CHARGE TO THE COMMITTEE
Public concern regarding possible health
risks from residential exposures to lowstrength, lowfrequency
electric and magnetic fields produced by power lines and the use
of electric appliances has generated considerable debate among
scientists and public officials.. In 1991, Congress asked that
the National Academy of Sciences (NAS) review the research literature
on the effects from exposure to these fields and determine whether
the scientific basis was sufficient to assess health risks from
such exposures. In response to the legislation directing the U.S.
Department of Energy to enter into an agreement with the NAS,
the National Research Council convened the Committee on the Possible
Effects of Electromagnetic Fields on Biologic Systems. The committee
was asked "to review and evaluate the existing scientific
information on the possible effects of exposure to electric and
magnetic fields on the incidence of cancer, on reproduction and
developmental abnormalities, and on neurobiologic response as
reflected in learning and behavior." The committee was asked
to focus on exposure modalities found in residential settings.
In addition, the committee was asked to identify future research
needs and to carry out a risk assessment in so far as the research
data justified this procedure. Risk assessment is a wellestablished
procedure used to identify health hazards and to recommend limits
on exposure to dangerous agents.
CONCLUSIONS OF THE COMMITTEE
Based on a comprehensive evaluation
of published studies relating to the effects of powerfrequency
electric and magnetic fields on cells, tissues, and organisms
(including humans), the conclusion of the committee is that the
current body of evidence does not show that exposure to these
fields presents a humanhealth hazard. Specifically, no conclusive
and consistent evidence shows that exposures to residential electric
and magnetic fields produce cancer, adverse neurobehavioral effects,
or reproductive and developmental effects.
The committee reviewed residential exposure
levels to electric and magnetic fields, evaluated the available
epidemiologic studies, and examined laboratory investigations
that used cells, isolated tissues, and animals. At exposure levels
well above those normally encountered in residences, electric
and magnetic fields can produce biologic effects (promotion of
bone healing is an example), but these effects do not provide
a consistent picture of a relationship between the biologic effects
of these fields and health hazards. An association between residential
wiring configurations (called wire codes, defined below) and childhood
leukemia persists in multiple studies, although the causative
factor responsible for that statistical association has not been
identified. No evidence links contemporary measurements of magneticfield
levels to childhood leukemia.
Epidemiologic studies are aimed at establishing
whether an association can be documented between exposure to a
putative diseasecausing agent and disease occurrence in
humans. The driving force for continuing the study of the biologic
effects of electric and magnetic fields has been the persistent
epidemiologic reports of an association between a hypothetical
estimate of electric and magneticfield exposure called
the wirecode classification and the incidence of childhood
leukemia. These studies found the highest wirecode category
is associated with a rate of childhood leukemia (a rare disease)
that is about 1.5 times the expected rate.
A particular methodologic detail in
these studies must be appreciated to understand the results. Measuring
residential fields for a large number of homes over historical
periods of interest is logistically difficult, time consuming,
and expensive, so epidemiologists have classified homes according
to the wire code (unrelated to building codes) to estimate past
exposures. The wirecode classification concerns only outdoor factors
related to the distribution of electric power to residences, such
as the distance of a home from a power line and the size of the
wires close to the home. This method was originally designed to
categorize homes according to the magnitude of the magnetic field
expected to be inside the home. Magnetic fields from external
wiring, however, often constitute only a fraction of the field
inside the home. Various investigators have used from two (high
and low) to five categories of wirecode classifications.
The following conclusions were reached on the basis of an examination
of the epidemiologic findings:
· Living in homes classified as
being in the high wirecode category is associated with about
a 1.5fold excess of childhood leukemia, a rare disease.
· Magnetic fields measured in the
home after diagnosis of disease in a resident have not been found
to be associated with an excess incidence of childhood leukemia
or other cancers. T h e link between wirecode rating and
childhood leukemia is statistically significant (unlikely to have
arisen from chance) and is robust in the sense that eliminating
any single study from the group does not alter the conclusion
that the association exists. How is acceptance of the link between
wirecode rating and leukemia consistent with the overall
conclusion that residential electric and magnetic fields not been
shown to be hazardous? One reason is that wirecode ratings
correlate with many factors-such as age of home, housing density,
and neighborhood traffic density-but the wirecode ratings
exhibit a rather weak association with measured residential magnetic
fields. More important, no association between the incidence of
childhood leukemia and magneticfield exposure has been found
in epidemiologic studies that estimated exposure by measuring
presentday average magnetic fields.
· Studies have not identified the
factors that explain the association between wire codes and childhood
Because few risk factors for childhood
leukemia are known, formulating hypotheses for a link between
wire codes and disease is very difficult.. Although various factors
are known to correlate with wirecode ratings, none stands
out as a likely causative factor. It would be desirable for future
research to identify the source of the association between wire
codes and childhood leukemia, even if the source has nothing to
do with magnetic fields.
· In the aggregate, epidemiologic
evidence does not support possible associations of magnetic fields
with adult cancers, pregnancy outcome, neurobehavioral disorders,
and childhood cancers other than leukemia.
The preceding discussion has focused
on the possible link between magneticfield exposure and
childhood leukemia because the epidemiologic evidence is strongest
in this instance; nevertheless, many epidemiologists regard such
a small increment in incidence as inherently unreliable. Although
some studies have presented evidence of an association between
magnetic field exposure and various other types of cancer, neurobehavioral
disorders, and adverse effects on reproductive function, the results
have been inconsistent and contradictory and do not constitute
reliable evidence of an association.
The purpose of exposure assessment is
to determine the magnitudes of electric and magnetic fields to
which members of the population are exposed.
The electromagnetic environment typically
consists of two components, an electric field and a magnetic field.
In general, for timevarying fields, these two fields are
coupled, but in the limit of unchanging fields, they become independent.
For frequencies encountered in electricpower transmission
and distribution, these two fields can be considered independent
to an excellent approximation. For extremelylowfrequency
fields, including those from power lines and home appliances and
wiring, the electric component is easily attenuated by metal elements
in residential construction and even by trees, animals, and people.
The magnetic field, which is not easily attenuated, is generally
assumed to be the source of any possible health hazard. When animal
bodies are placed in a timevarying magnetic field (as opposed
to remaining stationary in the earth's static magnetic field),
currents are induced to flow through tissues. These currents add
to those that are generated internally by the function of nerve
and muscle, most notably currents detected in the clinically useful
electroencephalogram and the electrocardiogram. The currents produced
by nerve and muscle action within the body have no known physiologic
function themselves but rather are merely a consequence of the
fact that excitable tissue (such as nerve and muscle) generate
electric currents during their normal operation.
General conclusions from the review
of the literature involving studies of exposure assessment and
the physical interactions of electric and magnetic fields with
biologic systems are the following:
· Exposure of humans and animals
to external 60hertz (Hz) electric and magnetic fields induces
The density of these currents is nonuniform
throughout the body. The spatial patterns of the currents induced
by the magnetic fields are different from those induced by the
electric fields. Electric fields generally are measured in volts
per meter and magnetic fields in microtesla (uT) or milligauss
(mG) (1 uT = 10 mG).
· Ambient levels of 60Hz
(or 50Hz in Europe and elsewhere) magnetic fields in residences
and most workplaces are typically 0.010.3 uT (0.13
Higher levels are encountered directly
under highvoltage transmission lines and in some occupational
settings. Some appliances produce magnetic fields of up to 100
uT (1 G) or more in their vicinity. For comparison, the static
magnetic field of the earth is about 50 uT (500 mG). Magnetic
fields of the magnitude found in residences induce currents within
the human body that are generally much smaller than the currents
induced naturally from the function of nerves and muscles. However,
the highest field strengths to which a resident might be exposed
(those associated with appliances) can produce electric fields
within a small region of the body that are comparable to or even
larger than the naturally occurring fields, although the magnitude
of the largest locally induced fields in the body is not accurately
· Human exposure to a 60Hz
magnetic field at 0.1 uT (1 mG) results in the maximum current
density of about 1 microampere per square meter (uA/m^2).
The endogenous current densities on
the surface of the body (higher densities occur internally) associated
with electric activity of nerve cells are of the order of 1 mA/m^2.
The frequencies associated with those endogenous currents within
the brain range from less than 1 Hz to about 40 Hz, the strongest
components being about 10 Hz. Therefore, the typical externally
induced currents are 1,000 times less than the naturally occurring
· Neither experimental nor theoretic
data on locally induced current densities within tissues and cells
are available that take into consideration the local variations
in the electric properties of the medium.
Because the mechanisms through which
electric and magnetic fields might produce adverse health effects
are obscure, the characteristics of the electric or magnetic fields
that need to be measured for testing the linkage of these fields
to disease are unclear. In most studies, the root mean square
(rms) strength of the field, an average fieldstrength parameter,
has been measured on the assumption that this measurement should
relate to whatever field characteristics might be most relevant.
As noted earlier, wirecode categories have been used in
many epidemiologic studies as a surrogate measurement of the actual
· Exposure levels of electric fields
and other characteristics of magnetic fields (harmonics, transients,
spatial, and temporal changes) have received relatively little
attention. Very little information is available on the ambient
exposure levels to environmental electric fields other than the
rms measurements of field strength. Those might vary from 5 to
10 volts per meter (V/m) in a residential setting to as high as
10 kilovolts per meter (kV/m) directly under power transmission
lines. Likewise magneticfield exposures are generally characterized
only in terms of their rms field strengths with little or no information
on such characteristics as the frequency and magnitude of transients
and harmonics. Residential exposures to powerfrequency electric
and magnetic fields are generally on the order of a few milligauss.
· Indirect estimates of human exposure
to magnetic fields (e.g., wiring configuration codes, distance
to power lines, and calculated historical fields) have been used
These estimates of magnetic fields correlate
poorly with spot measurements of residential 60Hz magnetic fields,
and their reliability in representing other characteristics of
the magnetic field has not been established. Because of the many
factors that affect exposure levels, great care must be taken
in establishing electric and magneticfield exposures.
· Unless exposure systems and experimental
protocols meet several essential requirements, artifactual results
are likely to be obtained in laboratory animal and cell experiments.
Many of the published studies either have used inferior exposure
systems and protocols or have not provided sufficient information
for their evaluation.
In Vitro Studies on Exposure to Electric
and Magnetic Fields
The purpose of studies of in vitro systems
is to detect effects of electric or magnetic fields on individual
cells or isolated tissues that might be related to health hazards.
The conclusions reached after evaluation of published in vitro
studies of biologic responses to electric and magneticfield
exposures are the following:
· Magneticfield exposures
at 50-60 Hz delivered at field strengths similar to those measured
for typical residential exposure (0.110 mG) do not produce
any significant in vitro effects that have been replicated in
When effects of an agent are not evident
at low exposure levels, as has been the case for exposure to magnetic
fields, a standard procedure is to examine the consequences of
using higher exposures. A mechanism that relates clearly to a
potential health hazard might be discovered in this way.
· Reproducible changes have been
observed in the expression of specific features in the cellular
signaltransduction pathways for magneticfield exposures
on the order of 100 uT and higher.
Signaltransduction systems are
used by all cells to sense and respond to features of their environments;
for example, signaltransduction systems can be activated
by the presence of various chemicals, hormones, and growth factors.
Changes in signal transduction are very common in many experimental
manipulations and are not indicative per se of an adverse effect.
Notable in the experiments using high magneticfield strengths
is the lack of other effects, such as damage to the cell's genetic
material. With even higher field strengths than those, a variety
of effects are seen in cells.
· At field strengths greater than
50 uT (0.5 G), credible positive results are reported for induced
changes in intracellular calcium concentrations and for more general
changes in gene expression and in components of signal transduction.
No reproducible genotoxicity is observed, however, at any field
strength. Again, effects of the sort seen are typical of many
experimental manipulations and do not indicate per se a hazard.
Effects are observed in very high field strength exposures (e.g.,
in the therapeutic use of electromagnetic fields in bone healing).
The overall conclusion, based on the
evaluation of these studies, is that exposures to electric and
magnetic fields at 5060 Hz induce changes in cultured cells
only at field strengths that exceed typical residential field
strengths by factors of 1,000 to 100,000.
In Vivo Studies on Exposure to Electric
and Magnetic Fields
Studies of in vivo systems aim to determine
the biologic effects of powerfrequency electric and magnetic
fields on whole animals. Studies of individual cells, described
above, are extremely powerful for elucidating biochemical mechanisms
but are less well suited for discovering complicated effects that
could be related to human health. For such extrapolation, animal
experiments are more likely to reveal a subtle effect that might
be relevant to human health. The obvious experiment is to expose
animals, say mice, to high levels of electric or magnetic fields
to observe whether they develop cancer or some other disease.
The experiments of this sort that have been done have demonstrated
no adverse health outcomes. Such experiments by themselves are
inadequate, however, to discount the possibility of adverse effects
from electric and magnetic fields, because the animals might not
exhibit the same response and sensitivities as humans to the details
of the exposure. For that reason, a number of animal experiments
have been carried out to examine a large variety of possible effects
of exposure. On the basis of an evaluation of the published studies
in this area, the committee concludes the following:
· There is no convincing evidence
that exposure to 60Hz electric and magnetic fields causes
cancer in animals.
A small number of laboratory studies
have been conducted to determine if any relationship exists between
powerfrequency electric and magneticfield exposure
and cancer. In the few studies reported to date, consistent reproducible
effects of exposure on the development of various types of cancer
have not been evident. One area with some laboratory evidence
of a healthrelated effect is that animals treated with carcinogens
show a positive relationship between intense magneticfield
exposure and the incidence of breast cancer.
· There is no evidence of any adverse
effects on reproduction or development in animals, particularly
mammals, from exposure to powerfrequency 50 or 60Hz
electric and magnetic fields.
· There is convincing evidence
of behavioral responses to electric and magnetic fields that are
considerably larger than those encountered in the residential
environment; however, adverse neurobehavioral effects of even
strong fields have not been demonstrated.
Laboratory evidence clearly shows that
animals can detect and respond behaviorally to external electric
fields on the order of 5 kV/m rms or larger. Evidence for animal
behavioral response to timevarying magnetic fields, even
up to 3 uT, is much more tenuous. In either case, general adverse
behavioral effects have not been demonstrated.
· Neuroendocrine changes associated
with magneticfield exposure have been reported; however,
alterations in neuroendocrine function by magneticfield
exposures have not been shown to cause adverse health effects.
The majority of investigations of magneticfield
effects on pinealgland function suggests that magnetic fields
might inhibit nighttime pineal and blood melatonin concentrations;
in those studies, the effective field strengths varied from 10
uT (0.1 G) to 5.2 mT (52 G). The experimental data do not compellingly
support an effect of sinusoidal electric field on melatonin production.
Other than the observed changes in pineal function, an effect
of electric and magnetic fields on other neuroendocrine or endocrine
functions has not been clearly shown in the relatively small number
of experimental studies reported.
Despite the observed reduction in pineal
and blood melatonin concentrations in some animals as a consequence
of magneticfield exposure, studies of humans provide no
conclusive evidence to date that human melatonin concentrations
respond similarly. In animals with observed melatonin changes,
adverse health effects have not been shown to be associated with
electric or magneticfieldrelated depression
· There is convincing evidence
that lowfrequency pulsed magnetic fields greater than 5
G are associated with bonehealing responses in animals.
Although replicable effects have been
clearly demonstrated in the bonehealing response of animals
exposed locally to magnetic fields, the committee did not evaluate
the efficacy of this treatment in clinical situations.
COMMITTEE ON THE POSSIBLE
EFFECTS OF ELECTROMAGNETIC FIELDS ON BIOLOGIC SYSTEMS
CHARLES F. STEVENS (Chair), Howard
Hughes Medical Institute, Salk Institute, La Jolla, Calif.
DAVID A. SAVITZ (Vice Chair), Department
of Epidemiology, University of North Carolina, Chapel Hill, N.C.
LARRY E. ANDERSON, Pacific Northwest
National Laboratory, Richland, Wash.
DANIEL A. DRISCOLL, Department of Public
Service, State of New York, Albany, N.Y.
FRED H. GAGE, Laboratory of Genetics,
Salk Institute, San Diego, Calif.
RICHARD L. GARWIN, IBM Research Division,
T.J. Watson Research Division, Yorktown Heights, N.Y.
LYNN W. JELINSKI, Center for Advanced
TechnologyBiotechnology, Cornell University, Ithaca, N.Y.
BRUCE J. KELMAN, Golder Associates,
Inc., Redmond, Wash.
RICHARD A. LUBEN, Division of Biomedical
Sciences, University of California, Riverside, Calif.
RUSSEL J. REITER, Department of Cellular
and Structural Biology, University of Texas Health Sciences Center,
San Antonio, Tex.
PAUL SLOVIC, Decision Research, Eugene,
JAN A.J. STOLWIJK, Department of Epidemiology
and Public Health, Yale University School of Medicine, New Haven,
MARIA A. STUCHLY, Department of Electrical
and Computer Engineering, University of Victoria, B.C., Canada
DANIEL WARTENBERG, UMDNJRobert
Wood Johnson, Medical School, Piscataway, N.J.
JOHN S. WAUGH, Department of Chemistry,
Massachusetts Institute of Technology, Cambridge, Mass.
JERRY R. WILLIAMS, The Johns Hopkins
Oncology Center, Baltimore, Md.