LEUKEMIA, BRAIN TUMORS, AND EXPOSURE TO EXTREMELY LOW FREQUENCY ELECTROMAGNETIC FIELDS IN SWISS RAILWAY EMPLOYEES.

Minder, C. E.; Pfluger, D. H.

Inst. for Social and Preventive Medicine, Univ. of Berne, Berne, Switzerland, e-mail: minder@ispm.unibe.ch (RR/C.E.M., D.H.P.)

The authors examined occupational exposure to extremely low-frequency electromagnetic fields (ELF EMFs) as a risk factor for leukemia and brain tumors among Swiss electrified railway engineers and train attendants. This study was an extension, using improved methods, of a previous study which identified an elevated standardized mortality ratio due to hematopoietic and lymphatic malignancies in a cohort of Swiss electric rail engineers (Balli-Antunes et al., Environmetrics 1:121-130, 1990; BENER Abstract No. 7564). Personnel and pension records of the Swiss Federal Railway were searched from 1972 through 1993 to identify a cohort of 18,070 males employed as or retired from work as line engineers, shunting yard (SY) engineers, train attendants, and station masters. The cohort was followed for mortality from January 1, 1972 through December 31, 1993 by linking the records to death certificates using a probabilistic record linkage technique. This provided 270,155 person-years of mortality follow-up. A total of 37 deaths due to leukemia, 98 due to other hematopoietic or lymphatic neoplasms, and 23 due to brain tumors were identified. Of this total of 158 cases, 123 linked cases could be manually verified for vital status against archived pension records. Magnetic field exposures of line and SY engineers were evaluated by measuring flux densities at 16-2/3 Hz (the primary frequency used for all Swiss trains) in the driver's seat of the engines using a portable recording meter (Bramur, Inc., Lee, MA). The instrument records rms flux densities in the 0-100 Hz frequency range at 10-sec intervals and stores the data in a battery-powered computer. The measurements were taken at 3 points just behind the driver's seat (at the level of the head, thorax, and feet) for each major type of engine manufactured in Switzerland since 1905. The measurements covered complete driving cycles (starting, acceleration phase, driving phase, braking, and stopping) that lasted from 20 min to 4 hr. Magnetic field exposure assessments of train attendants and stationmasters were made from spot measurements (lasting for 2 to 30 min) taken at their most frequent places of work such as various positions within the coaches (train attendants) and on platforms and in offices (station masters). Exposures were then estimated as time-weighted averages (TWAs) of location-specific field strengths. Historical magnetic field exposures of line and SY engineers were estimated as calculated weighted averages of engine-specific exposures (for each type of engine in service between 1905 and 1995) for each 5-yr calendar interval during the 90-yr period. For train attendants and station masters, historical magnetic field exposures were estimated by linearly interpolating between 0 uT for 1900 and actual exposures measured in 1993. Cumulative 16-2/3-Hz magnetic field exposures were calculated from the data and grouped into 3 exposure categories: 0 to 4.99 uT-yr, 5.00 to 74.99 uT-yr, and 75 uT-yr or greater. The fractions of each year the subjects were exposed to 16-2/3-Hz magnetic fields of 10 uT or greater were cumulated over life and aggregated into the following categories: less than 0.099 yr, 0.100-0.499 yr, and 0.500 yr or greater. Mortality rate ratios (MRRs) for leukemia and brain cancer were calculated by Poisson regression techniques using deaths from leukemia or brain cancer per unit of analysis (cumulative exposure or time fractions of each year exposed to magnetic fields of 10 uT or greater) as the dependent variable, and 5-yr age group, 10-yr calendar period, and job or exposure category as the independent variables. In the cohort of 18,070 workers a total of 6,879 were employed during the study period as line engineers; 1,314 as SY engineers; 5,720 as train attendants; and 4,157 were employed as station masters. Historical exposure estimates described the changing exposure patterns for the highest exposure group, line engineers, compared to progressively less exposed SY engineers, train attendants, and station masters (the reference group). The mean cumulative exposure of line engineers in the years 1930, 1960, and 1990 was 9.3, 17.9, and 25.9 uT-yr, respectively. Their mean percent time exposed to fields above 10 uT in these years was 9.9, 29.5, and 47.5, respectively. Mean cumulative exposure of the SY engineers in 1930, 1960, and 1990 was 2.6, 13.4, and 13.4 uT-yr, while mean percent time exposed to fields above 10 uT in these years was 1.4, 9.7, and 9.7, respectively. The mean cumulative exposures of train attendants in 1930, 1960, and 1990 were 0.4, 1.9, and 3.3 uT-yr and mean percent time exposed to fields above 10 uT in those years was 0.4, 2.2, and 3.6, respectively. Station masters had mean cumulative exposures in 1930, 1960, and 1990 of 0.1, 0.6, and 1.0 uT-yr, respectively, and were not exposed to magnetic fields stronger than 10 uT in any time period. The MRR from all causes (with 95% confidence interval (CI), using station masters as the referents) was similar and close to unity for all groups: 1.01 (CI 0.93-1.11), 1.08 (0.94-1.25), and 1.07 (CI 0.98-1.17) for line engineers, SY engineers, and train attendants, respectively. The MRR for leukemia was slightly but not significantly elevated for line engineers, 2.44 (CI 0.97-6.11), and SY engineers, 2.00 (CI 0.50-8.07), while it remained close to unity for train attendants, 1.09 (CI 0.39-3.05). The association between occupational leukemia mortality rate ratios and mean exposures for those occupations indicated a significant trend in 1930 (p=0.04), 1960 (p=0.03), and 1990 (p=0.03). Application of the cumulative year of exposure metric showed that there was an increase in leukemia mortality of 0.9% (CI 0.2-1.7%) per uT-yr of exposure. For years spent exposed to 10-uT fields or stronger, leukemia increased at the rate of 62% per yr (CI 15-129) and the trend was significant for all three years of exposure (p<0.05). The excess risk of leukemia by area of exposure reached significance only for highest exposure to the thorax (0.5 uT-yr or more, adjusted MRR of 2.43, CI 1.10-5.36 based on 22 cases). The MRR for brain tumors in the 3 groups was 1.02 (CI 0.23-4.55), 5.06 (CI 1.21-21.2), and 2.67 (CI 0.75-9.62), respectively. Brain cancer mortality did not show any dose-response relation for either cumulative year of exposure or years above 10 uT, but rather seemed to aggregate in the occupations of SY engineer and train attendant showing inconsistent patterns in relation to extremely low-frequency EMF exposure. This suggests that it is unlikely that EMF exposure is solely responsible for the finding. In a discussion of possible confounding exposures, the authors noted that engineers were not involved in cleaning and maintaining of railway engines. Cleaners used during the study period were phosphoric and sulfuric acids, and later formic acid. Benzene, a known leukomogen, was never used as a cleaning fluid and polychlorinated biphenyls were never used in Swiss Railway transformers. The authors concluded that these results support the hypothesis that prolonged exposure to high levels of ELF magnetic fields promotes or causes leukemia, and relate this conclusion to results of several other reports in the literature. From the standpoint of methodology, this study reinforces the necessity of accurately assessing EMF exposures through historical reconstruction since even a moderate exposure misclassification will result in insignificant results. Given the moderate size of the risks involved and the rarity of outcomes like leukemia and brain tumors, even a moderate degree of exposure misclassification will invariably lead to insignificant results. [An invited editorial on this study by David Savitz appears in the same issue (Am J Epidemiol 153:836-838, 2001, BENER Abstract No. 21827) accompanied by a reply by Minder and Pfluger (Am J Epidemiol 153(9):839-840, 2001, BENER Abstract No. 21828]. (30 Refs). [Copyright 2001, Information Ventures, Inc.]

Information Ventures, Inc. • 42 South 15th Street, Suite 700 • Philadelphia, PA 19102 • USA • Phone: +215-569-2300 • Fax: +215-569-2575 • E-mail: EMF-Link@infoventures.com