Invited Commentary: Electromagnetic Fields and Cancer in Railway Workers

The ideal study of occupational exposure to electromagnetic fields and cancer risk would have a clear exposure source, historically stable exposures, and comparable groups of exposed and unexposed workers. Cohorts of railway workers have marked exposure contrasts and limited job changes and provide marginally adequate study sizes, but there have been important changes in their exposures over time, and the field frequency involved is unusual. The results of Minder and Pfluger's study (Am J Epidemiol 2001;153:825-35) add modest support for an association between electromagnetic field exposure and leukemia. However, given the large size and high quality of a number of previous studies of occupational electromagnetic field exposure and cancer, additional studies similar to past ones are unlikely to yield important new insights. Am J Epidemiol 2001;153:836-8. brain neoplasms; electromagnetic fields; environmental monitoring; leukemia; occupational exposure Epidemiologic research on the relation between occupa- tional exposure to electromagnetic fields (EMFs) and risk of cancer (especially leukemia and brain cancer) has advanced considerably in the past 20 years. Starting with simple tabu- lations exploiting available mortality data (1) or cancer reg- istry data (2) on electrical occupations, studies have evolved to include increasingly sophisticated approaches to exposure assessment, detailed consideration of confounding, large study sizes, and refined analytical methods. Tens of millions of dollars have been spent, most of it wisely, to improve upon previous study methods in an attempt to resolve the question of whether occupational exposure to EMFs consti- tutes a health hazard. Unfortunately, the barriers to reaching definitive conclusions transcend financial limitations or a lack of ingenuity among those of us who have addressed the issue. Neither money nor creativity can overcome the inher- ent limitation of insufficient sizes of the populations avail- able for study or the challenge of assessing or reconstructing EMF exposures. The greatest source of ambivalence regard- ing continued pursuit of the hypothesized carcinogenic effects of EMFs continues to be the lack of experimental support and, some would argue, the lack of plausibility that such support will ever be found. What would constitute the ideal opportunity for a study of occupational EMFs and cancer? This benchmark of unattain- able perfection provides a referent for examining the current study by Minder and Pfluger (3), as well as other recent stud- ies of railway workers (4-6), electric utility workers (7-11), and community populations (12, 13). In the ideal setting, there would be a single, well-defined source of EMFs, allow- ing simple calculations of exposure based on the geometry of the workplace and the worker's location. The workers would be physically stable in their jobs relative to that source, not leaving their work stations for extended periods during the workday. This work environment would be historically sta- ble, with no innovation in technology or changes in the man- ner in which the work was done to alter exposure. Within a given industry, there would be other jobs that were identical in terms of training requirements, selection, and associated pay and prestige but free of elevated EMF exposure.