Authors
Otto Wong and Gerhard Raabe

Title
A critical review of cancer epidemiology in the petroleum industry, with a meta-analysis of a combined database of more than 350,000 workers

Journal
Regulatory Toxicology and Pharmacology 32, 78-98 (2000)

Summary
Several aliphatic and polycyclic aromatic hydrocarbons have been detected in the air of specific working environments within the petroleum industry. There is sufficient evidence for the carcinogenicity in experimental animals of several polycyclic aromatic hydrocarbons; however, limited evidence exists to indicate that working in petroleum refineries entails a carcinogenic risk. Another area of scientific uncertainty concerns the toxicological effects of various petrochemicals on the endocrine system. Several compounds produced from crude oil demonstrate androgenic or estrogenic properties (i.e. benzopyrene, thiram).
The data that exist for evaluating the postulated relationship between exogenous hormonally active agents (HAAs) and human cancers are essentially limited to studies involving organochlorine exposure (i.e. DDT, PCBs). Therefore, this study is valuable because it provides data on the relationship between prostate cancer mortality and occupational exposure.
In 1989, the authors of the present investigation published a review of cancer epidemiology in petroleum workers, which included a meta-analysis by cancer site. The purpose of the present study was to update the 1989 review and conduct a meta-analysis of cancers (other than leukemias and lymphomas) in cohort studies of petroleum workers. A total of 28 individual cohorts were included in the present investigation, representing all cohort mortality studies conducted in Canada, the United States, the United Kingdom, Australia, Sweden, Finland, and Italy. The combined database comprised over 350,000 petroleum workers and the observation period covered up to 60 years (1937-1996), with over 7 million person-years at risk. Workers employed at more than 60 refineries (~266,000) formed the majority of all cohort members. The remaining workers were employees in the distribution (gasoline transportation) or crude oil (upstream and production) divisions.
All the cohort studies included in the present review were similar with respect to study design, data collection, cohort identification, exposure, disease classification, and analysis. Various categories of length of employment were employed in different studies; thus, a meta-analyses by length of employment across studies was not possible. The purpose of the meta-analysis was to provide a summary measure of risk (SMR) for several cancer sites, including the stomach, large intestine, liver, pancreas, lung, skin, prostate, bladder, kidney, and brain. The authors concluded that the summary SMR was not significantly increased for these cancer sites. In fact, all summary SMRs were below unity.
Overall, mortality from skin cancer was elevated for petroleum workers (SMR = 1.1, 95% CI 1.0-1.2). Melanoma mortality was significantly elevated in a group of UK refinery workers (SMR = 1.8, 95% CI 1.2-2.5) and in employees at upstream operations in Canada (SMR = 6.0, 95% CI 2.2-13). The authors of these two studies were unable to link the increased melanoma mortality to any specific sources of exposure. However, the authors of the Canadian study listed skin contact with several possible sources of complex polynuclear aromatic hydrocarbon (PAH) exposure, including crude oil, drilling mud, and natural gas.
A significant increase in prostate cancer mortality was reported in a group of U.S. refinery workers (SMR = 1.4, 95% CI 1.0-1.9) and in a group of short-term employees in several crafts including pumpers (SMR = 2.0, 95% CI 1.0-3.3) and roustabouts (SMR = 1.7, 95% CI 1.1-2.4). However, the absence of a significant trend by length of employment in these workers suggested that the increased risk was not associated with occupational exposure. Human epidemiological studies have implicated several HAAs as risk factors for various cancers. However, the mechanisms of action are not clear, and the results are not consistent. If environmental HAAs are in fact associated with cancer formation in humans, their activity should be most evident in tissues that are known targets for sex steroid hormones (i.e. prostate). This proposal however is not supported by the findings contained in this report.