Abstract
Arsenic is one of the few human carcinogens for which there is not yet a reliable animal cancer model. As such, the classification of arsenic as a carcinogen is based upon data derived from human epidemiologic studies. Although the mechanisms of action of arsenic as a toxic agent have been known for many years, the inability to produce cancer with arsenic in laboratory animals has confounded the operational characterisation of arsenic as initiator, promoter, complete carcinogen, or cocarcinogen for humans. Arsenic is clearly a genotoxic agent that induces chromosomal aberrations, micronuclei and sister chromatid exchange in mammalian cells as well as neoplastically transforms Syrian hamster embryo cells; however, it is not a classical point mutagen. This paper reviews some of the scientifically based issues relating to arsenic and risk assessment.