Printed in U.S.A. Response to "Evolution of Epidemiologic Evidence on Magnetic Fields and Childhood Cancers"

Dr. Poole has made a number of useful points about the interpretation of studies of the association of childhood cancers with extremely low frequency magnetic fields (1). We wish to comment on only two of his points. First, Dr. Poole takes issue with the appliance-specific relative risks that were reported in the two papers in this issue (2, 3) and several earlier ones, including another Los Angeles study (4). We agree with his desire for an exposure assessment that consolidates appliances with other sources of magnetic fields, including power lines. However, numerous difficulties preclude doing so at this time. Too little is known about the fields generated by each appliance at the target organs, variations between brands, duration and frequency of use, and other modifying influences needed for a sensible summary index. More important is the uncertainty over the appropriate "exposure metric, " a single number that represents a subject's magnetic field environment, including its frequency spectrum, time variation, polarization, spatial orientation, and high-frequency transients. Since fields from appliances and power lines differ in many of these qualities, weighting exposures from various electromagnetic field sources requires an exposure metric that is hypothetically carcinogenic. "Combining apples and oranges " makes sense in dietary epidemiology only because fructose has been identified as an etiologically relevant nutrient! Although summary electromagnetic field exposure indices would be useful, their development clearly poses formidable challenges to epidemiologists, exposure assessors, and biologists. During the presen