Objectives To test the association between occupational exposure to trichloroethylene (TCE) and risk of non Hodgkin lymphoma (NHL), we conducted a pooled analysis of four international case-control studies. Methods Studies were selected which included state-of-the art retrospective assessment of occupational exposure to TCE and histological information on lymphoma subtype. Overall, the pooled study population included 3788 NHL cases and 4279 controls. Summary indicators of exposure were harmonised across studies. We conducted unconditional logistic regression analysis to test the association between the harmonised TCE exposure estimates and NHL and its major subtypes, adjusting by age, gender, and study. Results Among the total study population, risk of follicular lymphoma, but not NHL overall or other subtypes, increased by probability (p = 0.02) and intensity level (p = 0.04) of TCE exposure. When the analysis was restricted to subjects most likely exposed to TCE, risk of NHL overall (p = 0.009), follicular lymphoma (p = 0.04), and chronic lymphocytic leukaemia (CLL) (p = 0.01) showed a linear increase by duration of exposure. No heterogeneity in NHL risk associated with high probability of exposure to TCE (all intensity levels combined) was detected. Conclusion With due caution because of several limitations, our pooled analysis supports the hypothesis of an increased risk of NHL, and particularly of the follicular lymphoma and CLL subtypes, associated with occupational exposure to TCE.

We compared the ability of the urinary excretion of trans,trans-muconic acid (t,t-MA), s-phenilmercapturic acid (s-PMA) and urinary benzene (U-Benz) to detect low level occupational and environmental exposure to benzene. Methods We monitored airborne benzene by personal air sampling, and U-Benz, s-PMA, t,t-MA and cotinine (U-Cotinine) in spot urine samples, collected at 8 am and 8 pm, in 32 oil refinery workers and 65 subjects, randomly selected among the general population of urban and suburban Cagliari, Italy. Results The median concentration of airborne benzene was 25.2 µg/m3 in oil refinery workers, and 8.5 µg/m3 in the general population subgroup. U-Benz in morning and evening samples was significantly more elevated among oil refinery workers than the general population subgroup (p = 0.012, and p = 7.4 x 10-7, respectively) and among current smokers compared to non-smokers (p = 5.2 x 10-8, and p = 5.2 x 10-5 respectively). Benzene biomarkers and their readings in the two sampling phases were well correlated to each other. The Spearman’s correlation coefficient with airborne benzene was significant for U-Benz in the evening sample, but not for t,t-MA and s-PMA in either sampling. Morning U-Cotinine excretion showed a good correlation with U-Benz in the morning and in the evening sampling (p < 0.001), and with s-PMA in the evening sample (p < 0.001), but not with t,t-MA in either samplings. t,t-MA in the evening sample was the only biomarker showing a moderate inverse correlation with BMI (p < 0.05). The multiple regression analysis adjusting by BMI and number of cigarettes smoked during the day confirmed the results of the univariate analysis. Discussion Our results suggest that unmetabolised U-Benz would allow a more reliable biomonitoring of low-level exposure to benzene than s-PMA and t,t-MA.