The objective of this study was to test the infiuence of genetic polymorphisms for metabolic enzymes (CYP2E1, mEH, GSTM1 and GSTT1) implicated in the biotransformation of styrene in humans on the interpretation of urinary biomarkers of exposure. Thirty workers from a fibreglass-reinforced plastics factory took part in the study. Ambient styrene concentration was determined during the whole workshift by passive sampling. Urine was collected at the end of the shift for the determination of mandelic acid (MA) and phenylglyoxylic acid (PGA) (major biotransformation pathway), N-acetyl-S-(1-phenyl-2-hydroxy)ethyl-L-cysteine (M1) and N-acetyl-S-(2-phenyl-2-hydroxy)ethyl-L-cysteine (M2) (minor metabolic pathway) and creatinine. The average airborne styrene concentration of 18.2 ppm (range: 0.9-68.9 ppm) was very close to the current threshold limit value (TLV-TWA) recently adjusted by ACGIH from 50 to 20 ppm. There was a better correlation between external and internal exposure as estimated by urinary MA + PGA (r=0.92; p<0.0001) compared with urinary M1 + M2 (r=0.74; p<0.0001). To investigate to what extent genetic polymorphisms in metabolic enzymes could explain interindividual variations observed in the concentration of urinary biomarkers related to a given external exposure, two 'metabolic indexes' (derived from the ratio between the sum of urinary metabolites for a specific pathway and ambient styrene concentration) were calculated for each worker and compared for different allelic combinations. Monovariate analyses showed that GSTM1 polymorphism was clearly the most significant parameter infiuencing urinary concentrations of mercapturic acids. Based on GSTM1 allelic status, two different biological exposure indexes (BEIs) for M1 + M2 in post-shift urinary samples corresponding to a 20 ppm styrene concentration are proposed (GSTM1null: 1330 µg g(-1) creatinine, GSTM1+: 2878 µg g(-1) creatinine). Multivariate regression analyses were also performed and revealed that the presence of the rare CYP2E1*1B allele linked to TaqI polymorphism (A1/A2) was associated with increased urinary concentrations of metabolites from both pathways. Two previously described polymorphisms for the EPHX gene were also tested but seemed not really relevant for interpretation of biomarkers. In conclusion, while CYP2E1 genotyping, particularly assessment of the CYP2E1*1B allelic status, is useful for a more accurate interpretation of the concentration of urinary biomarkers, GSTM1 genotyping is absolutely necessary when considering a biological monitoring programme based on determination of urinary mercapturic acids.