Abstract

A variety of exposure assessment methods have been used in epidemiological studies of benzene-exposed workers. The most usual metrics are duration, mean daily exposure intensity (ppm) and ppm-years (intensity multiplied by duration and summed over the exposure period.

Many of the early studies were of cancer and mortality in petroleum industry workers. These studies compared the cancer rates of the workers with those of the general population. Benzene was known to be present at such facilities but benzene exposure estimates were not provided.

In some papers, the cancer risk was compared between exposed and unexposed workers in the same industry or facility. Other papers compared the cancer risk for workers categorised into Low, Medium and High exposure groups, this categorisation may have drawn on measured exposure data. The categorisation may have used a general population or an industry-specific job-exposure matrix (JEM). The latter being likely to be more precise.

The strongest exposure estimates were those where benzene exposure estimates were attributed by experts to individual job titles/areas of work based on measured data, from an appropriate time frame. The exposure attribution was normally done case-blind for case-control studies to avoid bias. Exposure modifiers may have been used to extrapolate data to jobs and eras where measurements were not available. In particular personal benzene exposure data were not usually available before 1980, so extrapolations back in time were needed. In some cases hydrocarbon measurements have been used as a surrogate for benzene measurements. Validation of exposure estimates can add to confidence in the estimates.

In a few studies, biological exposure data was used to attribute exposure estimates to work. Where the data are from the individual concerned rather than attributed on a group basis, the exposure assessment is considered more robust.