Urine samples from humans occupationally exposed to styrene, with mandelic acid levels ranging from 400 to 1145 mg/g creatinine and from 68 to 400 mg/g creati nine for high and low exposure group, respectively, were analysed for N3 adenine DNA adducts, namely. 3-(2-hydroxy-1-phenylethyl)adenine (N3 alpha A) and 3-(2-hydroxy-2-phenylethyl)adenine (N3 beta A). A sensitive LC-ESI-MSMS method was developed with the limit of quantification of 1 pg/mL for both analytes Peaks corresponding to N3 alpha A and/or N3 beta A were found in seven of nine end-of-shift samples of the high exposure group and in six of 19 end-of-shift samples of the low exposure group Concentration of N3 alpha A + N3 beta A amounted to 2.8 +/- 1.6 pg/mL(mean +/- S.D.; n = 9) and 1.8 +/- 1 3 pg/mL (mean +/- S.D.; n = 19) in the high and low exposure group, respectively Of other 10 samples taken the next morning after exposure, two contained low but quantifiable concentrations of N3 alpha A and none contained N3 beta A However, interfering peaks were detected also in some control urine samples. Out of 22 controls, six and two samples contained peaks co-eluting with N3 alpha A and N3 beta A, respectively. Therefore, the method used was found insufficiently specific to be applicable for biological monitoring. Comparing the excretion of N3 alpha A + N3 beta A to that reported previously in mice it can be estimated that at the same absorbed dose, humans excreted not more than 1/30 of the amount of adenine adducts excreted by mice. As a consequence, the damage to DNA caused by styrene 7,8-oxide (SO), a reactive metabolite of styrene, appears to be much lower in humans than in mice (C) 2010 Elsevier Ireland Ltd All rights reserved.