The purpose of the experimental studies described here, using animal models is, in part, to answer the question as to whether extremely low frequency (ELF) magnetic field exposures at levels similar to those around VDU workstations can be shown to affect reproductive functions in animals in a manner that can be equated to a human health risk.
The studies considered here are limited to in vivo studies (those performed on live animals) of reproduction in mammals exposed to very low frequency (VLF) magnetic fields with appropriate frequencies, excluding, therefore, studies on the biological effects in general of VLF or ELF magnetic fields. These studies on experimental animals fail to demonstrate unequivocally that magnetic fields, such as are found around VDUs, affect reproduction. Moreover, as can be seen from considering the experimental studies described in some detail below, the animal data do not shed a clear light on possible mechanisms for human reproductive effects of VDU use. These data complement the relative absence of indications of a measurable effect of VDU use on reproductive outcomes from human population studies.
Studies of Reproductive Effects of VLF Magnetic Fields in Rodents
VLF magnetic fields similar to those around VDUs have been used in five teratological studies, three with mice and two with rats. The results of these studies are summarized in table 1. Only one study (Tribukait and Cekan 1987), found an increased number of foetuses with external malformations. Stuchly et al. (1988) and Huuskonen, Juutilainen and Komulainen (1993) both reported a significant increase in the number of foetuses with skeletal abnormalities, but only when the analysis was based on the foetus as a unit. The study by Wiley and Corey (1992) did not demonstrate any effect of magnetic field exposures on placental resorption, or other pregnancy outcomes. Placental resorptions roughly correspond to spontaneous abortions in humans. Finally, Frölén and Svedenstål (1993) performed a series of five experiments. In each experiment, the exposure occurred on a different day. Among the first four experimental subgroups (start day 1–start day 5), there were significant increases in the number of placental resorptions among exposed females. No such effects were seen in the experiment where exposure started on day 7 and which is illustrated in figure 1.
Table 1. Teratological studies with rats or mice exposed to 18-20 kHz saw-tooth formed magnetic fields
|
Magnetic field exposure |
|||||
|
Study |
Subject1 |
Frequency |
Amplitude2 |
Duration3 |
Results4 |
|
Tribukait and Cekan (1987) |
76 litters of mice |
20 kHz |
1 μT, 15 μT |
Exposed to day 14 of pregnancy |
Significant increase in external malformation; only if foetus is used as the unit of observation; and only in the first half of the experiment; no difference as to resorption or foetal death. |
|
Stuchly et al. |
20 litters of rats |
18 kHz |
5.7μT, 23μT, |
Exposed throughout |
Significant increase in minor skeletal malformations; only if foetus is used as the unit of observation; some decrease in blood cell concentrations no difference as to resorption, nor as to other types of malformations |
|
Wiley and Corey |
144 litters of |
20 kHz |
3.6 μT, 17μT, |
Exposed throughout |
No difference as to any observed outcome (malformation, |
|
Frölén and |
In total 707 |
20 kHz |
15 μT |
Beginning on various days of pregnancy in |
Significant increase in resorption; only if exposure starts on day 1 to day 5; no difference as to malformations |
|
Huuskonen, |
72 litters of rats |
20 kHz |
15 μT |
Exposed to day 12 of pregnancy |
Significant increase in minor skeletal malformations; only if foetus is used as the unit of observation; no difference as to |
1 Total number of litters in the maximum exposure category.
2 Peak-to-peak amplitude.
3 Exposure varied from 7 to 24 hours/day in different experiments.
4 “Difference” refers to statistical comparisons between exposed and unexposed animals, “increase” refers to a comparison of the highest exposed group vs. the unexposed group.
Figure 1. The percentage of female mice with placental resorptions in relation to exposure
The interpretations given by the researchers to their findings include the following. Stuchly and co-workers reported that the abnormalities they observed were not unusual and ascribed the result to “common noise that appears in every teratological evaluation”. Huuskonen et al., whose findings were similar to Stuchly et al., were less negative in their appraisal and considered their result to be more indicative of a real effect, but they too remarked in their report that the abnormalities were “subtle and would probably not impair the later development of the foetuses”. In discussing their findings in which effects were observed in the early onset exposures but not the later ones, Frölén and Svedenstål suggest that the effects observed could be related to early effects on reproduction, before the fertilized egg is implanted in the uterus.
In addition to the reproductive outcomes, a decrease in white and red blood cells were noted in the highest exposure group in the study by Stuchly and co-workers. (Blood cell counts were not analysed in the other studies.) The authors, while suggesting that this could indicate a mild effect of the fields, also noted that the variations in blood cell counts were “within the normal range”. The absence of histological data and the absence of any effects on bone marrow cells made it difficult to evaluate these latter findings.
Interpretation and comparison of studies
Few of the results described here are consistent with one another. As stated by Frölén and Svedenstål, “qualitative conclusions with regard to corresponding effects in human beings and test animals may not be drawn”. Let us examine some of the reasoning that could lead to such a conclusion.
The Tribukait findings are generally not considered to be conclusive for two reasons. First, the experiment only yielded positive effects when the foetus was used as the unit of observation for statistical analysis, whereas the data themselves actually indicated a litter-specific effect. Second, there is a discrepancy in the study between the findings in the first and the second part, which implies that the positive findings may be the result of random variations and/or uncontrolled factors in the experiment.
Epidemiological studies investigating specific malformations have not observed an increase in skeletal malformations among children born of mothers working with VDUs—and thus exposed to VLF magnetic fields. For these reasons (foetus-based statistical analysis, abnormalities probably not health-related, and lack of concordance with epidemiological findings), the results—on minor skeletal malformations—are not such as to provide a firm indication of a health risk for humans.
Technical Background
Units of observation
When statistically evaluating studies on mammals, consideration must be given to at least one aspect of the (often unknown) mechanism. If the exposure affects the mother—which in turn affects the foetuses in the litter, it is the status of the litter as a whole which should be used as the unit of observation (the effect which is being observed and measured), since the individual outcomes among litter-mates are not independent. If, on the other hand, it is hypothesized that the exposure acts directly and independently on the individual foetuses within the litter, then one can appropriately use the foetus as a unit for statistical evaluation. The usual practice is to count the litter as the unit of observation, unless evidence is available that the effect of the exposure on one foetus is independent of the effect on the other foetuses in the litter.
Wiley and Corey (1992) did not observe a placental resorption effect similar to that seen by Frölén and Svedenstål. One reason put forward for this discrepancy is that different strains of mice were used, and the effect could be specific for the strain used by Frölén and Svedenstål. Apart from such a speculated species effect, it is also noteworthy that both females exposed to 17 μT fields and controls in the Wiley study had resorption frequencies similar to those in exposed females in the corresponding Frölén series, whereas most non-exposed groups in the Frölén study had much lower frequencies (see figure 1). One hypothetical explanation could be that a higher stress level among the mice in the Wiley study resulted from the handling of animals during the three hour period without exposure. If this is the case, an effect of the magnetic field could perhaps have been “drowned” by a stress effect. While it is difficult to definitely dismiss such a theory from the data provided, it does appear somewhat far-fetched. Furthermore, a “real” effect attributable to the magnetic field would be expected to be observable above such a constant stress effect as the magnetic field exposure increased. No such trend was observed in the Wiley study data.
The Wiley study reports on environmental monitoring and on rotation of cages to eliminate the effects of uncontrolled factors which might vary within the room environment itself, as magnetic fields can, while the Frölén study does not. Thus, control of “other factors” is at least better documented in the Wiley study. Hypothetically, uncontrolled factors that were not randomized could conceivably offer some explanations. It is also interesting to note that the lack of effect observed in the day 7 series of the Frölén study appears to be due not to a decrease in the exposed groups, but to an increase in the control group. Thus variations in the control group are probably important to consider while comparing the disparate results of the two studies.
Studies of Reproductive Effects of ELF Magnetic Fields in Rodents
Several studies have been performed, mostly on rodents, with 50–80 Hz fields. Details on six of these studies are shown in table 2. While other studies of ELF have been carried out, their results have not appeared in the published scientific literature and are generally available only as abstracts from conferences. In general the findings are of “random effects”, “no differences observed” and so on. One study, however, found a reduced number of external abnormalities in CD–1 mice exposed to a 20 mT, 50 Hz field but the authors suggested that this might reflect a selection problem. A few studies have been reported on species other than rodents (rhesus monkeys and cows), again apparently without observations of adverse exposure effects.
Table 2. Teratological studies with rats or mice exposed to 15-60 Hz sinusoidal or square pulsed magnetic fields
|
Magnetic field exposure |
||||||
|
Study |
Subject1 |
Frequency |
Amplitude |
Description |
Exposure duration |
Results |
|
Rivas and Rius |
25 Swiss mice |
50 Hz |
83 μT, 2.3 mT |
Pulsed, 5 ms pulse duration |
Before and during pregnancy and offspring growth; total 120 days |
No significant differences at birth in any measured parameter; decreased male body weight when adult |
|
Zecca et al. (1985) |
10 SD rats |
50 Hz |
5.8 mT |
Day 6-15 of pregnancy, |
No significant differences |
|
|
Tribukait and Cekan (1987) |
35 C3H mice |
50 Hz |
1 μT, 15 μT |
Square wave-forms, 0.5 ms duration |
Day 0-14 of pregnancy, |
No significant differences |
|
Salzinger and |
41 off-springs of SD rats. Only male pups used |
60 Hz |
100 μT (rms). Also electric |
Uniform circular polarized |
Day 0-22 of pregnancy and |
Lower increase in operand response during training commencing at 90 days of age |
|
McGivern and |
11 offsprings of SD rats. Only male pups used. |
15 Hz |
800 μT (peak) |
Square wave-forms, 0.3 ms duration |
Day 15-20 of pregnancy, |
Territorial scent marking behaviour reduced at 120 days of age. |
|
Huuskonen et al. |
72 Wistar rats |
50 Hz |
12.6μT (rms) |
Sinusoidal |
Day 0-12 of pregnancy, |
More foetuses/litter. Minor skeletal malformations |
1 Number of animals (mothers) in the highest exposure category given unless otherwise noted.
As can be seen from table 2, a wide range of results were obtained. These studies are more difficult to summarize because there are so many variations in exposure regimens, the endpoints under study as well as other factors. The foetus (or the surviving, “culled” pup) was the unit used in most studies. Overall, it is clear that these studies do not show any gross teratogenic effect of magnetic field exposure during pregnancy. As remarked above, “minor skeletal anomalies” do not appear to be of importance when evaluating human risks. The behavioural study results of Salzinger and Freimark (1990) and McGivern and Sokol (1990) are intriguing, but they do not form a basis for indications of human health risks at a VDU workstation, either from the standpoint of procedures (use of the foetus, and, for McGivern, a different frequency) or of effects.
Summary of specific studies
Behavioural retardation 3–4 months after birth was observed in the offspring of exposed females by Salzinger and McGivern. These studies appear to have used individual offspring as the statistical unit, which may be questionable if the stipulated effect is due to an effect on the mother. The Salzinger study also exposed the pups during the first 8 days after birth, so that this study involved more than reproductive hazards. A limited number of litters was used in both studies. Furthermore, these studies cannot be considered to confirm each other’s findings since the exposures varied greatly between them, as can be seen in table 2.
Apart from a behavioural change in the exposed animals, the McGivern study noted an increased weight of some male sex organs: the prostate, the seminal vesicles and the epididymis (all parts of the male reproductive system). The authors speculate as to whether this could be linked to stimulation of some enzyme levels in the prostate since magnetic field effects on some enzymes present in the prostate have been observed for 60 Hz.
Huuskonen and co-workers (1993) noted an increase in the number of foetuses per litter (10.4 foetuses/litter in the 50 Hz exposed group vs. 9 foetuses/litter in the control group). The authors, who had not observed similar trends in other studies, downplayed the importance of this finding by noting that it “may be incidental rather than an actual effect of the magnetic field”. In 1985 Rivas and Rius reported a different finding with a slightly lower number of live births per litter among exposed versus nonexposed groups. The difference was not statistically significant. They carried out the other aspects of their analyses on both a “per foetus” and “per litter” basis. The noted increase in minor skeletal malformations was only seen with the analysis using the foetus as the unit of observation.
Recommendations and Summary
Despite the relative lack of positive, consistent data demonstrating either human or animal reproductive effects, attempts at replications of the results of some studies are still warranted. These studies should attempt to reduce the variations in exposures, methods of analysis and strains of animals used.
In general, the experimental studies performed with 20 kHz magnetic fields have provided somewhat varied results. If adhering strictly to the litter analysis procedure and statistical hypothesis testing, no effects have been shown in rats (although similar nonsignificant findings were made in both studies). In mice, the results have been varied, and no single coherent interpretation of them appears possible at present. For 50 Hz magnetic fields, the situation is somewhat different. Epidemiological studies which are relevant to this frequency are scarce, and one study did indicate a possible risk of miscarriage. By contrast, the experimental animal studies have not produced results with similar outcomes. Overall, the results do not establish an effect of extremely low frequency magnetic fields from VDUs on the outcome of pregnancies. The totality of results fails thus to suggest an effect of VLF or ELF magnetic fields from VDUs on reproduction.