A cohort mortality study of petrochemical workers.

A historical prospective cohort mortality study was conducted for a cohort of 6,588 white male employees of a Texas petrochemical plant because of a suspected increased incidence of malignant brain tumors. Mortality experience from 1941 to 1977 was determined and compared with that of the general U.S. white male population adjusting for age and time period. Overall and cause-specific standardized mortality ratios were calculated for various subgroups of the population defined by length of employment, latency and payroll status. Significant deficits in total cohort mortality were found for all causes of death, all circulatory diseases, all respiratory diseases and all digestive diseases. Although not statistically significant, fewer deaths were observed (O) than expected (E) for all malignant neoplasms. No statistically significant excess of malignant brain tumors was found in the overall plant population (O/E = 12/7.42 = 1.62). A borderline significant excess of brain cancer deaths was found among hourly employees with more than six months' employment based on 10 observed and five expected deaths. This excess was observed to occur among elderly employees (over 55 years) and in later follow-up years (post-1970). Risk did not appear to be related to length of employment. Because of the nature of the problem that prompted this study, the small number of cases involved and the lack of a suspect agent in the plant that could have produced this excess, insufficient evidence was found to conclude that these tumors were occupationally related.