The following chemicals resulting from chlorination of water supplies have been evaluated in the Guidelines: free chlorine (HOCI + OCI ), trihalomethanes, chlorinated acetic acids, halogenated acetonitriles, chloral hydrate (trichloroacetaldehyde), chlorophenols, and MX (3chloro-4-dichloromethyl-5-hydroxy-2(5H)-furanone). GVs have bcen recommended for 12 specific chemicals.

For countries wishing to control DBPS, it may not be necessary to establish standards for all DBPs for which GVs have been proposed. The trihalomethanes (THMs), of which chloroform is the major component, are likely to be the primary DBPS. In some water supplies the chlorinated acetic acids will also be important.

Other compounds that have been identified in chlorinated drinking water include chloro-alkenes, chloro-ketones, organic/inorganic chloramines, chloropicrin, unchlorinated carboxylic acids, and a variety of halo-alkanes, and chloro-aldehydes/furanones, in addition to those evaluated in the Guidelines. Main impurities in chlorine gas and liquid that are relevant to the nature of chlorination products are carbon tetrachloride and bromine (IARC 1991, Bull and Kopfler 1991).

Free Chlorine. In a carcinogenicity study in rats, chlorine in drinking water was administered at dose levels of 0, 7, or 15 mg/kg of body weight (bw)/day. The NOAEL was 15 mg/kg bw/day based on absence of toxicity at this level. Applying an uncertainty factor of 100 to this NOAEL gives a TDI of 0. 15 mg/ kg bw/day. This value is conservative since the NOAEL was the highest dose tested.

The major source of exposure to chlorine is drinking water. Therefore, 100% of theTDI was allocated to drinking water, giving a health-based GV of 5 mg/ L. lt should be noted, however, that most individuals are able to taste chlorine at concentrations below 5 mg/L, and some at levels as low as 0.3 mg/L. The GV of 5 mg/L would be unacceptable to most consumers because of taste and odor problems.

IARC has concluded that there is inadequate evidence for the carcinogenicity of hypochlorite in experimental animals, and no data were available as to its carcinogenicity to humans (Group 3).

Trihalomethanes. Because THMs usually occur together, it has been the general practice to consider total THMs as a group, and a number of countries have set guidelines or standards on this basis, ranging from 0.025 to 0.25 mg/L.

In the previous edition of the Guidelines, a GV of 0.03 mg/L was recommended for chloroform only, corresponding to an excess lifetime cancer risk of 10-5 . Few data existed for the remaining THMs. In the 1993 edition of the Guidelines, GVs were established for all four THMs. The four compounds are basically similar in toxicological action on the liver and kidney, but authorities wishing to establish a single GV for total THMs should not simply add the GVs for the individual compounds. Instead, the following approach is recommended:

Chloroform is the predominant THM, and IARC has classified chloroform in Group 2B as a possible human carcinogen. There was considerable discussion at the review group meeting as to whether chloroform is genotoxic or induces tumors through a nongenotoxic mechanism. The recommended GV of 0.2 mg/L for a 10-5 excess cancer risk is conservative and is based on extrapolation of observed kidney tumors in male rats exposed to chloroform in drinking water for 2 years. If the GV were to be established on the basis of a threshold concept, a 7.5-year study in beagle dogs identified an LOAEL of 15 mg/kg bw/day, based on increased frequency of hepatic fatty cysts and increased serum ensyme levels. Applying an uncertainty factor of 1000 to this LOAEL gives a TDI of 0.015 mg/kg bw/day. Allocatine 50% of the TDI to drinking water results in a GV of 0.2 mg/L. Chloroform concentrations of 0.2 mg/L. or higher are not uncommon in chlorinated drinking water.

The second most predominant THM is bromodichloromethane. This compound also occurs naturally in the tissue of a number of marine algae. IARC has classified bromodichloromethane in Group 2B, possibly carcinogenic to humans. In several carcinogenicity studies, bromodichloromethane caused an increase in tumors of the kidney in male mice, the liver in female mice, and the kidney and large intestine in male and female rats. The GV was therefore derived using the linearized multistage model and a value of 0.06 mg/L obtained for an excess cancer risk of 10-5 . Bromodichloromethane seldom occurs at this level; concentrations in treated drinking water typically range from 0.001 to 0.05 mg/L.

Dibromochloromethane also occurs naturally in marine algae and is released to sea water and air. In a 90-day feeding study in rats at dibromochloromethane doses of 0, 11, 21, 40, 90, or 180 mg/kg bw/day, the NOAEL was 21 mg/kg bw/day, based on hepatic vacuolization at 40 mg/kg bw/day and above. Using an uncertainty factor of 1000 (an extra 10 for the short duration of the study) and allocating 20% of the resulting TDI to drinking water results in a GV of 0. 1 mg/L. The 20% allocation to drinking water (rather than 50% or 100%) is rather conservative and was made to take into account possible exposure from air and food. Concentrations of dibromochloromethane in chlorinated drinking water are usually well below the GV. IARC has classified dibromochloromethane in Group 3, not classifiable as to its carcinogenicity to humans.

In a 90-day feeding study in rats at doses of bromoform of 0, 9, 18, 35, 70, or 140 mg/kg bw/day, the NOAEL was 18 mg/kg bw/day, based on hepatocellular vacuolization at 35 mglkg bw/day and above. A GV of 0. 1 mg/ L was derived for bromoform based on a TDI of 0.0 1 8 mg/kg bw/day and with an allocation of 20% of the TDI to drinking water. Concentrations of bromoform in drinking water seldom exceed 0.01 mg/L. IARC has classified bromoform in Group 3, not classified as to its carcinogenicity to humans.