Water Quality Standards

Summary information extracted from: Guidelines for drinking-water quality, 2nd ed. - Vol. 1. Recommendations. - Geneva, World Health Organization, 1993. pp. 64-65.

Benzene is used principally in the production of other organic chemicals. It is present in petrol, and vehicular emissions constitute the main source of benzene in the environment. Benzene may be introduced into water by industrial effluents and atmospheric pollution. Concentrations in drinking-water are generally less than 5 µg/litre.

Acute exposure of humans to high concentrations of benzene primarily affects the central nervous system. At lower concentrations, benzene is toxic to the haematopoietic system, causing a continuum of haematological changes, including leukaemia. Because it is carcinogenic to humans, IARC has classified benzene in Group 1.

Haematological abnormalities similar to those observed in humans have been observed in animal species exposed to benzene. In animal studies, benzene was shown to be carcinogenic following both inhalation and ingestion. It induced several types of tumours in both rats and mice in a 2-year carcinogenesis bioassay by gavage in corn oil. Benzene has not been found to be mutagenic in bacterial assays but has been shown to cause chromosomal aberrations in vivo in a number of species, including humans, and to be positive in the mouse micronucleus test.

Because of the unequivocal evidence of the carcinogenicity of benzene in humans and laboratory animals and its documented chromosomal effects, quantitative risk extrapolation was used to calculate lifetime cancer risks. Based on a risk estimate using data on leukaemia from epidemiological studies involving inhalation exposure, it was calculated that a drinking-water concentration of 10 µg/litre was associated with an excess lifetime cancer risk of 10-5.

As data on the carcinogenic risk to humans following ingestion of benzene are not available, risk estimates were also carried out on the basis of the 2-year gavage study in rats and mice. The robust linear extrapolation model was used because there was a statistical lack of fit of some of the data with the linearized multistage model. The estimated range of concentrations in drinking-water corresponding to an excess lifetime cancer risk of 10-5, based on leukaemia and lymphomas in female mice and oral cavity squamous cell carcinomas in male rats, is 10–80 µg/litre. The lower end of this estimate corresponds to the estimate derived from epidemiological data, which formed the basis for the previous guideline value of 10 µg/litre associated with a 10-5 excess lifetime cancer risk. This guideline value of 10 µg/litre, for a 10-5 excess cancer risk, is therefore retained.