In regulatory toxicology, the dose-response relationship between occupational exposure and biomarkers is of importance in setting threshold values. We analyzed the relationships between occupational exposure to polycyclic aromatic hydrocarbons (PAH) and various biomarkers of internal exposure and DNA damage with data from 284 highly exposed male workers. Personal exposure to phenanthrene and other PAHs was measured during shift and correlated with the sum of 1À, 2+9À, 3À, and 4-hydroxyphenanthrenes in post-shift urine. PAHs and hydroxyphenanthrenes were associated with DNA damage assessed in WBC as 8-oxo-7,8-dihydro-2 ¶-deoxyguanosine/10 6 dGuo and strand breaks by Comet assay as Olive tail moment. Hydroxyphenanthrenes correlated with phenanthrene (Spearman r s = 0.70; P < 0.0001). No correlations could be found between strand breaks and exposure (r s = 0.01, P < 0.0001 for PAHs; r s = À0.03, P = 0.68 for hydroxyphenanthrenes). Correlations with 8-oxo-7,8-dihydro-2 ¶-deoxyguanosine/10 6 dGuo were weakly negative (r s = À0.22, P = 0.004 for PAHs) or flat (r s = À0.08, P = 0.31 for hydroxyphenanthrenes). Linear splines were applied to assess the relationships between the log-transformed variables. All regression models were adjusted for smoking and type of industry. For hydroxyphenanthrenes, 51.7% of the variance could be explained by phenanthrene and other predictors. Up to 0.77 Mg/m 3 phenanthrene, no association could be found with hydroxyphenanthrenes. Above that point, hydroxyphenanthrenes increased by a factor of 1.47 under a doubling of phenanthrene exposure (slope, 0.56; 95% confidence interval, 0.47-0.64). Hydroxyphenanthrenes may be recommended as biomarker of occupational PAH exposure, whereas biomarkers of DNA damage in blood did not show a dose-response relation to PAH exposure. (Cancer Epidemiol Biomarkers Prev 2007;16(9):1863 -73)