In 1978, a leukemia cluster in highway maintenance workers (HMW) in Wheaton, MN was identified. Six leukemia cases occurred in the male population of Wheaton. Based on cancer rates from the Minneapolis-St. Paul component of the Third National Cancer Survey, only one would have been expected. Five of these six men were HMW. A follow-up case-control study was unable to elucidate the problem. Recently it has been alleged that these cases were the result of benzene contamination in asphalt materials. Concern has also been expressed about exposure to 2,4,5,-T and 2,4-T as well as increased deaths due to injury. Based on Minnesota data there are an estimated 250,000 to 500,000 city, county and state HMW nationwide. There has never been a comprehensive study of the health or mortality of these individuals. This retrospective cohort study will examine all cause and specific causes of mortality in Minnesota HMW. There are an estimated 9,600 people who served as HMW between January 1, 1945 and December 31, 1985. All of these workers will be abstracted from old payroll records and a sub-cohort of approximately 5,600 people working one or more years identified. These workers will be traced to determine vital status. Death certificates will be obtained on the deceased. After death certificates on the deceased have been coded to ICD-9-CM their mortality experience will be compared to that of the Minnesota and U.S. populations by indirect standardization. The """"""""Monson Program"""""""" for the calculation of standardized mortality ratios will be rewritten to do these analyses on an IBM-AT computer. Person years of follow-up will begin after one year of work or approximately 1945. These years will be calculated in five year increments for each of three exposure groups: highway maintenance workers, light equipment operators, and others. Standardized mortality ratios will be calculated and statistical significance determined based on the Mantel-Haensel approximation of the Poisson distribution. Data will be interpreted based on the following criteria: biologic plausibility, strength of association, dose-response, latency, statistical significance, and confounding. Plausible increases in mortality will be followed by case-control studies.
