In contrast to acute or chronic dosing experiments with a single chemi
cal in animals, man is exposed to thousands of chemicals during a life
time. Each of these may act alone, additively, synergistically or anta
gonistically in terms of biological effects, but most current risk ass
essment procedures fail to recognize such interactions. In carcinogene
sis, a mutational process that is thought to occur through DNA damage
by endogenous and/or exogenous agents, a wide variety of host factors
is involved in disease outcome. These include absorption of chemicals,
their distribution, metabolism and excretion. In addition, once metab
olic activation has occurred, there is an array of protective mechanis
ms that cells have evolved to maintain DNA integrity, such as DNA repa
ir, genetic redundancy and programmed cell death. One approach to risk
assessment is to regard all DNA-damaging events as potentially leadin
g to cancer and to measure DNA damage as the biologically relevant end
point. The main method, if not the only method, presently available to
assay a wide range of DNA adducts is P-32-postlabelling. This method
has high sensitivity (limit of detection > 1 adduct per 10(10) nucleot
ides) and is capable of visualizing many different DNA adducts in a si
ngle analysis. Postlabelling is best suited for detecting hydrophobic
adducts-low molecular weight adducts usually need a preliminary separa
tion procedure prior to being postlabelled. This chromatographic proce
dure has been used to study DNA samples from human tissues of cigarett
e smokers, occupationally exposed groups and individuals living in pol
luted environments. Correlations have been found between the severity
of exposure and the level of DNA adducts detected for human samples. H
owever, most studies are single-time point studies, whereas for risk a
ssessment purposes it may be better to use more quantitative and repre
sentative measures of long-term exposure, for example the number of ad
ducts formed per annum. This article reviews methods of DNA adduct mea
surement, with particular reference to the P-32-postlabelling techniqu
e, which has been used to determine DNA adduct levels in populations e
xposed to complex mixtures. Copyright (C) 1996 Published by Elsevier S
cience Ltd