Surface water disinfection can lead to the formation of mutagenic/carcinogenic by-products derived from reactions with naturally occurring inorganic compounds. We investigated the feasibility and potential usefulness of an integrated approach to genotoxicity analysis of drinking water. The approach employed the Comet and micronucleus (MN) assays to evaluate the DNA and chromosomal damage produced by water extracts in human blood cells. Surface water samples from Lago Trasimeno (Italy) were collected in different seasons (July 2000, October 2000, February 2001, and June 2001), and samples were disinfected with sodium hypochloride (NaClO), chlorine dioxide (ClO(2)), or peracetic acid (PAA). Extracts of untreated and treated water were incubated with primary human leukocytes. The Comet assay revealed both strong seasonal variations and differences between samples processed by the three disinfection protocols. The three disinfectants increased the genotoxicity of the water collected in July 2000 and October 2000, with PAA producing the greatest amount of DNA damage. Extracts of raw water collected in February 2001 produced so much DNA damage that the relative genotoxic potentials of the three disinfectants could not be evaluated. No increase in MN frequency was detected in any of the samples. The multi-endpoint MN assay indicated, however, that our study samples (especially the sample collected in the February 2001) were cytotoxic. We conclude that this integrated approach to genotoxicity assessment may be useful both for the quality control of raw drinking water and to help compare the potential health risks associated with alternative disinfection processes.