Many epidemiologic studies have since evaluated associations of a number of cancers, particularly leukemias and brain tumors, with estimated or measured exposures to electric fields (EFs) and MFs. Studies of cancer in children have focused on residential power-frequency magnetic field exposure; both residential and occupational exposures of adults have been examined. The reported associations are inconsistent from study to study, and effects range from none to weak. With the passage of time, such studies have generally improved in quality with respect to exposure assessment, population size, and outcome assessment, but this has not resulted in greater consistency of results, stronger associations, greater acceptance, or greater consensus about interpretation.

Epidemiology can be a powerful tool for identifying potential risk factors when there is a strong correlation between increased risk of disease and specific environmental conditions. Epidemiologic studies also have been effective in identifying relatively weak associations between putative risk factors and some cancers. Epidemiology is most successful in cases where there are large differences in exposure, where the adverse effects are not rare, and when large samples can be studied prospectively. However, when association is weak, interpretations are more difficult, and conclusions concerning risk less convincing. Epidemiologic studies are at a serious disadvantage if they are used in an effort to prove that weak associations exist or do not exist.

Recent epidemiologic studies of power-frequency MFs and cancer have been conducted in larger populations and with more rigorous methods, but they have not produced evidence that argues persuasively for a quantitative relationship between increased exposure to power-frequency MFs and increased risk of any particular form of human cancer (NRC 1997).

The results of epidemiologic studies of the association between ELF-MF exposures and cancer are difficult to interpret for a number of reasons:

  • There is considerable uncertainty in exposure assessment. Measurements are usually made after the time period of interest, and the relevant exposure metric is not known.

  • Little is known about the risk factors for different leukemias and central nervous system cancers; therefore, possible confounders cannot be identified, measured.

  • There is no accepted mechanism that can plausibly account for the causality of any association.

  • The cancer outcomes identified are rare, as is high-MF exposure in a population; detection of a causal relationship between exposure and health effects at a level of high statistical significance therefore is difficult to achieve.

For all the above reasons, quantitative estimates of the relationship between exposure level and excess cancer risk are not persuasive. People evaluating epidemiologic findings in this field can arrive at different conclusions, depending on their starting viewpoints. Those concerned about protecting public health might lean toward acceptance of a possible association between MFs and cancer risk, whereas others might reject such an association based on the lack of a plausible mechanism and the inability to identify possible confounders. All will agree that the evidence supporting an association is limited.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement