Bitumens and bitumen emissions, and some heterocyclic polycyclic aromatic hydrocarbons.

1190 www.thelancet.com/oncology Vol 12 December 2011 In October, 2011, 16 experts from nine countries met at the International Agency for Research on Cancer (IARC) in Lyon, France, to reassess the carcinogenicity of bitumens and their emissions, and of some N-heterocyclic and S-heterocyclic polycyclic aromatic hydrocarbons (PAHs). These assessments will be published as Volume 103 of the IARC Monographs. Bitumens are produced by distillation of crude oil during petroleum refi ning, and also occur naturally. Bitumens can be divided into six broad classes according to their physical properties and specifi cations required for diff erent applications (webappendix p 1). The main use (about 80%) of bitumens is for road paving; other uses include roofi ng, waterproofi ng, sealing, and painting. The term bitumen should not be confused with asphalt, which refers to the mixture of bitumen (4–10% by weight), small stones, sand, and fi ller used for road paving. Four major occupational exposures to bitumens and bitumen emissions have been evaluated for cancer risks in epidemiological studies: roofi ng, road paving, mastic asphalt work, and other occupations such as manufacturing of bitumen products and asphalt mixing. Roofers are often exposed to relatively high levels of bitumen emissions, due to high application temperatures (webappendix p 1). A meta-analysis of seven studies published before 1994 reported an increased risk for lung cancer (relative risk 1·78 [95% CI 1·50–2·10]), with risk estimates similar for cohort and case–control studies, the latter being adjusted for smoking. A subsequent Euro pean cohort study among workers exposed to bitumens reported an increas ed risk for lung cancer among roof ers and waterproofi ng workers (standardised mortality ratio [SMR] 1·33 [95% CI 0·73–2·23]). However, roofers can also be exposed to other lung carcinogens, such as coal tar from removing old roofs, and potential confounding is diffi cult to rule out. Four cohort studies reported increased risks for cancers of the upper aerodigestive tract, but potential confounding by tobacco smoking, alcoholic beverages, or other occupational exposures could not be excluded. A meta-analysis of four studies of road pavers reported no increase in lung cancer risk. The European cohort study, the largest study of road pavers published subsequently, reported a small increase in lung cancer mortality among road pavers compared with the general population (SMR 1·17 [95% CI 1·04–1·30]); however, the eff ect was attenuated when other construction workers were used as internal referents. Of the quantitative estimates of exposure to bitumen emissions modelled (cumulative, average, and duration of exposure), average exposure was signifi cantly associated with lung cancer mortality. In a nested case–control study that excluded pavers who were likely to have been exposed to coal tar and higher levels of bitumen, there was no suggestion of an increase in lung cancer risk with various metrics of bitumen exposure. Overall, the evidence for lung cancer risk among road pavers and road maintenance workers exposed to bitumens is inconsistent. Additionally, potential confounding by exposures to other carcinogens such as coal tar, diesel engine exhausts, or silica dust could not be ruled out. Work with mastic asphalt involves exposure to the highest reported levels of bitumen emissions, because of its use at high temperatures (webappendix p 1). Two studies investigated cancer risks associated with exposure to bitumens during mastic asphalt work, and both showed an increased risk for lung cancer. All informative studies of carcinogenicity of bitumens in experimental animals have been done in mice (apart from one inhalation study in rats), with bitumen applied dermally either neat, in a solvent (benzene, acetone, toluene, mineral spirits, mineral oil), or as a condensate of emissions generated from the bitumens. One of three studies with oxidised bitumens (Class 2) and all four studies with their fume condensates showed an increased risk for skin tumours (webappendix p 2). By contrast, none of the studies with straight-run bitumens (Class 1) or their fume condensates, including the inhalation study in rats, showed an increased risk. Of the two studies that investigated cut-back bitumens (Class 3), the initiation-promotion study showed evidence of a promoting eff ect on skin tumours. Bitumens of Bitumens and bitumen emissions, and some heterocyclic polycyclic aromatic hydrocarbons