In my commentary on the two California studies of magnetic fields and miscarriage, 1 I had hypothesized that persons who move around more in their home, workplace, and community are more likely by chance alone to encounter one or more sources of relatively high magnetic fields and have more variable exposure over time. The rationale is that we are surrounded by magnetic field sources, ranging from pencil sharpeners to sewing machines to photocopiers, and that the only difference between someone who encounters such an exposure and someone who does not is the extent of random movement through their otherwise similar environments. If, indeed, moving around were the issue, then there are reasons to believe that nausea of early pregnancy or advanced size of later pregnancy would inhibit such random motion.
In their letter, Li and Neutra 2 provide evidence from the prospective study by Li and colleagues 3 that directly addresses the question of nausea. They report that nausea is not related to exposure to a maximum magnetic field of 16 or more mG or to the rate-of-change metric, counter to my hypothesis. The additional data on changes in magnetic fields from early to late pregnancy in the substudy by Lee et al4 suggest that the anticipated immobility of late pregnancy does not result in lower exposure. Either my anticipated causes of reduced mobility (nausea, increased girth) do not truly affect mobility, or mobility does not influence maximum magnetic field or the rate-of-change metric.
We now know a great deal about why some people have relatively high time-weighted average magnetic fields, eg, average above 2 or 3 mG. They spend extended time at home or work near a specific magnetic field source, such as an external power line or energized equipment. Such averages do not seem to be greatly affected by sporadic, brief exposures to appliances or passing by power lines. However, to identify sources of brief, more intense exposures, the environmental determinants become notably more complex. Imagine trying to study noise exposure in relation to hearing loss, focusing not on averages or duration of time above a given threshold, but on brief instantaneous bursts of noise (eg, slamming a door or walking near a leaf blower). Although people must vary to some extent in the risk of encountering a single loud noise each day, the random nature of such encounters would make it difficult to assess the consequences, and there would likely be little correlation of an individual’s exposure over time. In fact, the virtual lack of correlation reported by Lee et al4 between exposures in early and late pregnancy for the maximum magnetic field (r = 0.09) and rate-of-change metric (r = 0.19), far lower than for the time-weighted average (r = 0.64), would be consistent with a nearly random phenomenon. However, a random event would not be expected to be associated with risk of miscarriage, as these exposure measures were, both in the retrospective study and in the small prospective study.
Putting aside the biological plausibility of an effect of maximum magnetic fields of 16 mG or higher or an increased rate-of-change metric, also applicable to time-weighted average magnetic fields, we are left with the question of what these measures capture about a woman’s environment and behavior. Before taking on the extremely challenging goal of replicating the association with magnetic fields, the more modest and readily attainable goal should be to determine what is driving these indices. Such information, which could be uncovered to some extent within the completed studies, as Li and Neutra 2 have begun, would serve as the basis for specifying candidate explanations to compete with the possible etiologic influence of these particular magnetic field exposure metrics.
David A. Savitz
1. Savitz DA. Magnetic fields and miscarriage. Epidemiology 2002; 13: 238.
2. Li D-K, Neutra RR. Magnetic fields and miscarriage (Letter). Epidemiology 2002; 13: 237–238.
3. Li D-K, Odouli R, Wi S, et al
. A population-based prospective cohort study of personal exposure to magnetic fields during pregnancy and the risk of miscarriage. Epidemiology 2002; 13: 9–20.
4. Lee GM, Neutra RR, Hristova L, Yost M, Hiatt RA. A nested case-control study of residential and personal magnetic field measures and miscarriages. Epidemiology 2002; 13: 21–31.