Numerous government, public and private organizations have funded studies of the mechanisms by which extremely-low-frequency (ELF) electric and magnetic fields (EMF) might interact with biologic material at the molecular, cellular, tissue, and organism levels. Several thousand publications and reports on the topic have been released in the past 15 years. The investigations have included characterization of the physical aspects of EMF exposure and measurements of the response of biologic systems, so research in this field requires strong multidisciplinary teams to properly record and control the full range of experimental variables and to reliably interpret the results. Research efforts must combine the talents of electrical engineers and physicists to describe the characteristics of exposure, cellular and molecular biologists and biochemists to explore the mechanisms of interactions at the fundamental level, physiologists and toxicologists to evaluate effects in intact functioning animals, and epidemiologists and physicians to interpret observed effects in the context of human health. Brief reviews of the status of different research components contributing to the study of biologic effects appear in the April 1995 issue of the Journal of the American Industrial Hygiene Association.
The role of engineering studies in the overall response to concerns about the possibility of adverse health consequences from exposure to ELF-EMF falls into two categories: first they support biomedical research to explain the relationship between exposure and biologic and health end points. Second, they can be used to assess sources of exposure and to develop mitigating technology such that any adverse health effects that are confirmed can be avoided.
SUPPORT OF BIOMEDICAL RESEARCH
To date, most of the engineering and physical science effort has appropriately centered on the support of biomedical research, seeking to identify and characterize interactions between ELF-EMF and biologic systems. That support has occurred in four primary areas.
Identification of the Range of Fields in the Environment
Early measurement programs provided an understanding of the magnitude and characteristics of EMF in two specific environments that were the focus of public concern: high-voltage electric power transmission lines and the Navy ELF communications system. Subsequently, EMF measurements have been made in residences, workplaces, on transportation systems, and elsewhere. These measurements have sufficient detail to at least bracket the range of exposure for large segments of the population.
Development of Exposure Facilities
A second important way in which engineers and physical scientists support their counterparts in biomedical research is in the development and characterization of appropriate laboratory exposure facilities. The realization of the importance of this function early in the history of ELF-EMF effects research is a credit to the field and, with the exception of a few studies, has helped to produce a literature in which exposure conditions are appropriate and well defined. Recently, some have criticized past research as having insufficiently characterized exposure conditions because characteristics and attributes of EMF not suspected as being important were not fully investigated: transient components, intermittency, the static geomagnetic field5 strength and orientation, for example. Unfair as such retrospective criticism is, it emphasizes the need for a continuing effort in the area of exposure system development and characterization. New studies will be required for systematic
The earth possesses a static magnetic field of about 300 to 500 mG—the geomagnetic field. This is north seeking field that is used for compass readings. More specifically, the field direction and inclination with respect to vertical varies from point to point on the earth’s surface. These parameters are considered important in relationship with ac magnetic fields for some biological experiments.
evaluation of potentially important field attributes, such as frequency, intermittency, coherency, transients, and other changes, as well as coexposures to static fields, radio frequency fields, and visible light.
Human Exposure Assessment
Epidemiology is an important component of biomedical research to identify and quantify possible adverse health effects of ELF-EMF exposure. A critical part of all epidemiologic studies is the identification of “exposed” and “unexposed” populations. Early occupational studies relied primarily on job titles, and early residential studies frequently rated a subject ’s exposure by scoring observable characteristics of the electric power distribution system outside the home (wire codes). Because reliance on such surrogates of field exposure has exhibited more convincing associations with disease than have quantitative exposure assessments of the actual electric and magnetic fields, subsequent epidemiology studies have placed progressively more emphasis on measurements of additional EMF characteristics, particularly magnetic field characteristics, other than spot measurements of average root-mean-square (rms) field strength. Surprisingly, the increased sophistication of magnetic field exposure assessment in recent epidemiology studies has not significantly improved the correlation between estimated total exposure and the incidence of disease, even though the assessed exposures correlate much more favorably with contemporary exposures measured with a recording device worn on the body. Job titles and wire codes continue to associate more strongly with disease than do measured fields. This paradox has caused some to question whether time-weighted rms field exposure is the relevant measure of dose, or even whether magnetic fields are implicated at all.
Early concepts of ELF-EMF interactions with biologic tissue focused almost entirely on the activity of electric fields and the currents induced within the body; time-varying magnetic fields, in effect, produce internal electric currents. Because the magnitudes of those induced quantities are, for most environmental exposures, small compared with endogenous electric fields and current densities,
and because the forces exerted on ions by those small fields result in motions that are far smaller than thermal motion, some researchers have stated categorically that environmental levels of ELF magnetic fields must be biologically inactive. The inconsistency between such a conclusion and some laboratory results has stimulated work to develop new theoretical models for interactions between fields and biologic material. Models invoking quantum mechanical resonances during concurrent exposure to specific alternating current and static magnetic field conditions, or response to spatial and temporal coherency of the field, have received some limited experimental support. Discrepancies between model predictions and experiments that use average rms field strength as the independent variable also have stimulated questions about whether there are other attributes of field exposure, such as transients, that might give rise to induced fields above the noise threshold. Continuation of this healthy interaction between theoretical and experimental research will be critical, especially if fields interact with biological tissues only when certain combinations of field attributes or specific coexposure combinations exist, because such complex interactions would not be detected with normal toxicologic screening approaches.
EXPOSURE AVOIDANCE AND MITIGATION
Industrialized society receives the vast majority of its ELF-EMF exposure from artificial sources. Consequently, if research reveals that exposure to these fields, either in general or as some specific combination of field attributes, precipitates adverse health effects, the relevant exposures can, in all likelihood, be identified and then eliminated or controlled through a variety of engineering approaches.
There is growing social pressure to begin practices of exposure avoidance or field mitigation, even though a risk to public health has not been demonstrated and the exposure conditions that give rise to some speculative risk are unknown. Lacking both quantified risk and any concept of the benefit of mitigative actions, effective and efficient mitigative procedures cannot be specified.
Biologic studies are designed first to detect and characterize the general response of a biologic system to EMF; that is, to determine whether exposure to power frequency (50-60 Hz) EMF functionally or structurally alters organisms. If effects are observed, these data can be used to determine whether the processes involved could lead to a health detriment. Risk assessment studies are normally organized to achieve three goals: They characterize the toxic properties of the agent, evaluate the health hazard, and make a quantitative characterization of risk.
In the first phase of a risk assessment, studies explore biologic response over a wide range of doses and exposure pathways. The essential goal is to determine a dose-response relationship. A critical requirement for these studies is a rigorous definition of the agent itself. With chemicals, the molecular identity, structure, and purity must be determined so that the dose actually administered can be accurately assessed. These requirements are difficult to apply to ELF-EMF studies for two primary reasons. First, it is not possible to apply an arbitrarily large exposure field. Second, the term “EMF” can be applied to field conditions that vary in several respects—even when limited to frequencies of 50 or 60 Hz. It has been suggested, for example, that transients, large excursions from the normal field strength that can occur on a fast time scale, are by nature more “biologically detectable” than are smooth-varying power frequency sinusoidal electric or magnetic fields. The recent attention to transients might result, however, as much from the general failure of sinusoidal fields, free of transients, to induce effects in experimental systems as from any theoretical biophysical analysis. This consideration is based on the assumption that transients might produce a detectable biologic signal with a significant signal-to-noise ratio, whereas the energy content of a smoothly varying sinusoidal field of low intensity is well below the average thermal “noise” of the biologic system. Additionally, real world
measurements show that transients are common; the “pure” electromagnetic environment created in the laboratory is the unusual condition. The possibility remains that the “noise” eliminated in the laboratory may be the signal that influences the biologic process.
DEFINE THE HAZARD
For a hazard to be identified, health detriment must be demonstrated in an appropriate animal surrogate. Documenting surrogate appropriateness entails identifying the biologic response of the target cells of the model, determining dose-response patterns in relationship to the animal’s environment, determining methods of delivering the putative toxic agent, and comparing these to appropriate conditions in humans. Once the interspecies correlation is established under well-defined exposure conditions, the process can move to the next step.
CONDUCT A RISK ASSESSMENT
Risk assessment combines knowledge of the toxic properties of the agent in the surrogate with what is known for humans. For EMF exposure it has been difficult to identify the toxic properties, if any, of the fields or to place the results that appear contradictory for both laboratory and epidemiologic studies into a risk-hazard framework. If EMF is a toxic agent, it is unlike any other toxic agent that has been extensively studied.
Even though it seems unlikely that the three goals of the toxicologic approach are immediately achievable for the case of EMF exposure, in vitro and in vivo studies have an intrinsic value; the determination of biologic sensitivity to EMF exposure should be considered as an important component of toxicology, independent of the direct application to risk assessment.
An additional factor that must be considered in the evaluation of the role of EMF exposure on human health effects is the regular appearance of publications suggesting new biologic effects of EMF, particularly in combination with other agents known to be detrimental to health, such as chemical mutagens. Such reports should be carefully analyzed and, if deemed
significant in the assessment of risks, replicated in other laboratories. The process of funding such work through grants, however, tends to be a poor vehicle for ensuring rapid confirmation of preliminary studies by independent investigators.
IN VITRO STUDIES
The current status of in vitro studies of EMF effects is one of uncertainty. Overall, in vitro studies seek to determine whether cultured cells respond to exposures to EMF delivered in temporal and intensity patterns believed equivalent to, or exceeding, human environmental exposures. The literature includes two general types of in vitro studies. First, several systems used historically to evaluate the toxic effects of radiation and chemicals have been exposed to a relatively wide spectrum of EMF patterns and intensities. Overall, these studies have been negative—they have shown no effects—particularly in assays relevant to genotoxicity. Second, some studies measure more subtle changes, such as gene expression or changes in intracellular calcium levels. These show both positive and negative effects for EMF, again for a fairly wide spectrum of EMF exposures and mostly at field strengths significantly exceeding those in the normal environment. The studies have examined several effects, including the following:
Membrane effects with emphasis on biophysical mechanisms of charge distribution and inhibition of membrane functions, such as the operation of ion channels and of transmembrane receptors.
Ion effects with emphasis on the intracellular distribution of ions and effects on specific ions such as calcium.
Gene expression with emphasis on genes associated with the signal transduction pathway, oncogenes, and other genes previously reported altered by EMF.
Changes in intracellular levels of proteins, such as ornithine decarboxylase, in response to EMF exposure.
The major emphasis of the in vitro component of the Electric and Magnetic Fields (EMF) Research and Public Information Dissemination (RAPID) program has had the following goals:
Confirm or refute reports of EMF exposure elevating or decreasing several end points in cultured cells. These comparative studies must determine whether a positive response has been peculiar to specific laboratories, exposure apparatus, experimental techniques or cell systems.
Identify one or more cellular systems that produce a robust response, that is, a response that is significant and reproducible for systems in which the dynamic range of response is large compared with uncertainties of the measurement. Such systems are needed to characterize those components of EMF that are essential to elicit response, as well as to determine temporal patterns of the response. Because the combinations and permutations of exposure characteristics are extremely large, the absence of a robust response denies the ability to identify those particular field and exposure characteristics that elicit responses. Because there are some reports in the literature that suggest that responses are limited to “windows” in intensity and/or frequency or to the effects of aperiodic transients, this is an extremely important goal.
Identify and characterize possible mechanisms that would lead to establishing clear hypotheses for biologic effects of EMF at all levels of biologic activity from biophysics to the interaction of EMF with biologic structures; induced response at the subcellular and cellular levels; and sequelae of such response in terms of possible health
detriments in animals and humans. Some of these studies should be designed to determine whether a cellular basis for positive epidemiologic studies can be identified.
IN VIVO STUDIES
In vivo experiments complement and bridge our knowledge base gained from in vitro studies and observations from epidemiology. In vivo studies have the advantage of placing potential target cells within the anatomical and physiologic context in which toxic effects are manifested. For some end points—including cancer, developmental toxicity, and neurobehavioral effects—this is the level at which end points must be assayed, quantified as to the magnitude of effect, and characterized mechanistically. Although in vitro studies can identify the toxic properties of agents and can be useful in studying mechanisms, observations of toxic effects in exposed animals remain the major tool to characterize potential toxicity for any given agent, free of the confounding variables intrinsic to epidemiologic studies of human populations.
When applications of in vivo data are used in a risk assessment context, there are three points that should be considered:
The experimental animal must be an appropriate surrogate for human response to the tested agent. This limitation is especially relevant for toxic effects that require physiologic manifestation of the initial insult, which is almost always the rule. Without detailed knowledge of the extent to which appropriate substitution occurs, both in the initial insult and in the subsequent manifestation of that insult, animal response cannot be interpreted as indicative of probable human response.
The ability to detect small toxic effects can be severely limited by the number of animals that can be assayed. Lifetime bioassay studies of cancer are extremely labor intensive, and significant resources are required to perform large studies. Realistically,
most animal experiments cannot detect effects that occur at less than about a 10% occurrence rate. Epidemiologic studies are interpreted to indicate a potential hazard at a rate of about 1 in 10,000 for a nominal residential exposure to a 60 Hz magnetic field of 2 milliGauss (mG). Because animal studies are limited in the number of animals that can be reasonably handled (normally not more than a hundred or so animals), animal studies might not detect such rare effects unless studies are conducted with much higher levels of exposures, say, up to 2,000 mG. This estimate is based on the dose-response relationship being linear and that both species (humans and animals) have similar sensitivity. For deterministic effects, such as neurotoxicity or developmental effects, higher statistical power can be obtained with smaller numbers of animals.
Chronic toxic effects, such as cancer, must be interpreted in animal studies through a knowledge of the dose-response relationship for the toxic effect of concern in the species of animal used. This knowledge not only assists interpretation of effects induced by high doses in a limited number of animals, but also places this information qualitatively and quantitatively within a structure of results from an extremely large number of toxic agents. An essential part of any chronic study of toxicity is identifying doses that induce the acute and chronic toxicity. For EMF, such toxicity has not been identified, limiting confidence that results from studies with animals can be extrapolated to predict human response to residential or occupational exposures. The use of animal species used classically in biomedical research is the preferred approach in dealing with this uncertainty, because their responses to many other agents are generally predictive of human response.