INTRODUCTION
The potential association between childhood leukemia development and exposure to power frequency magnetic fields has been investigated
extensively both in laboratory settings and in epidemiologic studies. Evidence supporting a possible link between childhood leukemia and magnetic
fields, however, has been provided mostly by epidemiology. In fact, the entire magnetic field-childhood leukemia research field was initiated by an
epidemiologic study published by Wertheimer and Leeper in 1979, which described an association between childhood cancer and electrical wiring
configurations of children’s homes.
Epidemiology and epidemiologic observations have contributed to the identification of a number of associations between
various adverse health outcomes and various exposures. As historical examples, we can mention Percival Pott’s conclusion from 1775 on the soot’s
causative role in scrotal cancer development among the London chimneysweepers or John Snow’s description of the relationship between a cholera
outbreak and the water supply system in London in 1853. More recent examples include the identification of a causal link between smoking and lung
cancer in the 1960s, and the recognition of the acquired immunodeficiency syndrome epidemic in the 1980s.
In this paper we give a short overview of how epidemiologic evidence regarding a possible link between childhood
leukemia and magnetic field exposure developed over the past two and a half decades. We also discuss possible future research directions we feel
could lead to the resolution of the magnetic field–childhood leukemia question.
EVOLUTION OF EPIDEMIOLOGIC EVIDENCE
Exposure assessment
A very important aspect of all epidemiologic studies is reliable exposure assessment. For studies of power frequency electric and magnetic
field exposures, it has not been a simple task. Exposure is ubiquitous in modern societies wherever electricity is used, but exposure cannot be
detected (i.e., seen, heard or sensed) without instrumentation. For both electric and magnetic fields many physical characteristics can be measured,
but due to lack of any known and accepted biophysical mechanism to explain a carcinogenic effect, it is not known which parameter or combination of
parameters should be measured in an epidemiologic study. Exposures to electric and magnetic fields are also highly variable both in space and time.
Intensity of exposure decreases exponentially with distance from many sources; therefore, sharp differences in exposure could be observed within
relatively short distances. Temporal changes in exposure may occur rapidly within seconds or minutes (short-term variability), but exposure patterns
may also change with seasons and years (long-term variability) (Foliart et al., 2002). Exposure to electric and magnetic fields may also exhibit
variation by behavioral characteristics; more active individuals may encounter sources of high exposures more frequently (EPRI, 2003b).
Exposure assessment methods evolved greatly in epidemiologic studies examining potential health effects of electric and
magnetic fields. The first published study used a crude dichotomous categorization (high current configuration code versus low current configuration
code) based on characteristics of transmission and distribution lines near the subjects’ homes and the estimated distance between the lines and homes
both obtained through visual inspection and simple measurements (Wertheimer and Leeper, 1979). In later studies, wire code categories were further
refined to include four, and – with the inclusion of buried cables – five categories (Savitz et al, 1988; Linet et al., 1997). Some studies also
used, as exposure measures, the simple distance from the subjects’ homes to transmission and distribution lines and substations (UKCCS, 2000).
In comparison to wire code categories, historical magnetic field calculations near transmission lines represented further
refinement in exposure assessment (Feychting and Ahlbom, 1993). Calculated fields relied not only on the power lines’ physical characteristics and
their distance from the homes, but it also incorporated information on the lines’ recorded historical load over the study years.
With the development and introduction of appropriate instrumentation, actual measurements of electric and magnetic fields
became feasible in epidemiologic studies. Measurements were first taken for short durations in various fixed locations in the home (spot
measurements) (Savitz et al., 1988). To diminish exposure misclassification due to short-term variability, the measurements were later extended to
include 24 or 48 hours or even longer time intervals (Linet et al., 1997). Measurements taken in fixed locations, however, do not capture all
exposures encountered by subjects actively moving around in various environments. To account for all exposures experienced by study subjects,
investigators in more recent studies employed personal magnetic field exposure monitoring (McBride et al., 1999). Personal exposure monitoring means
that magnetic field meters – measuring and recording various characteristics of the encountered magnetic fields – are worn and carried by the study
subjects for 24 to 48 hours.
Both descriptive exposure assessment methods and exposure assessment methods relying on actual measurements have their
own advantages for use in epidemiologic studies. Descriptive methods, in general, do not require subject participation or consent. This
characteristic may make these methods very appealing for epidemiologists since subject participation and the resulting potential for selection and
participation bias are usually serious concerns in epidemiologic studies. With the use of wire codes and historical calculated fields it is also
possible to assess exposures that occurred in the past, in an etiologically relevant period, prior to the development of the investigated diseases.
Unfortunately, however, non-measurement exposure assessment methods provide only very rough and approximate measures of exposures. Wire code
categories and historical calculated fields tend to correlate, but only poorly with actual measured fields (Kavet, 1995).
Measurements, on the other hand, provide us with very accurate data on ambient magnetic field. In the case of measured
fields, however, the main question becomes that how well a short-term (minutes to days) measurement, taken after diagnosis, represents a long-term
(months to years) target period prior to or during disease development. Measurements also provide the investigators with data on a number of
physical characteristics of the measured fields. In the absence of any known biophysical mechanism for the carcinogenic effect of magnetic fields,
however, the investigators face the question of with metric to choose for analysis. Most frequently measures of central tendency, such as
time-weighted average, have been used in epidemiologic studies; these metrics are among the ones easiest to measure, they are fairly reproducible
over time and it is also easy to understand what they mean.
Questionnaire-based data on use of electric household appliances, and assessed parental occupational exposure to m
agnetic fields have also been used as exposure measures in a number of studies. Questionnaire-based data on appliance use, however, proved to be
fairly unreliable (Mezei et al., 2001) and parental occupational exposure may only be relevant for preconceptional and intrauterine physiologic and
pathologic changes.
Selected epidemiologic studies of childhood leukemia and magnetic field exposure
The epidemiologic study by Wertheimer and Leeper published in 1979 (Wertheimer and Leeper, 1979) was the first to raise the question of a potential carcinogenic effect of low-level power frequency magnetic fields. The study examined and found an association between childhood cancer mortality and indicators of potentially higher than average magnetic field exposure. The study was not limited to leukemia cases only, but it included all types of cancer cases, and it examined mortality rather than incident events. Exposure was assessed with a crude two-category classification of exposure based on power lines outside the children’s homes. In spite of its limitations, the study showed a 2-3-fold increase in risk of cancer development in association with higher estimated residential exposure.Expert panel evaluations
In recent years, several national and international expert panels completed their evaluations on potential health effects of power frequency magnetic field exposures. All recent expert evaluations concluded that there is an association between childhood leukemia development and exposure to power frequency magnetic fields. The United States National Institute of Environmental Health Sciences Working Group reported that there is limited evidence that residential exposure to extremely low-frequency magnetic fields is carcinogenic in children (Portier and Wolfe, 1998). The National Radiological Protection Board (NRPB) in the UK stated that relatively high average exposure to extremely low-frequency magnetic fields (0.4 µT or more) is associated with a doubling of the risk of childhood leukemia (NRPB, 2001). The International Agency for Research on Cancer (IARC) classified power-frequency magnetic field as a “possible carcinogen” (class 2B) in June 2001 (IARC, 2002). The International Commission for Non-Ionizing Radiation Protection (ICNIRP) Standing Committee on Epidemiology concluded that among all the health outcomes evaluated in epidemiologic studies of extremely low frequency of magnetic fields, the strongest evidence for an association exists between childhood leukemia and postnatal exposure to magnetic fields above 0.4 µT (Ahlbom et al., 2001).
RESEARCH DIRECTIONS
Overall, there appears to be an epidemiologic association between childhood leukemia and exposure to power frequency magnetic fields. This
epidemiologic association, however, is difficult to interpret without clear supporting evidence from experimental laboratory work and without known
biophysical mechanism (IARC, 2002). Chance was largely discounted as a likely explanation for the observed association mainly based on the results of
the pooled analyses (Ahlbom et al., 2000; Greenland et al., 2000). Although it has not been experimentally or theoretically substantiated, a potential
causal relationship, likely or not, remains a hypothetical explanation for the association.
Alternative explanations for the epidemiologic association also include confounding and bias. Confounding could occur in
an epidemiologic study if a second exposure (other than magnetic field) had a real carcinogenic effect (i.e., it could induce childhood
leukemia) and this second exposure was also associated with magnetic field exposure. If the investigator has no knowledge on the
second exposure and cannot adjust for its effect then a spurious association between magnetic field and childhood leukemia may appear in the
study, even if there is no causal relationship between them.
Bias occurs in an epidemiologic study if, due to systematic errors, the effect of exposure on the risk of the disease
is erroneously over- or underestimated. Most common concern in case-control epidemiologic studies is control selection bias, which occurs if the
ratio of the selection probabilities of exposed cases and unexposed cases is not equal to the ratio of selection probabilities of exposed controls
and unexposed controls.
Currently there is a scientific uncertainty regarding the nature of the association between magnetic field and childhood
leukemia. This uncertainty could only be resolved with new insights and new results provided by targeted and focused scientific research. Research
work being conducted at the EMF Health Assessment Program of the Electric Power Research Institute (EPRI) targets two potential explanations for
the observed epidemiologic association: 1) control selection bias; and 2) confounding by exposure to contact current.
Control selection bias hypothesis
The majority of epidemiologic studies of magnetic fields and childhood leukemia have been case-control studies where control selection bias
could be a legitimate and serious concern. Most of the expert panels have acknowledged that control selection bias cannot be ruled out as a possible
explanation for the observed epidemiologic association (Ahlbom et al., 2001; NRPB, 2001; IARC, 2002). In spite of the commonly expressed concern of
control selection bias, there is only limited data available to assess whether this type of bias really occurred in these studies, and if it
occurred then whether it explains the entire observed association or is only partially responsible.
Selection bias is hypothesized to occur through differential participation of controls and cases based on their
socio-economic status. Healthy subjects are, in general, less likely to participate in epidemiologic studies as controls than subjects with
specific diseases as cases. Participation may also vary by socioeconomic status; economically disadvantaged individuals are usually less likely to
volunteer to partake in an epidemiologic study. In addition, people in lower socioeconomic classes are thought more likely to live in homes with
higher than average magnetic field exposure. If there is a preponderance of non-participation among poorer and higher exposed controls than among
cases then it could result in a biased overestimation of relative risk in a case-control epidemiologic study.
Some evidence already exists showing that selection bias may be – at least partially – responsible for the association
found in epidemiologic studies. In their 1988 study, Savitz and colleagues (Savitz et al., 1988) reported that those subjects who did not allow
indoor measurements in their homes more frequently lived in high wire code homes than subjects who did allow indoor measurements. Hatch and
colleagues (Hatch et al., 2000), in their reanalysis of the 1997 U.S. study (Linet et al., 1997), also showed that full participants (allowing more
complete exposure assessment) were more likely to have higher socioeconomic status (higher education, higher income, and more likely to own
residence), but were less likely to live in homes with high wire codes (VHCC) or high measured fields (measurements at front door above 0.2 µT).
In addition, Hatch et al. (2000) also showed that a positive association between higher wire code (VHCC) and childhood leukemia could only be shown
when the analyses were restricted to full participants (relative risk, 1.2). When partial participants were also included in the analyses, the
association diminished (relative risk, 1.0). Michaelis et al. (1997) and Spinelli et al. (2001) reported that differential participation of cases
and controls based on their socioeconomic status might also have occurred in the Canadian and German epidemiologic studies of childhood leukemia and
magnetic field exposure (Spinelli et al., 2001; EPRI, 2003a). It was also reported, although not in childhood leukemia studies, that lower
socioeconomic status (lower family income) and higher residential mobility are also associated with higher residential wire codes (Jones et al.,
1993; Gurney et al., 1995).
In spite of the available, sporadic evidence, it is still not clear whether control selection bias could fully or
partially be responsible for the observed epidemiologic association. In a 2001 expert workshop organized in Whistler (British Columbia, Canada),
by the Electric Power Research Institute and the World Health Organization, further research to elucidate the exact role of selection bias was
recommended in several areas (EPRI, 2003a). Selection bias related research conducted by the EMF Health Assessment Program of the Electric Power
Research Institute has been developed based on the 2001 Whistler workshop recommendations and focuses on three main issues: 1) reexamining the
association between socioeconomic status and childhood leukemia, 2) inclusion of non-participants in the magnetic field–childhood leukemia analyses,
and 3) identification of highly exposed populations.
In the past, it has been commonly accepted that childhood leukemia occurs more commonly in families of higher
socioeconomic status (Greenberg and Shuster, 1985). In recent studies of magnetic fields and childhood leukemia, however, an inverse association
was observed between measures of social class and childhood leukemia (EPRI, 2003a). It needs to be clarified whether the inverse trend observed in
magnetic field studies is simply an indicator of control selection bias or it represents a real association between social class and leukemia
incidence. To shed light on the nature of the association between socioeconomic status and childhood leukemia, a thorough review of the relevant
literature is under way.
The best possible way to assess the role of selection bias and to evaluate unbiased magnetic field-related relative
risk estimates would be to include a truly random sample of the base population as controls and to include even those subjects in the analyses who
decline to participate in the study. This, however, is not possible since residential magnetic field exposure cannot be measured in the homes w
ithout the subjects’ consent. Wire code determinations and in limited instances field calculations, however, are possible for known addresses even
without the subjects’ participation. In two epidemiologic studies of childhood leukemia, the Canadian study published in 1999 (McBride et al., 1999)
and the ongoing Northern California Childhood Leukemia Study (Ma et al., 2002a, 2002b), it is possible to enlist an unbiased group of potential
controls. In these studies, controls were selected from provincial health insurance lists and lists of birth certificates, respectively, therefore
first-choice controls represent a close to ideal, and practically unbiased control group. Further analyses are planned in these studies to assess
the potential role of control selection bias by comparing wire code distributions among cases, first-choice controls, and participating controls.
In most epidemiologic studies of magnetic fields, the relative scarcity of highly exposed subjects limited the
statistical precision of relative risk estimates. In North America approximately 3%, and in Europe less than 1% of the population lives in homes
with average magnetic fields above 0.4 µT (Linet et al., 1997; UKCCS, 1999). Higher exposure prevalence may allow for a more accurate and
less bias prone assessment of exposure-related relative risks. Based on sporadic evidence, it was thought that in highly developed and densely
populated Asian regions this proportion may be significantly higher. To determine whether it is really the case, a residential magnetic field
exposure assessment survey is under way in Taiwan.
The inherent potential for control selection bias in case-control studies may even preclude scientists to arrive to a
definitive answer regarding the nature of relationship between magnetic fields and childhood leukemia using solely case-control epidemiologic studies.
To settle the question it may be necessary to conduct a large cohort study to examine the association. The rarity of childhood leukemia and the low
prevalence of exposure above 0.4 µT, however, would require such a large cohort size that such a study of this kind could be extremely
and inhibitively expensive.
Contact current hypothesis
Contact current occurs when two or more extremities (e.g., a hand and a foot) touch conductive surfaces that are at different electrical
potentials. Because the body is electrically conductive, current flows so long as contact is maintained. Much of the body’s electrical resistance
is concentrated in the epidermal layer under dry conditions; under moist or wet conditions, the skin’s resistance decreases dramatically and a
body’s resistance is on the order of 1 to 3 kohms (kΩ), depending on body dimensions, fat and muscle content, etc. (Reilly, 1998).
Research conducted over the past few years has identified contact current as an exposure that may plausibly explain the
association of childhood leukemia with residential magnetic fields. Thus far, in the United States (US) the research has considered conditions in
residential electrical systems that may predispose to contact current exposure, with an emphasis on single, detached homes. In the US, the National
Electrical Code requires, for electrical and fire safety, that a grounding wire connect the service panel neutral to the residential water line (if
the latter is conductive). Consequently, the water line acquires a small voltage, VW-E, relative to the earth (usually less than
1 volt). This voltage is affected by current in the primary and secondary grounding systems, as well as by inductive effects from heavily
loaded lines nearby.
If the drainpipe from the bathtub drain into the earth is conductive, then VW-E can serve as a voltage
source that drives current into a bathing child touching the faucet, the spout, or the water stream. This voltage from the point of contact to the
drain, Vbath, is a variable fraction of VW-E that depends on conditions in the soil beneath the residence and on
the child’s electrical resistance. In the absence of both a conductive water line and a conductive drainpipe, exposure does not occur.
Two measurement studies have examined the relationship between both VW-E and Vbath
and the average residential magnetic field, Bavg. A methods development study in a small sample of residences in Pittsfield, MA
reported a statistically significant positive correlation between VW-E and Bavg (Kavet and Zaffanella, 2002).
A larger study in the Denver area at the site of the studies of Wertheimer and Leeper (1979) and Savitz et al. (1988), reported a statistically
significant association between VW-E and Bavg and statistically significant upward trends of association across
Bavg strata for both VW-E and Vbath (Kavet et al., in press). Subsequent computer
modeling of typical US neighborhoods predicts positive association between VW-E and Bavg similar to those
measured (Kavet, submitted).
In the Denver study Vbath was estimated by inserting a 1,000-ohm resistor between the faucet and the
drain. About 12% of the homes had voltage levels that could produce up to 10s of microamperes into a child’s hand while making contact. Dosimetry
modeling estimates that in a small child, currents of this magnitude can produce dose to bone marrow of 100s of millivolts per meter (Dawson et al.,
2001), far greater than doses induced by ambient residential magnetic fields, and much greater than the lowest doses associated with plausible
experimentally induced biological effects. Both the National Institute of Environmental Health Sciences (NIEHS) in the US and the National
Radiological Protection Board (NRPB) in the United Kingdom (UK) identify 100 µT and greater as an exposure range associated with biological effects,
although the two differ with respect to their respective certainties regarding the validity of the experimental studies (NIEHS, 1999; NRPB, 2001).
This exposure level crudely translates to induced field strengths on the order of 1 mV/m. Although experiments relevant to leukemogenesis in
appropriate model systems are yet to be conducted, the exposure scenario described for contact currents projects electric fields in bone marrow on
the order of 100s of mV/m, which are dosimetrically plausible levels.
If the average residential magnetic field has been serving as a surrogate for another exposure, the latter must produce
adequate dose to the target tissue, and must be positively associated with the average residential magnetic field. Additionally, children must have
access to exposure on a reasonably regular basis. As the discussion above indicates, exposure to contact current has demonstrated features that
qualify it as a plausible factor that could be responsible for the association of childhood leukemia with residential magnetic fields. The
observation that young children bathe frequently and further engage in behavior that produces contact with potential sources of exposure
(P. Buffler and M. Does, personal communication) further legitimizes the contact current hypothesis’ plausibility. The role of contact
current is currently under investigation in the Northern California Childhood Leukemia Study (e.g., see Ma et al., 2002a, 2002b) with results
anticipated in the 2006–2007 time frame.
CONCLUSION
An epidemiologic association has been reported between residential magnetic field exposure and childhood leukemia development. Causal relationship, however, cannot be established without clear supporting evidence from experimental laboratory work and without known biophysical mechanism. Chance is now considered an unlikely explanation for the observed association. Potential alternative explanations include systematic epidemiologic error (i.e., control selection bias) and confounding due to a second exposure (i.e., contact current) related to magnetic field exposure. The scientific uncertainty regarding the nature of the association between childhood leukemia and residential magnetic field exposure could only be resolved with further targeted and focused scientific research work in these areas.
REFERENCES
AHLBOM, A., CARDIS, E., GREEN, A., LINET, M., SAVITZ, D., and SWERDLOW, A. (2001). “Review of the epidemiologic literature on EMF and health.” Environ. Health Perspect. 109 (Suppl. 6):911–933.
AHLBOM, A., DAY, N., FEYCHTING, M., ROMAN, E., SKINNER, J., DOCKERTY, J., LINET, M., MCBRIDE, M., MICHAELIS, J., OLSEN, J. H., TYNES, T, and VERKASALO, P. K. (2000). “A pooled analysis of magnetic fields and childhood leukaemia.” Br. J. Cancer 83:692–698.
DAWSON, T. W., CAPUTA, K., STUCHLY, M. A., and KAVET, R. (2001). “Electric fields in the human body resulting from 60-Hz contact currents.” IEEE Transact. Biomed. Eng. 48:1020–1026.
EPRI. (2003a). Selection Bias in Epidemiologic Studies of EMF and Childhood Leukemia. Proceedings of the 2001 Whistler Workshop. Technical Report 1008149. Electric Power Research Institute, Palo Alto, CA.
EPRI. (2003b). Analyses of Magnetic-Field Peak-Exposure Summary Measures. Technical Report 1008152. Electric Power Research Institute, Palo Alto, CA.
FEYCHTING, M. and AHLBOM, A. (1993). “Magnetic fields and cancer in children residing near Swedish high-voltage power lines.” Am. J. Epidemiol. 138:467–481.
FOLIART, D. E., IRIYE, R. N., SILVA, J. M., MEZEI, G., TARR, K. J., and EBI, K. L. (2002). “Correlation of year-to-year magnetic field exposure metrics among children in a leukemia survival study.” J. of Expos. Anal. Environ. Epidemiol. 12:441–447.
GREENBERG, R. S. and SHUSTER, J. L. (1985). “Epidemiology of cancer in children.” Epidemiol. Rev. 7:22–48.
GREENLAND, S., SHEPPARD, A. R., KAUNE, W. T., POOLE, C., and KELSH, M. A. (2000). “A pooled analysis of magnetic fields, wire codes, and childhood leukemia.” Epidemiol. 11:624–634.
GURNEY, J. G., DAVIS, S., SCHWARTZ, S. M., MUELLER, B. A., KAUNE, W. T., and STEVENS, R. G. (1995). “Childhood cancer occurrence in relation to power line configurations: a study of potential selection bias in case-control studies.” Epidemiol. 6:31–35.
HATCH, E. E., KLEINERMAN, R. A., LINET, M. S., TARONE, R. E., KAUNE, W. T., AUVINEN, A., BARIS, D., ROBISON, L. L., and WACHOLDER, S. (2000). “Do confounding or selection factors of residential wiring codes and magnetic fields distort findings of electromagnetic fields studies?” Epidemiol. 11:189–198.
IARC (2002). Static and Extremely Low-Frequency Electric and Magnetic Fields. IARC Monographs on the Evaluation of the Carcinogenic Risks to Humans. Vol. 80. International Agency for Research on Cancer. Lyon, France.
JONES, T. L., SHIH, C. H., THURSTON, D. H, WARE, B. J., and COLE, P. (1993). “Selection bias from differential residential mobility as an explanation for associations of wire codes with childhood cancer.” J. Clin. Epidemiol. 46:545–548.
KAVET, R. (1995). Magnetic field exposure assessment. In: Electromagnetic Fields: Biological Interactions and Mechanisms. M Blank (eds) American Chemical Society, Washington, DC, pp. 191–223.
KAVET, R. and ZAFFANELLA, L. E. (2002). “Contact voltage measured in residences: implications to the association between magnetic fields and childhood leukemia.” Bioelectromagnetics 23: 464–74.
KAVET, R., ZAFFANELLA, L. E., PEARSON, R. L., and DALLAPIAZZA, J. “The association of residential magnetic fields with contact voltage.” Bioelectromagnetics, In press.
LINET, M. S., HATCH, E. E., KLEINERMANN, R. A., ROBISON, L. C., KAUNE, W. T., FRIEDMAN, D. R., SEVERSON, R. K., HAINES, C. M., HEARTSOCK, C. T., NIVA, S., WACHOLDER, S., and TARONE, R. E. (1997). “Residential exposure to magnetic fields and acute lymphoblastic leukemia in children.” N. Eng. J. Med. 337:1–7.
LONDON, S. J., THOMAS, D. C., BOWMAN, J. D., SOBEL, E., CHENG, T-C., and PETERS, J. M. (1991). “Exposure to residential electric and magnetic fields and risk of childhood leukemia.” Am. J. Epidemiol. 134:923–937.
MA, X., BUFFLER, P. A., GUNIER, R. B., DAHL, G., SMITH, M. T., REINIER, K., and REYNOLDS, P. (2002a). “Critical windows of exposure to household pesticides and risk of childhood leukemia.” Environ. Health Perspect. 110:955–60.
MA, X., BUFFLER, P. A., SELVIN, S., MATTHAY, K. K., WIENCKE, J. K., WIEMELS, J, L., and REYNOLDS, P. (2002b). “Daycare attendance and risk of childhood acute lymphoblastic leukaemia.” Br. J. Cancer 86:1419–24.
MCBRIDE, M. L., GALLAGHER, R. P., THÉRIAULT, H. G., ARMSTRONG, B. G., TAMARO, S., SPINELLI, J. J., DEADMAN, J. E., FINCHAM, S., ROBSON, D., and CHOI, W. (1999). “Power-frequency electric and magnetic fields and risk of childhood leukemia in Canada.” Am. J. Epidemiol. 149:831–842.
MEZEI, G., KHEIFETS, L. I., NELSON, L. M., MILLS, K. M., IRIYE, R., and KELSEY, J. L. (2001). “Household appliance use and residential exposure to 60-Hz magnetic fields.” J. Expos. Anal. Environ. Epidemiol. 11:41–49.
MICHAELIS, J., SCHÜZ, J., MEINERT, R., MENGER, M., GRIGAT, J. P, KAATSCH, P., KALETSCH, U., MIESNER, A., STAMM, A. BRINKMANN, K., and KARNER, H. (1997). “Childhood leukemia and electromagnetic fields: Results of a population-based case-control study in Germany.” Cancer Causes Control 8:167–174.
NIEHS (1999). Health Effects from Exposure to Power-Line Frequency Electric and Magnetic Fields., NIH Publication No. 99-4493. National Institute of Environmental Health Sciences, Research Triangle Park, NC.
NRPB (2001). ELF Electromagnetic Fields and the Risk of Cancer. Documents of the NRPB. Vol. 12 (No. 1). National Radiological Protection Board. Clinton, UK.
PORTIER, C. J. and WOLFE, M. S. (1998). Assessment of Health Effects form Exposure to Power-Line Frequency Electric and Magnetic Fields. NIEHS Working Group Report. NIH Publication No. 98-3981. National Institute of Environmental Health Sciences, Research Triangle Park, NC.
REILLY, J. P. (1998). Applied Bioelectricity: From Electrical Stimulation to Electropathology. New York: Springer-Verlag.
SAVITZ, D., WACHTEL, H., BARNES, F. A., JOHN, E. M., and TVRDIK, J. G.. (1988). “Case-control study of childhood cancer and exposure to 60-Hz magnetic fields.” Am. J. Epidemiol. 128:21–38.
SPINELLI, J. J., MCBRIDE, M. L., ABANTO, Z. U., TAMARO, S., BAJDIK, C. D., and GALLAGHER, R. P. (2001). “Assessing response bias as an explanation for the observation of low socioeconomic status (SES) as a risk factor for childhood leukemia.” Am. J. Epidemol. 153: S254.
UKCCS (1999). UK Childhood Cancer Study Investigators. Exposure to power-frequency magnetic fields and the risk of childhood cancer. Lancet 354:1925–1931.
UKCCS (2000). UK Childhood Cancer Study Investigators. Childhood cancer and residential proximity to power lines. Br. J. Cancer 82:1073–1102.
WERTHEIMER, N. and LEEPER, E. (1979). “Electrical wiring configurations and childhood cancer.” Am. J. Epidemiol. 109:273–284.