Risk of Cancer and Exposure to Gasoline Vapors

Elsebeth Lynge 3 4 Aage Andersen 1 4 Ralph Nilsson 0 4 Lotti Barlow 4 8 Eero Pukkala 4 7 Rotf Nordlinder 0 4 Paolo Boffetta 4 6 Philippe Grandjean 4 5 Pirjo Heikkila 4 9 Lars-Gunnar Hbrte 4 Robert Jakobsson 2 4 Ingvar Lundberg 2 4 Bente Moen 4 11 Timo Partanen 4 12 Trond Riise 4 10 0 Sahlgrenska University Hospital, Dept. of Occupational Medi- cine , Gothenburg , Sweden 1 Norsk Kreftregister , Montebello, Oslo , Norway 2 Dept of Occupational Medicine, Karohnska Hospital , Stock- holm , Sweden 3 Danish Cancer Society , Copenhagen , Denmark 4 Received for publication May 2 , 1996, and accepted for publica- tion October 9, 1996 Abbreviations: Cl, confidence interval; SIR, standardized inci- dence ratio 5 Odense University, Institute of Community Health , Odense , Denmark 6 International Agency for Research on Cancer , Lyon , France 7 Finnish Cancer Registry , Helsinki , Finland 8 Socialstyrelsen, Centre for Epidemiology , Stockholm , Sweden 9 Institute of Forensic Medicine, University of Uppsala , Uppsala , Sweden 10 Division of Occupational Medicine, Department of Public Hearth, University of Bergen , Bergen , Norway 11 Division of Occupational Hearth, Institute of Public Hearth and Pnmary Hearth Care , Bergen , Norway 12 Institute of Occupational Health , Topeliuksenkatu 41 aA, Hel- sinki , Finland Until the introduction of self-service around 1970, service station workers in the Nordic countries were exposed to gasoline vapors. Based on measurements reported in the literature, the 8-hour time-weighted average benzene exposure was estimated to be in the range of 0.5-1 mg/m3. We studied the cancer incidence in a cohort of 19,000 service station workers from Denmark, Norway, Sweden, and Finland. They were identified from the 1970 censuses and followed through 20 years, where 1,300 incident cancers were observed. National incidence rates were used for comparison. The incidence was not increased for leukemia (observed = 28, standardized incidence ratio (SIR) = 0.9, 95% confidence interval (Cl) 0.6-1.3) nor for acute myeloid leukemia (observed = 13, SIR = 1.3, 95% Cl 0.7-2.1). The incidence was slightly elevated for kidney cancer (observed = 57, SIR = 1.3, 95% Cl 1.0-1.7) and for pharyngeal, laryngeal, and lung cancer. A 3.5-fold risk of nasal cancer was found (observed = 12, SIR = 3.5, 95% Cl 1.8-6.1). This cohort exposed to gasoline vapors with benzene levels estimated to be 0 5-1 mg/m3 showed no excess risk of leukemia or acute myeloid leukemia, a 30% elevated nsk of kidney cancer, and a previously unnoticed risk of nasal cancer. Am J Epidemiol 1997; 145:449-58. gasoline; incidence; kidney neoplasms; leukemia; nose neoplasms The possible risk of cancer following low dose exposure to petroleum products and, in particular, exposure to gasoline vapors containing benzene is an issue of public concern (1, 2). Until around 1970, service station workers in the Nordic countries were employed to fill customers' cars. Since then, selfservice increased gradually and became predominant around 1990. Benzene has been used as an indicator of - exposure to gasoline. The volume percentage of benzene in the various gasoline blends ranged from 2 to 6 volume percent in the Nordic countries. The available measurements indicate that the 8-hour time-weighted average exposure to benzene for the Nordic service 3 station workers was on average 0.5-1 mg/m , as described below. Biologic monitoring data based on urine samples (3, 4) and peripheral lymphocytes (4, 5) from service station workers have indicated that early effects of exposures occur at such exposure levels. A group of scientists and representatives from the petrochemical industry therefore met in December 1993 at the International Agency for Research on Cancer in Lyon, France, and decided to study the possible cancer risk associated with low dose exposure to gasoline vapors as a joint Nordic analysis of cancer incidence in service station workers. We report here on the outcome of this study. MATERIALS AND METHODS Population The aim was to assess the cancer risk among persons occupationally exposed to gasoline and gasoline vapors. We did not focus on persons occupationally exposed to exhaust fumes and combustion products, although many of the people exposed to gasoline would unavoidably also have been exposed to exhaust fumes. The service station workers were identified from the data files of the 1970 census in Denmark, Norway, Sweden, and Finland. They were followed up for deaths, emigrations, and incident cancer cases through a 15- to 20-year period. The use of personal identification numbers and central population registers in the Nordic countries ensured a virtually complete followup. At the 1970 census, self-administered questionnaires were completed by each household in free text, where the education, job title, and name and address of the work place at the time of the census had to be listed for each member of the household. The questionnaires were checked by the municipalities in Denmark and Norway and centrally coded by the Central Bureau of Statistics in all four countries (6). The industry for a given person was coded according to nationally adapted versions of the International Standard Industrial Classification of 1958 (7). In Norway, Sweden, and Finland, the occupation was coded according to nationally adapted versions of the International Standard Occupational Classification (8). In Denmark, occupation was coded according to a national classification. Service station workers were identified from the census files using the combination of codes for industry and occupation, as shown in table 1. In Denmark, 5,250 service station workers (4,055 men and 1,195 women) were registered at the census on November 9, 1970. They were followed up for deaths and emigrations through November 8, 1987, by linkage with the Central Population Register and for Gasoline consumption, 1.000 tons, average annual 1959- 19691963 1973 total total 1,038 1,555 0.32 657 (retail trade lue) andoU) 6242 (fuel retailing) incident cancer cases by linkage with the National Cancer Register. For each cancer site, person-years were counted from November 9, 1970, until the date of death, emigration, diagnosis of the studied cancer, or November 8, 1987, whichever came first. The observed number of cancer cases was thus equivalent to the number of observation periods terminated by a date of diagnosis. Person-years and observed cancer cases were counted in the same way for all persons who were economically active on the census date and used for the generation of cancer incidence rates by sex and 5-year age groups (defined as age at the time of the census). These rates were used for calculation of the expected number of cancer cases among the service station workers (9). In Norway, 3,561 persons (3,181 men and 380 women) were service station workers on the census D o date of November 1, 1970. They were followed up for nw deaths and emigrations for 1971-1 (...truncated)