schultz et al., 1999). The maximum blood concentration of bromochloroacetate in F344/N rats was reached 1.5 h after administration by gavage (Schultz et al., 1999).

Dihaloacetates exhibit low binding to rat plasma proteins: in the plasma of treated F344 rats, 93% of the measured bromochloroacetate was in the unbound fraction (Schultz et al., 1999).

The oral bioavailability of bromochloroacetate was reported to be 47% in male F344/N rats (Schultz et al., 1999). The lower bioavailability of bromochloroacetate compared with other dihaloacetates is due to a greater first-pass metabolism in the liver.

Elimination half-lives of dihaloacetates in the blood of male F344/N rats are less than 4 hours; the plasma half-life of bromochloroacetate after intravenous injection is approximately 45 minutes. Elimination of dihaloacetates occurs primarily by metabolism; after an intravenous dose of 500 μmol/kg bw [86.7 mg/kg bw] bromochloroacetate, excretion as the parent compound was less than 3% in the urine and less than 0.1% in the faeces. Bromine substitution of dihaloacetates increases the rate of metabolic clearance (Schultz et al., 1999).

(b) Metabolism

The metabolism of bromochloroacetic acid has been reviewed (NTP, 2009). Biotransformation of dihaloacetates to glyoxylate occurs primarily in the liver cytosol of rats and humans by a glutathione-dependent process (James et al., 1997) catalysed by glutathione S-transferase-zeta (GST-zeta) (Tong et al., 1998a).

During GST-zeta-mediated oxygenation of dihaloacetates to glyoxylate, glutathione is required but not consumed. GST-zeta-mediated biotransformation of dihaloacetates (Fig. 4.1) involves displacement of a halide by glutathione to form S-(α-halo-carboxymethyl)glutathione, hydrolysis of this intermediate to form S-(α-hydroxy-carboxymethyl)glutathione and elimination of glutathione to produce glyoxylate (Tong et al., 1998b). Among the brominated/chlorinated dihaloacetates, the relative rates of glyoxylate formation catalysed by purified GST-zeta are: bromochloroacetate > dichloroacetate > dibromoacetate (Austin et al., 1996). Glyoxylate can undergo transamination to glycine, decarboxylation to form carbon dioxide and oxidation to oxalate. Glyoxylate may induce toxicity by reacting covalently with proteins, e.g. N-terminal amino groups or lysine ε-amino groups (Anderson et al., 2004).

Fig. 4.1

Biotransformation of dihaloacetates.

Bromochloroacetic acid is a suicide substrate for GST-zeta; 12 hours after a single injection (0.30 mmol/kg bw), GST-zeta activity in the rat liver is reduced to 19% of that in controls (Anderson et al., 1999). Hydrolysis of S-(α-halocarboxymethyl)glutathione forms a hemithioacetal that eliminates glutathione and yields glyoxylate. Because this intermediate may inactivate GST-zeta by covalently binding to a nucleophilic site on the enzyme (Anderson et al., 1999; Wempe et al., 1999), its hydrolysis and GST-zeta inactivation are competing reactions. Recovery of GST-zeta activity occurs via de-novo synthesis of the protein. Because GST-zeta is identical to maleylacetoacetate isomerase, the enzyme that catalyses the penultimate step of the tyrosine degradation pathway, its loss by exposure to dihaloacetates leads to the accumulation of maleylacetoacetate and maleylacetone which may cause tissue damage by reacting with cellular nucleophiles (Ammini et al., 2003).

The elimination half-life of (–)-bromochloroacetic acid in male F344 rats is approximately sixfold shorter than that of (+)-bromochloroacetic acid, indicating that the rate of GST-zeta-catalysed metabolism of bromochloroacetic acid is much faster for the (–)-stereoisomer (Schultz & Sylvester, 2001). [The Working Group noted that the carcinogenicity studies in animals were performed using a racemic mixture of bromochloroacetic acid.] Because the metabolism of bromochloroacetic acid stereoisomers in naive and GST-zeta-depleted cytosol of rats is dependent on the presence of glutathione, Schultz and Sylvester (2001) suggested that an additional GST isoenzyme that is not inactivated by dihaloacetates might provide a minor contribution to the formation of glyoxylate in non-pretreated animals.

(c) Toxicokinetic models

No data were available to the Working Group.

(a) DNA adducts

Oxidative stress can result in oxidative damage to DNA, most commonly measured as increases in 8-hydroxydeoxyguanosine (8-OHdG) adducts. After acute oral administration of bromochloroacetic acid to male B6C3F₁ mice, a significant increase in 8-OHdG/deoxyguanosine ratios in liver nuclear DNA was observed (Austin et al., 1996). After administration of bromochloroacetic acid in the drinking-water to male B6C3F₁ mice at concentrations of 0.1, 0.5 or 2.0 g/L for 3–10 weeks, the 8-OHdG content in liver nuclear DNA was increased (Parrish et al., 1996). These findings demonstrate that bromochloroacetic acid causes oxidative stress/damage.

(b) DNA damage

Bromochloroacetic acid induced DNA damage in Chinese hamster ovary cells, as measured in the Comet assay (Plewa et al., 2010).

(c) Mutations

In two bacterial mutagenicity assays, bromochloroacetic acid gave positive results in Salmonella typhimurium strain TA100 regardless of the presence of a metabolic activation system. It was not mutagenic in strain TA98 or in Escherichia coli WP2 uvrA/pKM101 regardless of the presence of a metabolic activation system (NTP, 2009). Glyoxylate (a metabolite of dihaloacetate biotransformation) was mutagenic in S. typhimurium strains TA97, TA100 and TA104 in the absence of a metabolic activation system and in strain TA102 in the presence of a metabolic activation system (Sayato et al., 1987).

(d) Chromosomal effects

No increase in chromosomal damage (micronucleus formation in blood lymphocytes) was reported after administration of bromochloroacetic acid in the drinking-water to mice for 3 months (NTP, 2009).