Exposure to polycyclic aromatic hydrocarbons (PAHs) in foundry workers has been evaluated by determination of benzo(a)pyrene-serum albumin adducts and urinary 1-hydroxypyrene. Benzo(a)pyrene binding to albumin and 1-hydroxypyrene were quantitatively measured by enzyme linked immunosorbent assay (ELISA) and reverse phase high performance liquid chromatography (HPLC), respectively. 70 male foundry workers and 68 matched controls were investigated. High and low exposure groups were defined from breathing zone hygienic samples, consisting of 16 PAH compounds in particulate and gaseous phase. Mean total PAH was 10.40 micrograms/m3 in the breathing zone, and mean dust adsorbed PAH was 0.15 microgram/m. All carcinogenic PAH was adsorbed to dust. Median benzo(a)pyrene-albumin adduct concentrations (10-90% percentiles) were similar in foundry workers (smokers 0.55 (0.27-1.00) and non-smokers 0.58 (0.17-1.15)) pmol/mg albumin and age matched controls (smokers 0.57 (0.16-1.45) and non-smokers 0.70 (0.19-1.55) pmol/mg albumin). Median 1-hydroxypyrene concentrations were significantly higher (P < 0.0001) in smoking and non-smoking foundry workers (0.022 (0.006-0.075) and 0.027 (0.006-0.164)) mumol/mol creatinine than in smoking and non-smoking controls (0 (0-0.022) and 0 (0-0.010) mumol/mol creatinine). Dose-response relations between total PAH, pyrene, carcinogenic PAHs, and 1-hydroxypyrene for smokers, and polycyclic aromatic hydrocarbons adsorbed to dust for non-smokers are suggested. Exposure to PAHs adsorbed to dust showed an additive effect. There was no correlation between the concentrations of 1-hydroxypyrene and benzo(a)pyrene-albumin adducts. The change in 1-hydroxypyrene over a weekend was also studied. Friday morning median 1-hydroxypyrene concentrations were significantly higher in both smokers and non-smokers (0.021 (0-0.075) and 0.027 (0.06-0.164)) mumol/mol creatinine than Monday morning median concentrations (0.007 (0-0.021) and 0.008 (0-0.021) mumol/mol creatinine). Smoking did not affect the concentrations of 1-hydroxypyrene or benzo(a)pyrene-albumin adducts. These data suggest that 1-hydroxypyrene is a sensitive biomarker for low dose PAH exposure. Exposure to PAHs may be aetiologically related to increased risk of lung cancer in foundry workers.