Evaluation of urinary biomarkers of exposure to benzene: correlation with blood benzene and influence of confounding factors

trans,trans-Muconic acid (t,t-MA) is generally considered as a useful biomarker of exposure to benzene. However, because of its lack of specificity, concerns about its value at low level of exposure have recently been raised. The aim of this study was (a) to compare t,t-MA, S-phenylmercapturic acid (SPMA) and benzene (B-U) as urinary biomarkers of exposure to low levels of benzene in petrochemical workers and, (b) to evaluate the influence of sorbic acid (SA) and genetic polymorphisms of biotransformation enzymes on the excretion of these biomarkers. A total of 110 workers (including 24 smokers; 2-10 cigarettes/day) accepted to take part in the study. To assess external exposure to benzene, air samples were collected during the whole working period by a passive sampling device attached close to the breathing zone of 98 workers. Benzene was measured in blood (B-B) samples taken at the end of the shift, and was considered as the reference marker of internal dose. Urine was collected at the end of the shift for the determination of B-U, SPMA, t,t-MA, SA and creatinine (cr). B-U and B-B were determined by head-space/GC-MS, SPMA and SA by LC-MS, t,t-MA by HPLC-UV. Most (89%) personal measurements of airborne benzene were below the limit of detection (0.1 ppm); B-B ranged from < 0.10 to 13.58 mu g/l (median 0.405 mu g/l). The median (range) concentrations of the urinary biomarkers were as follows: B-U 0.27 mu g/l (< 0.10-5.35), t,t-MA 0.060 mg/l (< 0.02-0.92), SPMA 1.40 mu g/l (0.20-14.70). Urinary SA concentrations ranged between < 3 and 2,211 mu g/l (median 28.00). Benzene concentration in blood and in urine as well as SPMA, but not t,t-MA, were significantly higher in smokers than in non-smokers. The best correlation between B-B and urinary biomarkers of exposure were obtained with benzene in urine (mu g/l r = 0.514, P < 0.001; mu g/g cr r = 0.478, P < 0.001) and SPMA (mu g/l r = 0.495, P < 0.001; mu g/g cr r = 0.426, P < 0.001) followed by t,t-MA (mg/l r = 0.363, P < 0.001; mg/g cr r = 0.300, P = 0.002). SA and t,t-MA were highly correlated (r = 0.618, P < 0.001; corrected for cr r = 0.637). Multiple linear regression showed that the variation of t,t-MA was mostly explained by SA concentration in urine (30% of the explained variance) and by B-B (12%). Variations of SPMA and B-U were explained for 18 and 29%, respectively, by B-B. About 30% of the variance of B-U and SPMA were explained by B-B and smoking status. Genetic polymorphisms for biotransformation enzymes (CYP2E1, EPHX1, GSTM1, GSTT1, GSTP1) did not significantly influence the urinary concentration of any of the three urinary biomarkers at this low level of exposure. At low levels of benzene exposure (< 0.1 ppm), (1) t,t-MA is definitely not a reliable biomarker of benzene exposure because of the clear influence of SA originating from food, (2) SPMA and B-U reflect the internal dose with almost similar accuracies, (3) genetically based inter-individual variability in urinary excretion of biomarkers seems negligible. It remains to assess which biomarker is the best predictor of health effects.