In living cells reactive oxygen species (ROS) are formed continuously as a consequence of metabolic and other biochemical reactions as well as external factors. Some ROS have important physiological functions. Thus, antioxidant defense systems cannot provide complete protection from noxious effects of ROS. These include oxidative damage to DNA, which experimental studies in animals and in vitro have suggested are an important factor in carcinogenesis. Despite extensive repair oxidatively modified DNA is abundant in human tissues, in particular in tumors, i.e., in terms of 1-200 modified nucleosides per 10(5) intact nucleosides. The damaged nucleosides accumulate with age in both nuclear and mitochondrial DNA. The products of repair of these lesions are excreted into the urine in amounts corresponding to a damage rate of up to 10(4) modifications in each cell every day. The most abundant of these lesions, 8-oxo-7,8-dihydro-2'-deoxyguanosine (8-oxodG), is also the most mutagenic, resulting in GT transversions which are frequently found in tumor relevant genes. A series of other oxidative modifications of base and sugar residues occur frequently in DNA, but they are less well studied and their biological significance less apparent. The biomarkers for study of oxidative DNA damage in humans include urinary excretion of oxidized nucleosides and bases as repair products and modifications in DNA isolated from target tissue or surrogate cells, such as lymphocytes. These biomarkers reflect the rate of damage and the balance between the damage and repair rate, respectively. By means of biomarkers a number of important factors have been studied in humans. Ionizing radiation, a carcinogenic and pure source of ROS, induced both urinary and leukocyte biomarkers of oxidative DNA damage. Tobacco smoking, another carcinogenic source of ROS, increased the oxidative DNA damage rate by 35-50% estimated from the urinary excretion of 8-oxodG, and the level of 8-oxodG in leukocytes by 20-50%. The main endogenous source of ROS, the oxygen consumption, showed a close correlation with the 8-oxodG excretion rate although moderate exercise appeared to have no immediate effect. So far, cross-sectional study of diet composition and intervention studies, including energy restriction and antioxidant supplements, have generally failed to show an influence on the oxidative DNA modification. However, a diet rich of Brussels sprouts reduced the oxidative DNA damage rate, estimated by the urinary excretion of 8-oxodG, and the intake of vitamin C was a determinant for the level of 8-oxodG in sperm DNA. A low-fat diet reduced another marker of oxidative DNA damage in leukocytes. In patients with diseases associated with a mechanistically based increased risk of cancer, including Fanconi anemia, chronic hepatitis, cystic fibrosis, and various autoimmune diseases, the biomarker studies indicate an increased rate of oxidative DNA damage or in some instances deficient repair. Human studies support the experimentally based notion of oxidative DNA damage as an important mutagenic and apparently carcinogenic factor. However, the proof of a causal relationship in humans is still lacking. This could possibly be supported by demonstration of the rate of oxidative DNA damage as an independent risk factor for cancer in a prospective study of biobank material using a nested case control design. In addition, oxidative damage may be important for the aging process, particularly with respect to mitochondrial DNA and the pathogenesis of inflammatory diseases.