Carcinogen-DNA adducts may represent an intermediate end-point in the carcinogenic cascade and may reflect exposure to chemical carcinogens, as well as susceptibility and, ultimately, cancer risk. Interindividual variability in activity of enzymes involved in the metabolism of polycyclic aromatic hydrocarbons to mutagenic diol epoxides may predict adduct levels and, indirectly, lung cancer risk. Using 32P-postlabeling methods, the levels of bulky DNA adducts were determined in macroscopically normal bronchial tissues obtained from resected lobes of 143 Hungarian patients with lung malignancy and other pulmonary conditions. DNA from normal tissue was also evaluated for polymorphisms in cytochrome P450 2C9 (CYP2C9) at two sites, codons 144 (Arg/Cys) and 359 (Ile/Leu), for glutathione S-transferase P1 (GSTP1) at codon 105 and for NAD(P)H:quinone oxidoreductase (NQO1) at codon 187 (Pro/Ser). Using the Mann-Whitney U-test and analysis of variance, levels of adducts were evaluated in relation to variant genotypes, separately for smokers and non-smokers. As previously reported, bulky DNA adduct levels in smokers (n = 104) were estimated to be 54% higher than in non-smokers (n = 39) (8.6 +/- 4.2 versus 5.6 +/- 3.3 per 10(8) nucleotides, respectively, P < 0.01). Adduct levels were 16-29% higher in individuals with the homozygous Ile359/Ile359 CYP2C9 allele than in those heterozygous for the variant allele (Ile359/Leu359) [8.8 +/- 4.3 (n = 84) versus 7.6 +/- 3.5 (n = 20) for smokers and 5.8 +/- 3.5 (n = 32) versus 4.5 +/- 1.3 (n = 7) for non-smokers], although differences were not statistically significant. There were no clear differences in adduct levels in relation to genotypes of NQO1 or GSTP1. Although numbers of patients in this study are large in relation to many studies of carcinogen-DNA adducts, it is still possible that significant differences were not noted for polymorphisms in xenobiotic metabolizing enzymes due to relatively small numbers in stratified data.