AUTHORS: S.J. London, et al.
American Journal of Industrial Medicine. 1994:26 p. 47-60.
STUDY DESIGN: A registry-based case-control design was used to study the hypothesis that electrical workers are exposed to higher magnetic fields and are at higher risk of leukemia than non-electrical workers. Cases included 2,355 men aged 20-64, diagnosed with leukemia between 1972 and 1990 who had an occupation listed at time of diagnosis; controls, 67,212 men aged 20-64, diagnosed with other cancers between 1972 and 1990. Men with cancers of the central nervous system were excluded. Exposure assessment involved magnetic field measurements on non-study subjects in each electrical occupation and on a random sample of non- electrical occupations to estimate magnetic field exposure. Exposure assessment was also obtained for other potential leukemia-causing agents.
CONCLUSIONS: The results demonstrate that workers in electrical occupations have a higher exposure to magnetic fields than workers in non-electrical occupations. One hundred and twenty-one (121) leukemia cases were identified in the group of men with electrical occupations. Workers in electrical occupations had a slightly elevated risk of leukemia as compared to workers in non-electrical occupations. (OR=1.3, CI=1.1-1.6). Results were not changed by controlling for other potential leukemia-causing agents. These results are consistent with findings from similar studies which demonstrate that electrical workers may be at a slightly higher risk for developing leukemia.
LIMITATIONS: The power of the study is decreased as a result of assigning "group mean" EMF exposure data to individuals. The precision of exposure estimates is limited resulting in possible misclassification, which can either lessen or increase the observed dose-response relationship. Data on exposure to other leukemogenic agents was not directly measured. Other risk factors associated with cancer, such as smoking, were not controlled for in the analysis. The exposure assessment was based on occupation at time of diagnosis which may differ from occupation of longest exposure. With respect to the controls, it is important to note that cancer controls may not be representative of the population.