teratogen update: electromagnetic fields

TERATOLOGY 54:305-313 (1996)

Teratogen Update: Electromagnetic Fields ELISABETH ROBERT* Institut European des Genomutations, 69005 Lyon, France

ABSTRACT Public concern is increasing about the potential health effects of extremely low frequency (ELF) electromagnetic fields (EMFs) naturally present or generated by electrical appliances and those of very low frequency (VLF) fields, like those generated by video display terminals (VDTs). There are arguments in favour of EMFs being biologically active but no mechanism has been identified that explains the link between EMFs and bioeffects. More than 50 studies on exposures of animals to EMFs have been performed within the last few years. Although there were a few statistically significant effects in the studies reviewed, no replicable results were found among laboratories. The extent of effects observed, if any, was always low, and deserves further investigation before relevance to humans can be considered. Human data reviewed concern the potential reproductive effects (mainly spontaneous abortions, low birthweight, and congenital malformations) of exposure to various sources of EMFs: maternal residence, heated waterbeds, electric blankets, and ceiling heating coils, occupational exposure (mainly video display terminals), and magnetic resonance imaging. The totality of the evidence that is thus far available provides no convincing evidence to indicate that low frequency EMFs of the sort that might be met in occupational or daily life exposures does any harm to the human reproductive process. It is suggested that those counseling pregnant women follow the guidelines established by WHO in agreement with the International Non-Ionizing Radiation Committee. This group does not consider that the results of published studies provide a basis for restricting human exposure to electromagnetic fields and radiation. Tera- tology 54:305-313, 1996. o 1997 ~ i ~ e y - ~ i s s , Inc.

~ ~

Worldwide, as society has industrialized and needs for electrical power have increased, so has public concern about the potential health effects of extremely low frequency (ELF) electromagnetic fields (EMFs), naturally present or generated by a large number of common electrical appliances in the home and office (Breysse et al., '94), listed in Table 1.

Within the last decade several epidemiologic studies have suggested that EMFs could have health effects, mostly increase the risk of cancer. A more limited literature is available on potential reproductive hazards associated with EMFs. Juutilainen ('91), Chernoff et al. ('92), and Brent et al. ('93) have assembled detailed reviews of the available animal and human

studies. The reader is invited to refer to these articles for completeness, as the goal of the present paper is to make a summary of the available information and to stimulate ideas as to the directions needed in future studies.

PHYSICAL CHARACTERISTICS OF ENVIRONMENTAL EMFs

Electric and magnetic fields propogate at different frequencies (energies) and the biologic effects vary with frequency. Referring to Figure 1, it can be seen that these frequencies span the range from ionizing radiations, such as cosmic and X-rays (over 10l8 Hz), which are known to damage cells, to ELF magnetic fields (below 300 Hz), such as those generated by electric power transmission, distribution, and use. This review also deals with very low frequency (VLF) fields (3,000- 30,000 Vlm), such as those generated by video display terminals (VDTs), not including frequencies associated with microwave ovens, satellite and radio or television waves, or cellular telephones (over lo5 Hz). Naturally occurring electric and magnetic fields are present every- where, and include the earth's magnetic field and fields associated with natural deposits of magnetite, the flux density of which ranges from 30 to 70 pT, i.e., lower than artificially produced EMFs by several orders of magnitude. Table 2 indicates the different units used for strength or intensity of electric and magnetic fields.

An electric field exists near a voltage source irrespec- tive of whether an electrical current is flowing through it, A magnetic field exists near a voltage source only when electric charge flows through the source to consti- tute an electrical current. The intensity of a magnetic field near the source varies directly with the amount of current and inversely with the distance from the source (Fig. 2). Magnetic fields are either pulsed or sinusoidal. Pulsed magnetic fields usually have higher frequencies than sinusoidal ones, and induced currents that they generate are of greater density, with potentially more biologic effects.

BIOLOGICAL BACKGROUND Bassett ('93) indicated that EMFs can affect biologi-

cal processes in beneficial ways. For example, it is

*Correspondence to: Elisabeth Robert, Institut European des Genomu- tations, 86 Rue Docteur Edmond Locard, 69005 Lyon, France.

Received 5 December 1996; Accepted 9 December 1996

o 1997 WILEY-LISS, INC.

306 E. ROBERT

TABLE 1. Electric and magnetic field sources (ELF and VLF)

Examples of exposure levels

kVlm pT

Earth magnetic field Solar flares Van Allen belt Magnetic resonance imaging Power lines (400 kV, at 30 m from the axis) Motors and generators

0.02-6 1-30

Transformers, electrical distribution panels, switch gear

Electric appliances Electric tools <0.1 0.1-800 Electric blankets, heating pads, water

Electric resistance heating Fluorescent lighting Electric (analog) clocks

Metal water pipes, gas lines, cable TV,

bed heaters 0.25 3.5

Home and commercial building wiring

telephone cables Radio and television transmit towers Computers and video displav terminals 0.03 0.6-1.4

postulated that therapeutic devices based on low frequency pulsed EMFs accelerate the healing of bone fractures; however, this idea remains controversial. To establish plausibility for both therapeutic and other biologic phenomena due to EMFs, it is then essential to show an effect at the cellular level. Specifically, the effect varies with the square of the distance from the source, which means that doubling the distance lowers the intensity by l/4. Although few hypotheses have been tested thus far, there are a number of biological effects of EMFs reported in the literature that might provide the basis for designing experiments and epidemiologic studies. According to Stevens ('93), the cellular effects could be of three types: effects on DNA transcription and translation, calcium balance in cellular inter- communications, and pineal production of melatonin.

Alterations in DNA transcription and translation could have pleiotropic effects that may affect the embryo at any phase of cell division. It is generally accepted that EMFs do not have enough energy to cause mutagenic damage to DNA (Desjobert et al., '95). Goodman and Henderson ('88, '91) demonstrated changes in the transcription and translation stages of protein synthesis following exposure to weak EMFs, but neither Nafziger et al. ('93) nor Lacy Hulbert et al. ('95) could confirm these effects.

Calcium ion flow across the cell membrane governs many physiologic processes, such as egg fertilization, and cell division (Blackman et al., '88). Disruption of calcium homeostasis has many implications including increasing oxidative stress and apoptosis which are suspected of representing the embryotoxicity caused by a number of xenobiotics (Davies et al., '90; DeSesso et al., '94) and of being a probable cause of retarded

embryo development. No causal linkage between EMF and these cellular effects has ever been substantiated.

The pineal gland is normally responsive to visible EMFs (i.e., light), since the eyes are reactionally con- nected to the pineal gland by a series of neurons (Reiter, '93). Melatonin has many functions in the organism, including a potent antioxidant. Any perturbation which causes levels of melatonin to be lower than normal may have significant physiological consequences.

ANIMAL STUDIES Potential biologic effects of EMFs are likely to be

proportional to induced currents in the living organism. Kaune and Phillips ('80) have shown that, with the same field, the currents within a person are considerably higher than those in a quadruped. An average- sized adult placed in a 1 kV/m electric field collects approximately 15 pA with a current density at the waist of 0.15-0.25 mA/m2. In contrast, current density in analogous anatomic sites in the rat and pig would be about 0.002 mA/m2, respectively. To extrapolate experimental electric fields to current density levels induced in humans from environmental electric fields, values must be scaled upward to appropriate levels. Similarly, since magnetically induced current density is proportional to the linear dimension of the exposed subject, experiments with laboratory species must scale up magnetic field strength to achieve the induced current levels equivalent to those that magnetic fields induce in humans. Chernoff et al. ('92) reviewed about 20 articles on laboratory studies on electric fields, performed in nonmammals and mammals. The most detailed work was performed by Rommerein et al. ('go), who exposed pregnant rats to electric fields of different intensities up to 150 kV/m for 19 Wday for a period beginning 1 month prior to breeding and continuing through breeding, completion of gestation, and the rearing of offspring until weaning. Male rats were also tested with respect to fertility and male-mediated teratogenesis. Reproduc- tive outcomes were unaffected by all of these exposures. Most of the other studies were similarly negative. One exception was that of Cerretelli et al. ('79), who exposed male rats to 100 kV/m to study fertility and fetal outcome after their mating to control females. Male reproductive function was not affected, but the investigators did find a significant decrease of fetal weight in litters sired by males exposed to the fields for 8 Wday.

Almost 30 studies of exposures of animals to magnetic fields have been performed within the last 5 years. A study of dairy cattle living near a high voltage power line in Minnesota revealed no apparent reproductive effects related to proximity to the line (Martin et al., '86). Ryan et al. ('95) performed a study using magnetic fields of up to 1 mT generated by a 60 Hz power source and found no adverse effects on rat embryonic development.

Wilson et al. ('86) and Reiter et al. ('88) reported a reduction in the nocturnal rise of pineal melatonin levels, in adult rats and their offspring, respectively,

TERATOGEN UPDATE: ELECTROMAGNETIC FIELDS 307

INFRA 9 RED m ULTRA

VIOLET MICROWAVES

X-RAYS

I 014 1015

GAMMA RAYS

Fig. 1. The electromagnetic spectrum. Nonionizing radiations are found in the frequency range below X-rays. The electromagnetic waves that have been the subject of the work reviewed in this paper are those well below the visible range.

TABLE 2. Magnetic and electric field units1

SI CGS ~

Electric field Field intensitv Volts Der meter Volts Der centimeter

(vim) (V/im) 1 kV/m = 1,OOOVlm 1VIcm = 10OVlml

Mametic field Flux density Tesla (T) Gauss (G)

1 pT = 0.000001 T (1 pT = 10 mG)

Field intensity Amperes per meter Oersted (Oe)

l N m = 1.26uT

1 mG = 0.001 G

W m ) 1 Oe = 79.6Alm

%I = International System of Units; CGS = Centimeter-Gram- Second Unit.

when exposed to 60 Hz electric fields in utero and for 23 days after birth, but this finding was not confirmed in other species. The physiologic significance of this observation, if any, has not been established. Delgado et al. (’82) exposed chick embryos to weak-pulsed magnetic fields and concluded that a field of 100 Hz, 1 pT, had a “powerful” effect on inducing abnormalities. It has not been possible to repeat these experiments in mammals. Additional studies investigated the effects of time- varying magnetic fields of the type associated with VDTs. Pregnant rats and mice were exposed t o field intensities much higher than those produced by terminals. There was no evidence of developmental toxicity associated with this exposure in rats (Stuchly et al., ’88) or mice (Wiley et al., ’90, ’92).

z n -I

LL w u

40 I 30 c P O 0 kV

20

10

I) - 0 5 10 * 15 20

DISTANCE FROM CENTRE LINE IN METRES 25

Fig. 2. The magnetic field decreases rapidly with distance from the center line of overhead power lines.

Picazo et al. (’95) studied the effects of chronic exposure to EMFs on the reproductive system of fetal mice. These investigators observed a tendency toward heavier and larger ovaries in exposed animals as a consequence of thick corpora lutea. These differences between exposed and nonexposed animals were interpreted as possibly related to a less developed endome- trium, which could increase the risk of lower birth weight in neonates and result in an increased mortality rate. None of the alleged links between the finding at different stages of the reproductive process are biologically proven, and if links exist, there would be no proof of causality.

308 E. ROBERT

Experimental evidence showing safety of magnetic resonance imaging (MRI) for the embryo has been published infrequently. Studies of bacteria and of human lymphocytes in culture failed to show a mutagenic effect (Thomas and Morris, '81; Cooke and Morris, '81). Belton et al. ('84) showed that MRI of chick eggs is compatible with normal development. Pregnant mice subjected to MRI did not show adverse effects on litter size or growth (McRobbie and Foster, '85). Tyndall and Sulik ('91) studied a strain of mice genetically prone to eye malformations and found an increase in eye abnormalities after exposure to MRI fields. This study was limited, however, because it failed to quantitate the exposure received by the animals and did not control rigorously for the possibility of tissue heating, which has previously been associated with the induction of malformations in experiments with nonionizing radiation. Foster et al.'s study ('83) of MRI in pregnant goats is often cited as providing evidence of safety; however, only four animals were used and reproductive toxicity evaluation of the offspring was not reported. Mevissen et al. ('94) exposed rats to static and alternating magnetic fields with a flux density of 30 mT from day 1 to day 20 of gestation to simulate exposures of pregnant operators of MRI devices. For comparison with effects of this static field, a time-varying (50 Hz) magnetic field of the same flux density was used. A total of 48 rats were studied: 24 were exposed (6 to static fields, 6 to alternating fields) and 24 were unexposed. No teratogenic effect was found, and the only observed effect was a slight, but significant, decrease in the mean number of living fetuses with an increase in the number of resorptions (14% to 4% in controls) in the group with static EMF exposure but not in the group with alternating EMF exposure (6% VS. 3% in controls). The authors interpret their findings as linked with the fact that static EMF of such flux density may induce embryotoxic effects, while alternating field effect exposure does not seem to be associated with any severe reproductive risk.

Although there were a few statistically significant effects of EMF on laboratory animals in the studies reviewed, no replicable results were found among laboratories, and even within the same laboratory when two subsequent studies were published. The extent of effects observed, if any, was always low, and deserves further investigation before relevance to humans can be considered.

HUMAN DATA The following studies concern the potential effect of

exposure to various sources of EMFs that concern primarily spontaneous abortions, low birth weight, and congenital malformations.

Maternal residential magnetic field exposure The possibility of an association of early pregnancy

loss (EPL) with residential exposure to ELF magnetic fields was investigated by Matilainen et al. ('90) and Juutilainen et al. ('93) in case-control studies. In the

latter, 89 cases and 102 controls were obtained from an earlier study aimed at investigating the occurrence of EPL in a group of women attempting to become pregnant. Magnetic field exposure was characterized by measurements in residences. Strong magnetic fields were measured more often in case than in control residences. In an analysis based on fields measured at the front door, a cutoff score of 0.63 pT resulted in an odds ratio (OR) of 5.1 [95% confidence interval (CI): 1.0-251. The results should be interpreted cautiously due to the small number of highly exposed subjects and other limitations of the data. Robert ('93) performed a preliminary study in France that did not identify an excess of birth defects among children whose parents live within 500 m of a high voltage power line (HVPL). It was pointed out that the field strength drops off rapidly with distance from the line and that few people live directly below a power line; therefore, it is unlikely that most of these children were exposed in utero to EMFs much different from "nonexposed" children.

Savitz and Ananth ('94) used data collected for a study of childhood cancer to examine the relationship between measured residential EMFs and wire codes to pregnancy outcome. Data consisted of interviewed cases and controls. Pregnancies in homes with measured fields above 0.2 pT or high wire codes were not more likely than others to end in miscarriage, low birth weight, or preterm delivery. According to the authors, as the sample was not specifically collected to test this hypothesis, lack of data on potential confounders and small numbers of cases limit the study's conclusion.

More recently, Robert et al. ('96) conducted a matched case-control study in the same region of France to explore in further detail whether living closer to HWLs increased the risk of congenital anomalies. For every case and control the distance from the HVPL to the maternal residence at the time of birth was measured as a surrogate of EMF exposure. Using 100 m as the cutoff point between exposure and nonexposure yielded an OR of 0.95 (95% CI: 0.45-3.22). Among the 11 malformed infants within 100 m, there were 2 children with Down's syndrome, but otherwise there was no pattern in the occurrence of specific anomalies. Pa- tients included in this study were not actually exposed to EMFs if they lived at a distance more than 25 m from the center line of overhead power lines. People living at a shorter distance from a line are so few that no epidemiological study can have enough statistical power to determine whether the prevalence of a specific congenital anomaly is significantly increased as a result of living near a HVPL. One can only conclude that, in this sample, there were no differences in residential proximity to HVPLs between malformed and control infants.

Heated waterbeds, electric blankets, and ceiling heating coils

Wertheimer and Leeper ('86) studied a sample of 1,256 births to 4,271 women who delivered in two


Denver area hospitals in 1982. They obtained information on use of electrically heated beds (electric blankets and heated waterbeds) from a telephone survey and attempted to determine whether such exposures influ- enced reproductive parameters: gestational length, birth weight, miscarriage rate, and birth defects. They found an increase in spontaneous pregnancy loss during seasons of use and concluded that either thermal or EMF effects might be involved in human reproductive wastage. The same authors ('89) showed subsequently a similar seasonal variation in pregnancy loss associated with the use of ceiling heating. Both of these papers were reviewed extensively by Chernoff et al. ('92) and severe design problems were underlined: more than 70% of eligible births in the study were missing, the rate of congenital malformations in the reference group (1 in 335) was abnormally low, and the method for studying the miscarriage rate was inappropriate. In both papers the authors tried to adapt their observa- tions to a number of speculations about biologic effects of EMFs on fetal loss, but their findings do not demon- strate that the use of electric blankets or heated waterbeds increases the risk of adverse reproductive outcome.

Dlugosz et al. ('92) in an epidemiologic study in New York State did not find that electric bed heating increased the risk of congenital defects or fetal loss. Li et al. ('95) interviewed 118 pregnant women who had had a child born with a congenital urinary tract anomaly, and a comparison group of 369 mothers of control infants. They found an association between electric blanket use during the first trimester of pregnancy and congenital urinary tract abnormalities (OR = 10; 95% CI: 1.2-85.5) in women with a history of subfertility. The participation rate in this study was relatively low (62.5% for cases and 67.6% for controls) and the results relied only on three exposed cases; a selection bias cannot be excluded. No association was found in this study with electrically heated waterbeds or work with VDTs. The authors concluded that only a very sensitive population of subfertile women would be likely to be affected by electric blankets which generate magnetic fields 5 times higher than electrically heated waterbeds or VDTs and are used longer and have a closer proximity to the target (in utero embryo). Many maternal characteristics were analyzed as subsamples. For the above reasons, the conclusions of the study are questionable and confirming data are needed.

Bracken et al. ('95) conducted a prospective study to evaluate the relation of birth weight and fetal growth retardation with use of electrically heated beds (electric blankets or heated waterbeds) during pregnancy. None of the exposure measures showed a dose-response relation to these outcomes.

Occupational exposure Nordstrom et al. ('83) found an increase in the

incidence of congenital malformations in a survey of 372 married couples in which the male worked at one of

two Swedish power companies between 1953 and 1979. The finding could not be explained by confounding or reporting biases, but it is not clear by what mechanism such adverse outcomes might have been transmitted via the father. No consistent pattern of malformations was evident. Coleman and Beral ('88) suggested that this was a chance association. Lundsberg et al. ('95) conducted a nested case-control study to compare men with abnormal semen parameters and controls. They showed no association between occupationally related categories of magnetic field exposure and male subfertility, as evaluated by morphology, motility, and concentra- tion of sperm. These findings do not support theories of deleterious effects to male reproductive health from magnetic field exposure.

Working with VDTs The issue of VDTs as possible reproductive hazards

has been much discussed in recent years. The first statements made to that effect were met with absolute denial from officials and scientists because there was no measurable ionizing radiation around the VDTs which could reasonably affect the embryo or the fetus. VDTs are similar to the video screens of television sets. Many VDTs involve the modulation of a scanning electron beam targeted on the surface of a fluorescent tube. The magnetic flux density a t the very low frequency corresponding to the screen scanning beam (15-30 kHz) would be between 0.01 and 0.2 pT with occasional values up to 0.6 pT and between 0.07 and 0.7 pT with occasional values above 1 pT at the extremely low frequencies corresponding to the electric main supply of alternating current (50-80 Hz) (Advisory Group on Non-Ionizing Radiations, '94). These ELF fields are no greater than those produced by other domestic appliances. Although early television sets were found to be a significant source of ionizing radiation, beginning in the early 1960s the use of leaded glass to fabricate picture tubes eliminated this source of radiation. AS explained by Blackwell and Chang ('88), some newer television and computer systems are using smaller and lighter liquid crystal displays, which are much simpler than VDTs and emit little radiation of any sort. These authors, as well as Marha and Charron ('83, note that other forms of electromagnetic radiation emitted by VDTs have been measured for a variety of the commercial units on the market. By measuring radiation at the back and sides of the units, attention has also been given to the possible exposure to people positioned near the VDT. These two studies, as well as a number of others, confirm the finding that nonionizing radiation and magnetic fields associated with these units are not produced in quantities that are significant biologically (Liden et al., '86; Juutilainen and Saali, '86; Marriott and Stuchly, '86; Murray, '86).

The first major concern about the potential teratogenicity of VDTs in Canada occurred in 1980 when a cluster of four infants with severe malformations was described. The mothers had worked at the same place, a

310 E. ROBERT

newspaper department in Toronto. The cluster was linked to the fact that the women had worked with VDTs during pregnancy. The publication of this cluster in the Toronto Globe and Mail soon brought forward reports on other clusters of reproductive failure in North America, reviewed by Bergqvist ('84). A cluster of miscarriages was subsequently investigated at the General Telephone Company of Michigan (Lichty, '85), where 6 of 29 pregnancies in VDT-exposed women aborted spontaneously compared to 8 miscarriages in 97 pregnancies not exposed to VDTs. Although this difference was statistically significant, the author left open the possibility that work-related factors other than VDT exposure might be involved because the jobs performed by VDT-exposed and nonexposed women at this company differed considerably.

Clusters of adverse reproductive outcomes should be expected to occur at different work places, including numerous places where VDT work is common. Bergqvist ('84) published mathematical models of clusters trying to estimate whether the reported number of clusters was larger than reasonable from mere random distribu- tions and this latter model was judged to be true. Abenhaim and Lert ('91) performed an analysis of case clusters in an office setting that included VDT exposures and reached the same conclusions.

In addition to the preceding report on clusters of miscarriages, four case-control studies have been un- able to find an association between congenital anomalies and exposure of the mother to VDTs (Kurppa et al., '85; McDonald et al., '86; Ericson and KiillBn, '86a; Tikkanen et al., '90). A significant excess of hydrocepha- lus was reported as a result of a case-control study performed by Brandt and Nielsen ('go), but no similar excess was reported by any of the other investigators. The case-control study by Ericson and G l l h ('86b) included miscarriages, perinatal deaths, severe malformations, and low birth weight infants as index cases; a total of 522 such cases and 1,032 matched controls were studied. When the possible effect of VDT work on poor reproductive outcome was analyzed, a significantly increased risk was registered for birth defects, but these effects were reduced and lost statistical significance when stratification for smoking and stress was made. There was no detectable effect on miscarriage rate. This epidemiologic study illustrated how hazardous it is to draw quick conclusions from a few positive findings. For instance, in a large case-control study by Goldhaber et al. ('881, there was a small, but significantly elevated risk of miscarriage for women who reported using VDTs for more than 20 Wweek during the first trimester of pregnancy. As is true for any retrospective method of investigation, there was concern that recall bias may have impaired data collection in this study. Women were questioned about their VDT use more than 2 years after the pregnancies in question. It is also possible that estimates regarding exposure to VDTs were not accurately recalled by the subjects, and as stated by Robinson ('89), pregnancy outcome may have distorted the recollection of VDT expo-

sures. The increased miscarriage rate may have also been due to unmeasured factors confounded with high VDT use, such as poor ergonomic conditions or job-related stress. This concept was supported by the hding in the case- control study that miscarriages were increased in certain job categories without regard to VDT exposure. McDiarmid et al. ('94) made a comparison between apparent associa- tions with VDT exposure that were established with either retrospective or prospective data collection. The prospec- tively collected data did not support the data colleded retrospectively, which implicated recall bias as a codound- ing factor in the retrospective data.

Schnorr et al. ('91) performed a study at the National Institute for Occupational Safety and Health in which they monitored the incidence of spontaneous abortion in 882 pregnancies that included occupational use of VDTs during gestation. The data in this very thorough investigation did not indicate any association between the use of VDTs and exposure to the accompanying EMFs with an increased risk of spontaneous abortion. Similar negative findings were also reported by Lind- bohm et al. ('92), and by Roman et al. ('921, who performed a case-control study that was designed to minimize the possible role of nonoccupational factors in the incidence of spontaneous abortions. Although the study by Lindbohm et al. ('92) did not find an overall increased risk of spontaneous abortion associated with use of VDTs, they did find an increased risk for women who worked with terminals that emitted a high level of ELF magnetic fields. The small number of subjects (less than 20 per group) makes the significance of this observation questionable.

In 1987, the Council on Scientific Maim ('87) stated officially that available data were not sufficient to rule out all possibility that a factor associated with VDT use may be hazardous to pregnancy. The meta-analysis was interpreted by Parazzini et al. ('93) as showing that the magnitude of the risk from the VDT itself, if a risk exists at all, is quite small. With the growing use of computer technology around the world, we should expect the continuing appearance of isolated and anec- dotal case reports of a malformed infant born to a woman who had VDT exposure (Kultursay et al., '94). Rodriguez-Pinilla and Martinez-Frias ('951, who had rejected the hypothesis in a case-control analysis of a sample of malformations drawn from the Spanish registry, urged caution in interpreting case reports.

Infante-Rivard ('95) carried out another case-control study in Spain and suggested that in utero exposure to EMFs generated by sewing machines that were used by pregnant women might be a risk factor for childhood leukemia. This would indicate that EMFs would have a transplacental carcinogenic action, which needs to be further explored.

MRI According to the International Non-Ionizing Radia-

tion Committee of the International Radiation Protec- tion Association ('911, MRI is a tissue imaging technique that uses a magnetic field and radiofrequency


radiation. Originally called “nuclear magnetic resonance (NMR),” this name was abandoned to avoid the misleading implication that this imaging method in- volves nuclear radiation. There is no potential for ionization of chemical species with this technique (as there is with X-ray). Therefore, MRI may entail less theoretical risk for the developing embryo. The background magnetic field ranges from 30 to 70 pT. The patient’s exposure during MRI is up to about 2 T, and the exposure level of operators of MRI devices ranges from 5 to 100 mT (Mevissen et al., ’94). MRI examinations are avoided in patients with metallic implants, such as some intrauterine contraceptive devices.

During the last decade, obstetrical use of MRI imaging has increased (Lowe et al., ’85). This technique has proved efficient in the evaluation of the position of the placenta (Powell et al., ’86) and in the assessment of fetal anatomy, e.g., ventriculomegaly (Hanigan et al., ’861, growth retardation (Stark et al., ’851, or conjoined twins (Turner et al., ’86). First trimester exposures were, however, performed on women coming for termi- nation of pregnancy. Second and third trimester MRI examinations have been reported by Baker et al. (‘94) not to increase the incidence of adverse pregnancy outcome or of abnormal hearing tests in a small number of children. Evans et al. (’93) designed a questionnaire study of MRI workers that did show that employment exposure to these magnetic fields had adverse effects on fertility or infant birth weight. The relative risk for miscarriage in comparison with women pregnant a t other jobs or at home was 1.27 (95% CI: 0.92-1.77; P = 0.07). The completeness and validity of the self- reported data in studies like this are open to question.

CONCLUSIONS The public concern about EMFs is motivated mainly

by the fact that they are ubiquitous; many pregnant women cannot avoid totally this type of exposure. In agreement with Sir R. Doll (’96), we believe the totality of the evidence available at this time provides no convincing evidence that low frequency EMFs of the sort that might be met in occupational or daily life exposures have a harmful effect on the human reproductive process. There are reasonable arguments in favor of EMFs being biologically active, but no mechanism has been identified that explains the link between EMFs and bioeffects. Moreover, frequencies related to electrically heated beds, occupational exposures in power companies, VDTs, and MRI exposures have not been shown to be associated with any significant teratogenic activity. It is usual in such situations to state that further research is needed to assess the reality of a risk, but since there are no plausible mechanisms for this alleged teratogenicity, further research would not be productive at this time.

It is suggested that those counseling pregnant women follow the guidelines established by the World Health Organization (WHO) in agreement with the Interna- tional Radiation Protection Association (IRPA) (Table 3). This group does not consider that the results of

TABLE 3. Permissible levels at 50 Hz: IRPA/WHO recommendations (May 1993)

Electric Magnetic Type of exposure fields (kV/m) fields (pT) Occupational

Per working day 10 500 Short duration 30 5,000

Home Up to 24 h 5 100 A few hours per day 10 1.000

published studies provide a basis for restricting unduly human exposure to EMFs and radiation.

ACKNOWLEDGMENTS We thank Dr. J. Lambrozo, Dr. A.R. Scialli, and the

reviewers for their helpful suggestions.

LITERATURE CITED Abenhaim, L., and F. Lert (1991) Methodological issues for the

assessment of clusters of adverse pregnancy outcomes in the workplace: The case ofvideo display terminal users. J. Occup. Med., 33:1091-1096.

Advisory Group on Non-Ionizing Radiations. Health effects related to the use of visual display units. British National Radiation Protec- tion Board, 1994, 5, No 2.

Baker,P.N., I.R. Johnson, P.R. Harvey,P.A. Gowland, andP. Mansfield (1994) A three-year follow-up of children imaged in utero with echoplanar magnetic resonance. Am. J. Obstet. Gynecol., 170:32-33.

Bassett, C.A.L. (1993) Medical benefits of electric and magnetic fields: Challenges for broader use. In: Electricity and Magnetism in Biology and Medicine. M. Blank, ed. San Francisco Press, San Francisco, pp. 9-11.

Belton, P.S., R.E. Gordon, J.M. Jones, and D. Shaw (1984)A31P topical magnetic resonance study of embryonic development in hens’ eggs. Br. Poult. Sci., 24:429-433.

Bergqvist, U.O. (1984) A model study on the occurrence of adverse pregnancy outcomes in clusters of workers using video display terminals. Scand. J. Work Environ. Health, 10:80-82.

Blackman, C.F., S.G. Benane, D.J. Elliott, and D.E. House (1988) Influence of electromagnetic fields on the efflux of calcium ions from brain tissue in vitro: A three model analysis consistent with the frequency response up to 510 Hz. Bioelectromagnetics, 9:215-227.

Blackwell, R., and A. Chang (1988) Video display terminals and pregnancy. Areview. Br. J. Obstet. Gynaecol., 95:44€453.

Bracken, M.B., K. Belanger, K. Hellenbrand, L. Dlugosz, T.R. Holford, J.E. McSharry, K. Adesso, and B. Leaderer (1995) Exposure to electromagnetic fields during pregnancy with emphasis on electrically heated beds: Association with birthweight and intrauterine growth retardation. Epidemiology, 6:263-270.

Brandt, L.P.A., and C.Y. Nielsen (1990) Congenital malformations among children of women working with video display terminals. Scand. J. Work Environ. Health, 16:229-233.

Brent, R.L., W.E. Gordon, W.R. Bennett, and D.A. Beckman (1993) Reproductive and teratogenic effects of electromagnetic fields. Re- prod. Toxicol., 7:535-580.

Breysse, P., P.S. Lees, M.A. McDiarmid, and B. Curbow (1994) ELF magnetic field exposures in an office environment. Am. J. Ind. Med.

Cerretelli, P., V. Veicsteinas, A. Margonato, D. Cantone, C. Viola, C. Malaguti, and A. Previ (1979) 1,000 kV project: Research on the biological effects of 50-Hz electric fields in Italy. In: Biological Effects of Extremely Low Frequency Electromagnetic Fields. Proceedings of the 18th Annual Hanford Life Sciences Symposium Series. Confer- ence No. 781016. R.D. Phillips, M.F. Gillis, W.T. Kaune, and D.D. Mahlum, eds. NTIS, Springfield, VA, p. 241.

25:177-185.

312 E.ROBERT

Chernoff, N., J.M. Rogers, and R. Kavet (1992) A review of the literature on potential reproductive and developmental toxicity of electric and magnetic fields. Toxicology (Amsterdam), 74:91-126.

Coleman, M., and V. Beral(1988)Areview ofepidemiological studies of the health effects of living near or working with electricity generation and transmission equipment. Int. J. Epidemiol., 17:l-13.

Cooke, P., and P.G. Morris (1981) The effects of NMR exposure on living organisms. 11. A genetic study of human lymphocytes. Br. J . Radiol., 154:622.

Council on Scientific Affairs (1987) Health effects of video display terminals. JAMA, 257: 1508-1512.

Davies, W.L., L.A. Crawford, O.J. Cooper, G.R. Farmer, D. Thomas, and B.L. Freeman (1990) Ethanol induces the generation of reactive free radicals by neural crest cells in vitro. J. Craniofac. Gen. Dev. Biol., 10.277-293.

Delgado, J.M.R., J . Leal, J.L. Monteoagudo, and M.G. Garcia (1982) Embryological changes by weak, extremely low frequency electromagnetic field. J . Anat., 134:533-551.

DeSesso, J.M., A.R. Scialli, and G.C. Goeringer (1994) D-mannitol, a specific hydroxyl free radical scavenger, reduces the developmental toxicity of hydroxyurea in rabbits. Teratology, 49:248-259.

Desjobert, H,, J, ~ i l l i ~ ~ , M, Adolphe, G, Averlant, and J. Nafziger (1995) Effects of 50 Hz magnetic fields on c-myc transcript levels in nonsynchronized and synchronized human cells. Bioelectromagnet- ics, 16:277-283.

Dlugosz, L., J. Vena, T. Byers, L. Sever, M. Bracken, and E. Marshall (1992) Congenital defects and electric bed heating in N~~ york State: A register-based case-control study, Am, J, Epidemiol,, 135: 1000-10 11.

Doll, R. (1996) Proceedings of the 1994 Conference of the French Society for Radiation Protection (in press).

Kaune, W.T., and R.D. Phillips (1980) Comparison of the coupling of grounded humans, swine and rats to vertical, 60 Hz electric fields. Bioelectromagnetics, 1: 117-119.

Kultursay, N., F. Koprubasi, and N. Kutukculer (1994) Video display terminal-The risk of trisomy 18 (letter). Clin. Genet., 45:270-271.

Kurppa, K., P.C. Holmberg, K. Rantala, T. Turminen, and L. Saxen (1985) Birth defects and exposure to video display terminals during pregnancy. Scand. J . WorkEnviron. Health, 11:353-356.

Lacy Hulbert,A., R.C. Wilkins, T.R. Hesketh, and J.C. Metcalfe (1995) No effect of 50 Hz electromagnetic fields on myc or beta-actin expression in human leukemic cells. Radiat. Res., 144:9-17.

Li, D.-K., H. Checkoway, and B.A. Mueller (1995) Electric blanket use in relation to the risk of congenital urinary tract anomalies among women With a history of SubfertilitY. EPidemioloD, 6:486439.

Lichty, P.D. (1985) Health Hazard Evaluation Report No. HETA-84-297- 1609. NIOSH, Department of Health and Human Services.

Liden, c., V. Bergqvist, and G. Wennersten (1986) Visual display terminals do not emit W - A radiation of clinical relevance. J. Theor. Biol., 122:491-492.

p. Nan- delstadh, H. Taskinen, M. Pekkarinen, M. Ylikoski, and K. Hem- mink (1992) Magnetic fields of video display terminals and spontaneous abortion.Am. J. Epidemiol.t 136:1041-1051.

T.w~* J. Weinreb, R. Santos-Ramos, and F'G' Cunningham (1985) Magnetic resonance imaging in human pregnancy. Obstet. Gyneco1'9 66:629-633'

ally related magnetic field exposure and male subfefiility. Fertil. Steril., 63:384-391.

Marha, K., and D. Charron (1985) The distribution of a pulsed very low frequency electric field around video display

Lindbohmy M.L.t M. Hietanen, p. Kmanen, M.

Lundsberg, L.S.3 M.B. Bracken, and K. Belanger (1995) OccuPation-

Health Phys .49:5 17-52 1. Ericson, A,, and B. K i m n (1986~3) An epidemiological study of work ~ ~ r r i ~ t t , I.A., and M,A, stuchly (1986) Health aspects of work with with video Screen and pregnancy outcomes: A registry study. Am. J.

Ericson, A., and B. Kiill6n (1986b) An epidemiological study of work with video screen and pregnancy outcomes. 11. A case-control study.

video display terminals, J, occup. Med,, 28,.833-848, Martin, F.B., A. Bender, G. Steuernagel, R.A. Robinson, R. Revsbech,

D.K, Sorensen, N. Williamson, and A. Williams (1986) Epidemio- logic study ofHolstein dairy cow and reproduction near a high-voltage direct-current powerline. J. 'lbxicol. Environ. Health, 19;303-324,

(1990) Early pregnancy loss and exposure to 50-Hz magnetic fields. ~~~~i~ (Finland): proceedings of the URSI/IEEE XVI National Convention on Radio science.

McDiarmid, M.A., P. Breysse, P.S. Lees, B. Curbow, and K. Kolodner (1994) Investigation of a spontaneous abortion cluster: Lessons learned, A,,,, J. ~ ~ d , ~ ~ d . , ~5~463-475,

McDonald, A.D., N.M. Cherry, C. Delorme, and J.C. McDonald (1986) Visual display units and pregnancy: Evidence from the Montreal sunrey. J, owup. Me& 28;122&1231,

McRobbie, D., and M.A. Foster (1985) Pulsed magnetic field exposure during pregnancy and implications for NMR foetal imaging: A study with mice, M ~ ~ , R ~ ~ ~ ~ , ~ ~ ~ g i ~ ~ , 3:231-234, 1985,

Mevissen, M., S. Buntenkotter, and W. Loscher (1994) Effects of static and time-varying (50 Hz) magnetic fields on reproduction and fetal development in rats, Rratology, 5():229-237,

Murray, W.E. (1986) Video display terminals: Radiation issues. Li- b r a v Hi Tech., 12:43-47.

Nafziger, J., H. Desjobert, B. Benamar, J .J . Guillosson, and M. Adolphe, (1993) DNA mutations and 50 H~ electromagnetic fields, Bioelectrochem. &energ,, 30;133-141,

Nordstrom, S., E. Birke, and L. Gustavsson (1983) Reproductive hazards among workers at high voltage substations. Bioelectromag- netics, 4:91-101.

Parazzini, F., L. Luchini, C. La Vecchia, and P.G. Crosignani (1993) Video display terminal use during pregnancy and reproductive outcome-A meta-analysis. J . Epidemiol. Community Health, 47: 265-268.

Picazo, M.L., V.A. Leyton, M.P. de Miguel, R. Paniagua, and J.L. Bardasan0 (1995) Effects ofchronic exposure to ELF magnetic fields on the reproductive system of fetal mice. Bioelectrochem. Bioenerg.

Ind. Med., 9:447-457.

Am. J. Ind. Med., 9:459-475. Evans, J.A., D.A. Savitz, E. Kanal, and J. Gillen (1993) Infertility and

pregnancy Outcome among magnetic imaging workers. J. Matilainen, p,, J. Juutilainen, S, Saarikoski, E, Laara, and S , Suonio Occup. Med., 35:1191-1195.

Foster, M.A., C.H. Knight, J.E. Rimmington, and J.R. Mallard (1983) Fetal imaging by nuclear magnetic resonance: A study in goats. Radiology, 149:193-195.

risk Of

miscarriage and birth defects among women who use visual display terminals during pregnancy. Am. J. Ind. Med., 13:695-706.

Goodman, R., and AS. Henderson (1988) Exposure of salivary glands to low frequency electromagnetic fields alters polypeptide synthesis. Proc. Natl. Acad. Sci. U.S.A., 853928-3932.

Goodman, R., and AS . Henderson (1991) Transcription in cells exposed to extremely low frequency electromagnetic fields: Areview. Bioelectrochem. Bioenerg., 25:335-355.

Hanigan, W.C., J . Gibson, N.J. Kleopoulos, T. Cusack, G. Zwicky, and R.M. Wright (1986) Medical imaging of fetal ventnculomegaly. J. Neurosurg., 64:575-580.

Infante-Rivard, C. (1995) Electromagnetic field exposure during pregnancy and childhood leukemia (letter). Lancet, 346:177.

International Non-Ionizing Radiation Committee of the International Radiation Protection Association (1991) Protection of the patient undergoing a magnetic resonance examination. Health Phys., 61: 923-928.

Juutilainen, J. (1991) Effects of low-frequency magnetic fields on embryonic development and pregnancy. Scand. J. Work Environ. Health, 17:149-158.

Juutilainen, J., and K. Saali (1986) Measurement of extremely low-frequency magnetic fields around video display terminals. Scand. J. Work Environ. Health, 12:609-613.

Juutilainen, J., P. Matilainen, S . Saarikoski, E. Laara, and S. Suonio (1993) Early pregnancy loss and exposure to 50-Hz magnetic fields.

M.K., M.R. and R.A. Hiatt (1988)

Bioelectromagnetics, 14:229-236. 38:173-177.


Powell, M.C., J. Buckley, H. Price, B.S. Worthington, and E.M. Symonds (1986) Magnetic resonance imaging and placenta previa. Am. J . Obstet. Gynecol., 154~565-569.

Reiter, R.J. (1993) Electromagnetic fields and melatonin production. Biomed. Pharmacother., 47~439-444.

Reiter, R.J., L.E. Anderson, R.L. Buschbom, and B.W. Wilson (1988) Reduction of the nocturnal rise in pineal melatonin levels in rats exposed to 60-Hz electric fields in utero and for 23 days aRer birth. Life Sci., 422203-2206.

Robert, E. (1993) Birth defects and high voltage power lines: An exploratory study based on registry data. Reprod. Toxicol., 7:283- 287.

Robert, E., J.A. Harris, 0. Robert, and S. Selvin (1996) Case-control study on maternal residential proximity to high voltage power lines and congenital malformations in France. Paediatr. Perinat. Epide- miol., 10:32-38.

Robinson, H. (1989) The risk of miscarriage and birth defects among women who use visual display terminals during pregnancy (letter). Am. J. Ind. Med., 15:357-358.

Rodriguez-Pinilla, E., and M.L. Martinez-Frias (1995) Video display terminals: Risk of trisomy 18? Clin. Genet., 47:335-336.

Roman, E., V. Beral, M. Pelerin, and C. Hermon (1992) Spontaneous abortion and work with visual display units. Br. J . Ind. Med., 49:507-512.

Rommereim, D.N., R.L. Rommereim, M.R. Sikov, R.L. Buschbom, and L.E. Anderson (1990) Reproduction, growth, and development of rats during chronic exposure to multiple field strengths of 6 0 - H ~ electric fields. Fundam. Appl. Toxicol., 14:608-621.

Ryan, B.M., E. Mallett, Jr., T.L. Bryan, M.A. Cahill, J.C. Findlay, L.C. Moore, L.E. Pomeranz, R.R. Symanski, J.D. Ter Molen, and D.L. McCormick (1995) Developmental toxicity (teratology) study of 60 Hz (power frequency) magnetic fields in Sprague-Dawley rats (abstract). Teratology, 51:196.

Savitz, D.A. and C.V. Ananth (1994) Residential magnetic fields, wire codes, and pregnancy outcome. Bioelectromagnetics, 15~271-273.

Schnorr, T.M., B. Grajewski, R. Hornung, M.J. Thun, G. Egeland, W. Murray, D. Conover, and W. Halperin (1991) Video display terminals and the risk of spontaneous abortion. N. Engl. J . Med., 324:727-733.

Stark, D.D., S.M., McCarthy, R.A. Filly, P.W. Callen, H. Hricak, and J.T. Parer (1985) Intrauterine growth retardation: Evaluation by magnetic resonance. Work in progress. Radiology, 155:425-427.

Stevens, R.G. (1993) Biologically based epidemiological studies of electric power and cancer. Environ. Health Perspect., 101:93-100.

Stuchly, M.A., D. Ruddick, D. Villeneuve, K. Robinson, B. Reed, D.W. Lecuyer, K. Tan, and J. Wong (1988) Teratological assessment of exposure to time-varying magnetic field. Teratology, 38:461-466.

Thomas, A., and P.G. Morris (1981) The effects of NMR exposure on living organisms. I. A microbial assay. Br. J . Radiol., 154:615.

Tikkanen, J., O.P. Heinonen, K. Kurppa, and K. Rantala (1990) Cardiovascular malformations and maternal exposure to video display terminals during pregnancy. Eur. J. Epidemiol., 6:61-66.

Turner, R.J., G.D. Hankins, J.C. Weinreb, P.R. Ziaya, T.N. Davies, T.W. Lowe, and L.C. Gilstrap (1986) Magnetic resonance imaging and ultrasonography in the antenatal evaluation of conjoined twins. Am. J . Obstet. Gynecol., 155645-649.

Tyndall, D.A., and K.K. Sulik (1991) Effect of magnetic resonance imaging on eye development in the C57BU6J mouse. Teratology, 43:263-275.

Wertheimer, N., and E. Leeper (1986) Possible effects of electric blankets and heated waterbeds on fetal development. Bioelectromag- netics, 7~13-22.

Wertheimer, N., and E. Leeper (1989) Fetal loss associated with two seasonal sources of electromagnetic field exposure. Am. J. Epide- miol., 129:220-224.

Wiley, M.D., D. Agnew, S. Harvey, M. Walsch, and P. Corey (1990) The effects of maternal exposure to VDT-like magnetic fields throughout pregnancy on the prenatal development of CD-1 mice. Teratology,

Wiley, M.J., P. Corey, K.J. Kavet, S. Harvey, D. Agnew, and M. Walsh (1992) The effects of continuous exposure to 20-lrHz sawtooth magnetic fields on the litters of CD-1 mice. Teratology, 46:391-398.

Wilson, B.W., E.K. Chess, and L.E. Anderson (1986) 60 Hz electric field effects on pineal melatonin rhythms: Time course for onset and recovery. Bioelectromagnetics, 7~239-249.

41:599-600.

teratogen update: electromagnetic fields

Documents