Sixty-three male cigarette smokers were entered into a cross-sectional study to determine whether inverse associations existed between polycyclic aromatic hydrocarbon (PAH)-DNA adduct levels and intake/serum levels of vitamin A, vitamin C and vitamin E. Associations between PAH-DNA adducts and intakes of carotene, as well as serum levels of beta-carotene, were also determined. Fasting blood samples were collected for assays of PAH-DNA adducts in circulating mononuclear cells, plasma cotinine and serum levels of vitamin A, beta-carotene, vitamin C and vitamin E. Since genetic deficiency in the detoxifying enzyme glutathione S-transferase M1 (GSTM1) has been associated with increased risk of lung cancer, GSTM1 genotype was also determined. Analysis of PAH-DNA adducts by competitive enzyme-linked immunosorbent assay (ELISA) indicated that 70% of the subjects had detectable adducts, with a mean of 4.38 adducts/10(8) nucleotides (range 1.00-24.1/10(8)). Pearson's method was utilized to determine whether any associations existed between the various host variables and PAH-DNA adducts. Previously, no significant associations were found between PAH-DNA adducts and cigarettes smoked/day, pack-years, daily/life-time tar exposures or plasma cotinine levels (Santella et al., Carcinogenesis, 13, 2041-2045, 1992). PAH-DNA adducts were inversely associated with serum cholesterol-adjusted vitamin E levels (r = -0.25, P < or = 0.05) and with smoking-adjusted vitamin C serum levels (r = -0.22, P < or = 0.09). Stratification by GSTM1 genotype indicated that these associations were limited to subjects with the null genotype. The relationship between adducts and serum cholesterol-adjusted vitamin E was significant in those of the null genotype (r = -0.38, P < or = 0.04), but not in those with the gene present (r = -0.12, P = 0.5). Similarly, for smoking-adjusted vitamin C, the relationship with adducts was stronger in subjects with the null genotype (r = -0.35, P < or = 0.06) than in those with GSTM1 present (r = -0.05, P = 0.77). These results are consistent with findings of prior epidemiological studies identifying significant inverse associations between anti-oxidant micronutrient status or GSTM1 genotype and the incidence of lung cancer. Additional studies should be conducted to confirm a possible role for vitamin E in PAH-DNA adduct formation and to explore further the possible roles of vitamin A, beta-carotene and vitamin C in modulating adduct formation and lung cancer risk.