A gas chromatography-mass spectrometry (GC-MS) method was developed for measuring 1,2-dihydroxynaphthalene (1,2-DHN) and 1,4-dihydroxynaphthalene (1,4-DHN) in urine. The method involves enzymatic digestion of urinary conjugates to release the DHNs which were then analyzed as trimethylsilyl derivatives by GC-MS. For 1,2-DHN and 1,4-DHN, respectively, the assay limits of detection were 0.21 and 0.15 microg/l, the assay limits of quantitation were 0.69 and 0.44 microg/l, and the coefficients of variation were 14.7 and 10.9%. This method was successfully applied to determine urinary levels of 1,2-DHN and 1,4-DHN in coke workers (14 top workers and 13 side-bottom workers) and 21 matching control workers from the steel industry of northern China. The geometric mean (GM) levels of 1,2-DHN were approximately 100 and 30 times higher than those of 1,4-DHN in exposed and control subjects, respectively. The GM levels 1,2-DHN and 1,4-DHN were significantly higher for coke workers (1,2-DHN: top workers--552 microg/l, side-bottom workers--260 microg/l; 1,4-DHN: top workers--3.42 microg/l, side-bottom workers--3.56 microg/l) than for controls (1,2-DHN: 38.8 microg/l; 1,4-DHN: 1.21 microg/l) (p<or=0.0031). In each exposure category, levels of the DHNs were marginally greater in smokers than in nonsmokers (p=0.0646). Strong correlations were observed among 1,2-DHN and 1,4-DHN and previously measured urinary levels of naphthalene, 1-hydroxynaphthalene, and 2-hydroxynaphthalene in these subjects (rs>or=0.623; p<0.0001). Also, levels of 1,2-DHN were significantly correlated with those of serum albumin adducts of l,2-naphthoquinone (rs=0.492, p=0.0004). These results indicate that 1,2- and 1,4-DHN are good biomarkers for assessment of naphthalene exposure in coke workers. Since the DHNs are precursors of the naphthoquinones, which have been implicated as toxic products of naphthalene metabolism, measurements of urinary DHNs may have toxicological significance.