Studies of a possible association between exposure to EMF and cancer are challenged by unidentified intensity and/or frequency parameters that can result in reproducible biological responses and dosimetric differences between animals and humans. The effects of EMF have usually been studied at intensities of exposure that are much higher than environmental levels, in order to determine if effects occur. Another approach, not yet used to any great extent, is to focus on exposure parameters at levels commonly experienced by humans. Several designs and animal models have been used in laboratory studies of cancer. The choice of model depends largely on the hypothesis for a particular underlying mechanism. Few carcinogenic agents exert their full effect after a single, brief exposure, and most agents act only after an extended exposure. During that time, exposure to other possibly confounding agents must be kept to a minimum. The long-term bioassay is designed to address this issue. In this design, the animals are observed for most of their lifetime, and the number, type, and time of appearance of tumors are the critical end-points. This type of study should include several doses and a relatively large number of animals, particularly if the natural incidence of the tumor type is low. As might be surmised, studies of complete carcinogenicity are expensive, due to both the length of time and the number of animals involved. One potential problem in such studies is the inadequacy of present knowledge about what aspects of the EMF signal are biologically active.
A consideration in using animal models to investigate the effects of EMF on cancer development is the appropriateness of the model itself in relation to human disease. For example, the rat mammary carcinoma model has been thought to be reasonably relevant in many ways for investigating human breast cancer (Russo et al., 1990), as many of the factors that promote tumors in the animal model also increase breast cancer risk in humans. Nevertheless, other aspects of the rodent model may not be directly relevant to human disease. Another commonly used animal cancer model is mouse skin. This represents a convenient, well-developed way of investigating mechanistic questions in multistage carcinogenesis. Although it allows the study of potential risk factors on the process of cancer development, its results may not translate easily into information on the specific cancer type in humans. Other models, specifically for investigating the risk for leukemia, are valuable because a variety has been developed. Thus, considerable information is available about leukemia and lymphoma and its carcinogenic process in animals, and some similarities exist between the development of the disease in these models and in humans (Pattengale & Taylor, 1983).
The most comprehensive study to date of EMF as a potential carcinogen was conducted at the IIT Research Institute for the National Toxicology Program (NTP, 1998b). In this study, which was conducted according to good laboratory practice (GLP), groups of 100 Fischer 344 rats and 100 B6C3F1 mice of each sex were exposed to one of several magnetic field conditions: 2, 200, or 1000 µT continuously or 1000 µT intermittently (1 h on, 1 h off), 60 Hz linearly polarized magnetic fields; one group received sham exposure. Neither exposed nor control animals were exposed to transients. Exposure began when the animals were 6-7 weeks of age and continued for 18.5 h/d for two years. The animals were monitored and evaluated over the course of their lifetime for survival, body weight, and clinical signs of neoplasia. At death (average age, 112-113 weeks), all animals underwent complete necropsy and histopathological evaluation.
The survival of exposed rats (47-59% of males and 58-68% of females) was no different from that of control animals (57% males and 59% females). There were no exposure-related clinical findings. The only significant increase in tumor incidence in field-exposed rats was for thyroid gland C-cell adenomas and carcinomas combined in male rats, with incidences of 16% in sham-exposed controls, 31% in those at 2 µT (< 1.01), 30% in those at 200 µT (p < 0.01), 25% in those at 1000 µT continuously (p = 0.06), and 22% in those at 1000 µT intermittently (p = 0.15). The incidence of mononuclear-cell leukemia in males was 50, 44, 47, 50, and 36% (p < 0.05, intermittent group) for the five groups, respectively.
The survival of exposed mice (72-84% in males and 74-79% in females) was similar to that of control animals (76% males and 70% females), except for male mice exposed to 1000 µT which had significantly reduced survival (62/100, p = 0.037) in comparison with control mice (76/100). There were no exposure-related clinical findings. Significant differences in the incidence of neoplasms between field-exposed and sham-exposed mice included decreased incidences of alveolar/bronchiolar adenoma in males exposed to 2 µT (11/99, p = 0.007) or 200 µT (9/100, p < 0.001) and in female mice exposed to 200 µT (0/99, p = 0.002). The incidences of adenoma and carcinoma combined were significantly lower in males (19/100, p = 0.041) and females (29/99, p = 0.008) exposed to 200 µT than in controls (males, 30/100; females, 11/95). The incidence of malignant lymphoma in female mice exposed intermittently to 1000 µT was significantly lower (20/100, p = 0.035) than that observed in the controls (32/100).
The authors concluded that under the conditions of these two-year studies with whole-body exposure there was equivocal evidence for the carcinogenic activity of 60 Hz magnetic fields in male Fischer 344/N rats, on the basis of the increased incidences of thyroid gland C-cell neoplasms in male rats exposed at 2 or 200 T . There was no evidence of carcinogenic activity in female rats or in male or female B6C3F1 mice exposed to 2, 200, or 1000 µT continuously or to 1000 µT intermittently.
A study similar to that described above was conducted in Canada (Mandeville et al., 1997). In this study, groups of 50 female Fischer 344/N rats were exposed to 2, 20, 200, or 2000 µT 60 Hz, linearly-polarized magnetic fields; there were also groups of sham-exposed and cage controls. The authors carefully excluded transients. Exposure began two days before birth and continued for 20 h/d for two years. The animals were monitored and evaluated over the course of their lifetime for survival, body weight, and clinical signs of neoplasia. At death, all of the animals were subjected to complete necropsy and histopathological evaluation of 50 organs and tissues from each animal, with specific attention to the incidences of mononuclear-cell leukemia, brain tumors, and mammary tumors.
The survival of rats at the end of the study was significantly lower among the sham-exposed controls (19/50, p = 0.03) and animals exposed to low field intensities (16/50, p = 0.005 for both 2 and 20 µT groups) than in cage controls (30/50). The survival rates were not different for exposed and sham-exposed animals (16/50, 16/50, 24/50, and 25/50 in rats at 2, 20, 200, and 2000 µT, respectively, versus 19/50 in sham-exposed). There were no consistent exposure-related clinical findings.
The two control groups had higher overall tumor incidences (45/50 in cage controls and 46/50 in sham-exposed) than exposed rats (42/50, 43/50, 43/50, and 41/50 in rats at 2, 20, 200, and 2000 µT). In all groups of animals, the commonest tumor types were pituitary adenomas, mammary gland fibroadenomas, and mononuclear-cell leukemia. Following the pattern for overall tumor incidence, that of pituitary adenomas in the exposed animals varied between 40% (2 µT; 20/50, p = 0.04) and 44% (2000 µT; 22/50, p = 0.07) and was statistically significantly lower than in the controls (31/49 in cage controls, 29/49 in sham-exposed). When the sham-exposed animals were compared with the exposed rats for mammary gland fibroadenomas, there was no significant difference (28/50 for sham-exposed and mice at 2 µT versus 27/50, 24/49, and 21/50 for those at 20, 200, at 2000 µT respectively). Mammary gland adenocarcinomas were rare in all groups of animals. The incidence of mononuclear-cell leukemia was relatively low and not significantly different in any group.
The authors concluded that there was no evidence for the carcinogenic activity of 60 Hz, linear sinusoidal, continuous waves at intensities of 2, 20, 200, or 2000 µT in female Fischer 344 rats, suggesting lack of carcinogenicity. There were also no statistically significant, consistent, dose-related trend in the number of tumor-bearing animals per study group that could be attributed to exposure to magnetic fields.
The results of a carefully conducted, well-documented study to determine the carcinogenic effect of 50 Hz sinusoidal magnetic fields in rats throughout their lifetime have been reported (Yasui et al., 1997). Groups of 48 Fischer 344/DuCrj rats of each sex were exposed to flux densities of 500 or 5000 µT or sham fields from 5 to 109 weeks of age. The high dose was much higher than those used in the other two studies. The exposure design eliminated transients, and exposure was for 22.6 h/d day for two years. The animals were monitored and evaluated over their lifetime for survival, body weight, and clinical signs of neoplasia. At death, all animals underwent complete necropsy and histopathological evaluation.
The survival of exposed rats was no different from that of control animals. There were no exposure-related clinical findings. The only significant difference between exposed and sham-exposed rats in the incidence of neoplasms was an increase in fibroma of the subcutis in males at 5 mT (9/48 versus 2/48 in controls, p = 0.05). This increase in the incidence of a benign tumor was not significant in a comparison with historical controls. The incidences of leukemia, lymphoma, and brain and of adenomas of the pituitary and thyroid (changes noted in other animal studies) were no different between field- and sham-exposed animals.
Marganato et al. (Margonato et al., 1995) investigated the exposure of groups of 256 animals male Sprague-Dawley rats to a 50 Hz field over 32 weeks. Exposure was for 22 h/d to 5 µT magnetic fields or sham exposure. Although this study was not designed to address the issue of cancer per se, the biological end-points included hematological examinations for each animal and morphological and histological evaluation of the liver, heart, mesenteric lymph nodes, and testes. In two identical sets of experiments involving 128 rats per group per experiment, no significant difference in the investigated variables was found between exposed and sham-exposed animals. Although certain parameters differed between the two experiments, the exposure conditions were the same. The authors concluded that the results did not indicate any harmful effects of prolonged exposure to magnetic fields comparable to those measured close to power lines. [This conclusion is limited to a selected number of markers within the portion of the life span of the animal corresponding to a high growth rate but not to the early developmental phase.]
Few long-term bioassays have been performed of the exposure of mice or rats to 50- or 60 Hz magnetic fields. In the largest and most comprehensive of the studies (Mandeville et al., 1997; NTP, 1998b; Yasui et al., 1997), few significant effects (both increases and decreases) of exposure on cancer development were seen, with the exception of isolated thyroid C-cell adenomas and carcinomas in male rats in the NTP study.
Neither the long-term studies of spontaneous tumor development studies summarized above nor shorter studies have given any indication that EMF are mutagenic. For this reason, investigations based on the multistage nature of carcinogenesis have focused on the promotional phases of the cancer process. The studies summarized below addressed the specific question of whether EMF acts as a promoter or co-promoter of tumorigenesis in models of mammary, skin, liver, and brain cancer and of lymphoma/leukemia and as an enhancer in a rat leukemia progression model.
In the earliest reported study of this type, Beniashvili et al. (Beniashvili et al., 1991) induced mammary tumors in five groups of 50 female rats by an intravenous injection of N-methyl-N-nitrosourea (MNU; 50 mg/kg body weight) at 55 days of age. After administration of the MNU, one group served as cage controls, and two groups were exposed to a 20 µT, 50 Hz magnetic field for 30 or 180 min. each day, and exposure continued for the lifetime of the animals. Animals exposed to the 50 Hz field for 180 min/d had a higher incidence (p < 0.05) of mammary tumors than cage controls (43/46 vs. 27/46) and a larger number (p < 0.05) of total tumors (75 vs. 31 in controls). In addition, the mean latent period was shorter in the exposed animals (45.5 d) than in the MNU-injected controls (74.4 d). No differences were observed for the animals exposed to magnetic fields for 0.5 h/d.
As part of this study, 25 female rats were exposed to 50 Hz magnetic fields for 0.5 or 3 h/d with no MNU treatment. At the conclusion of the study after two years, limited information was collected at necropsy and histological examination. A significant increase (p < 0.05) in the incidence of mammary gland tumors was observed in rats exposed to 20 µT for 3 h/d (7/25 animals) in comparison with those exposed for 0.5 h/d (1/25) and with unexposed animals (0/50). The latency for tumor appearance was 74 ± 15 d for controls, 65 ± 1 d for 0.5-h exposure, and 46 ± 12 d for 3-h exposure.
Anisimov et al. (Anisimov et al., 1996), conducted a replication experiment in which groups of 40 outbred white rats were field- or sham-exposed to 50 Hz, 20 µT magnetic fields for life (generally less than five months). Rats received 50 mg/kg MNU intravenously at three-week intervals with either magnetic field exposure or sham exposure for 3 h/d. Mammary adenocarcinomas (identified histologically) were found in 7/22 sham-exposed and 15/33 field-exposed (not significant). The mean latent period for tumor development was 166 4 d for sham-exposed and 125 7 (p < 0.05) for field-exposed.
[The inadequate reporting of the method and experimental details prevents an assessment of the significance of either result.]
In the initial study (Mevissen et al., 1993), female rats were exposed after DMBA administration to 30 mT (50 Hz) magnetic fields. The average mammary tumor incidences were 66% (range, 55-75%) in reference controls and 61% (range, 50-78%) in sham-exposed controls. Exposure to magnetic fields resulted in tumors in 19/36 sham-exposed and 20/33 field-exposed animals, and the total number of tumors was 51 in treated animals and 36 in sham-exposed (not statistically significant). When the latter experiment was repeated with only eight or nine animals per group, no difference was found between the exposed and control groups. When groups of 36 DMBA-treated rats were exposed to a gradient (0.3-1.0 µT) 50 Hz field, no difference in tumor incidence was seen, and the exposed animals had 47 tumors versus 60 in the sham-exposed (not statistically significant) (Löscher et al., 1994; Mevissen et al., 1993). The authors concluded that these experiments indicate that magnetic fields at high flux densities act as a promoter or co-promoter of breast cancer. [This conclusion must be considered only tentative because of the limitations of this study, particularly the small sample size used for exposure to magnetic fields.]
In the next set of experiments, a much larger number of animals (99 rats/group) was used (Löscher et al., 1993). All rats received the four weekly doses of 5 mg DMBA beginning at 52 days of age. One group of 99 rats was then exposed to magnetic fields at a flux density of 100 µT, while another group of 99 rats served as sham-exposed controls. After three months of exposure, the number of tumor-bearing animals (macroscopically visible tumors at necropsy) was 34 in the sham-exposed group and 51 in the exposed animals (p < 0.05). The tumors were not significantly larger (p = 0.08) in the exposed group (median, 707 mm3; interquartile range, 168-1885) than in the controls (367 mm3; 101-1178; p = 0.07), and no difference was found in the number of tumors per tumor-bearing rat. The authors stated that the data demonstrate that long-term exposure of DMBA-treated female rats to an alternating magnetic field of low flux density promotes the growth and increases the incidence of mammary tumors. [The study is well documented and adequately addresses the problem of the earlier study. No histopathological data are presented.]
An extension of the previous study was conducted in which a complete histopathological examination was performed (Baum et al., 1995). Histological examination of the tumors from the previous experiment revealed more tumors than were detected by palpation, the incidence of histologically verified DMBA-induced lesions being 57/99 in sham-exposed controls and 65/99 in animals exposed to magnetic fields. When tumors and hyperplasia were combined [an unusual combination], the numbers were 74/99 in sham-exposed controls and 65/99 in field-exposed rats; the incidence did not significantly differ between the groups. The authors concluded that exposure did not alter the incidence of mammary lesions but accelerated tumor growth, consistent with a co-promoting effect of EMF. There was a significant increase (p < 0.05) in the number of adenocarcinomas in exposed animals (62/99 in field-exposed versus 49/99 in sham-exposed controls). [Adenocarcinoma are not normally separated from carcinoma in situ for evaluation in the mammary gland.] After assessing the histological data, the authors concluded that long-term exposure of DMBA-treated female rats to EMF promotes the size and progression of mammary tumors (treated: median, 733 mm3; interquartile range, 183-2994; controls: 367 mm3; 101-1178; p < 0.05), while tumor incidence is not affected.
In the next experiments in this series of studies (Mevissen et al., 1996a), an effort was made to determine if a dose-response relationship exists with field intensity. Ninety-nine animals per group were exposed to 10 µT for 13 weeks. DMBA induced palpable tumors in 55/99 of sham-exposed and 60/99 field-exposed animals. At autopsy, these numbers were 61 and 67%, respectively; neither was statistically significant. The size, number per animal, incidence, and latency of tumors were similar in the two groups. The authors concluded that at this field intensity, magnetic fields had no effect.
In the next experiment in this series (Mevissen et al., 1996b), groups of 99 rats were exposed to 50 µT. At autopsy, 55% of sham-exposed controls and 69% (p < 0.05) of those exposed to magnetic fields had tumors, but the size and the number of tumors per animal were similar in the two groups. The author concluded that at this field intensity, magnetic fields had no effect.
Mevissen et al. (Mevissen et al., 1998a) conducted a replication experiment in which 99 female Sprague-Dawley rats were field-exposed or sham-exposed to a 100 µT, 50 Hz magnetic field. All of the rats received a total of 20 mg of DMBA (4 x 5 mg fractionated dose) beginning at 52 ± 2 days of age. After three months of exposure, the rats were sacrificed and mammary tumors identified by macroscopic evaluation at necropsy. Small tumors were examined histologically to confirm diagnosis as adenocarcinomas. Macroscopically visible tumors were found in 82 field-exposed and 61 sham-exposed rats (p < 0.05). The total number of tumors was 230 in controls and 297 in exposed animals; no difference was observed in the number of tumors per animal or in tumor size.
[In this series of studies of DMBA-initiated breast cancer in Sprague-Dawley rats promoted with 10-100 T magnetic fields, a higher number of total tumors was found in field-exposed groups in most studies. These effects often did not reach statistical significance, and in none of the studies was there a difference between field-exposed and sham-exposed animals in the number of tumors per tumor-bearing rat. The exposure resulted in a decreased latency in several but not all studies, and this effect was often seen only during part of the study. Increased tumor size was seen with exposure to magnetic fields but not consistently across the studies. Although the rats were initiated with a very high dose of DMBA (20 mg/rat), the 34-61% tumor incidence in the controls is much lower than would have been expected in Sprague Dawley rats in US studies. This may reflect a differences in the rat strain.]
In an effort to replicate these results (NTP, 1998a), Battelle Pacific Northwest Laboratories attempted to simulate the design and experimental method of the studies as closely as possible, with four weekly doses of 5 mg DMBA beginning at 52 days of age given by intragastric intubation to female Sprague-Dawley rats. Exposure was for 18.5 h/d, 7 d per week for 13 weeks at field intensities of 100 and 500 µT (50 Hz) and 100 µT (60 Hz) linearly polarized horizontal sinusoidal magnetic fields. The study was conducted under GLP conditions and included documentation of fields and exposure characteristics. A 26-week study at a lower single dose of 10 mg DMBA was also conducted. As all rats given DMBA in the first 13-week study had mammary gland neoplasms at incidences, determined by palpation, > 80%, a second 13-week study was conducted with four weekly doses of 2 mg DMBA.
Additional groups were included which were exposed at 500 µT (50 Hz) and 100 µT (60 Hz) for 18.5 h/d; all environmental conditions (temperature, light, relative humidity, and noise) were regulated and monitored continuously. Full gross pathological analysis and complete histopathological examination of mammary tumors were conducted.
In the first 13-week study, no difference in the onset of tumors or tumor size was found by gross palpation. The skin and the mammary glands were transluminated to identify all potential lesions; all gross lesions were counted, measured in two directions and sampled for histology. The incidences of mammary gland carcinomas were 92% in sham-exposed rats, 86% in those at 100 T 50 Hz, 96% in those at 500 T 50 Hz, and 96% in those at 100 T 60 Hz magnetic fields. The numbers of mammary gland carcinomas verified histologically were 691 in sham-exposed and 528 (p < 0.05 negative trend), 651 (not significant), and 692 (not significant) for rats at 100 and 500 µT 50 Hz and 100 µT 60 Hz, respectively.
The high tumor incidence in all groups (86-94%) seen with the same dosing regime as used in the Lscher studies decreased the sensitivity of the first 13-week study to detect a promoting effect. In the second 13-week study, at 8 mg DMBA (2 mg x 4 weekly doses), the mean body weights and clinical findings (attributable to DMBA administration) were not different between field-exposed and sham-exposed groups. There was no difference in the time to onset of tumors or in tumor size by gross palpation. The incidences of mammary gland carcinomas were 43, 48, and 38% for the sham-exposure and for exposure to 100 or 500 T, 50 Hz magnetic fields, respectively, and the numbers of mammary gland carcinomas verified histologically were 102, 90, and 79, respectively. At necropsy, 99% of the palpated tumors were shown histologically to be mammary gland carcinomas. The tumor incidence, total number of tumors, number of tumors per tumor-bearing rat (average, 1.7-2), and tumor size were not increased by exposure.
In the 26-week study, survival was similar in the sham- and field-exposed groups. In addition, no consistent differences in clinical findings were seen between groups exposed to DMBA plus magnetic fields and DMBA controls. Mammary gland carcinomas and multiple carcinomas were observed in all groups, but the rats exposed to magnetic fields had consistently fewer mammary tumors than the DMBA controls (Table 4.3); the 100 µT, 60 Hz group had a significantly lower incidence than controls (p < 0.05). The numbers of tumors per tumor-bearing animal showed a similar pattern, the 100 µT, 60 Hz group having lower values than controls. Most of the palpable tumors in all groups were shown histologically to be mammary gland carcinomas and fibroadenomas. The numbers of mammary gland carcinomas verified histologically were 649 (sham-exposed), 494 (100 T, 50 Hz) (p < 0.05 negative trend for poly-3 test), 547 (500 T, 50 Hz), and 433 (100 T 60 Hz) (p < 0.05 negative trend for poly-3 test). The tumor sizes were similar in all groups.
| Hyperplasia | ||||
| Adenoma | ||||
| Carcinoma | ||||
| Fibroadenoma |
In both the 26-week and the 13-week studies, the authors found no evidence that magnetic fields promote the development of mammary gland neoplasms. [The rats exposed to EMF showed a decrease in the number of tumors and in the incidence of DMBA-initiated tumors. Because of the large number of tumors generated by the dose of DMBA used in the first 13-week study, the sensitivity of the assay to pick-up small promotor effects was limited.]
A further study was conducted to examine the effects of magnetic fields on mammary tumor development in rodents (Ekström et al., 1998). Although the intensities of the magnetic fields used in this study were similar to those in the studies described above, 250 and 500 µT at 50 Hz, there were also some significant differences. In this study, a transient-producing, intermittent field was used (15 s on, 15 s off). In addition, the fields were used in a strictly 'promotional' design, exposure beginning one week after DMBA administration rather than simultaneously with the DMBA. Fifty-two-day-old Sprague-Dawley female rats were treated with DMBA (7 mg/animal) and, once started, the exposures continued for 19-21 h/d for 25 weeks. As in the other studies, the end-points were tumor incidence, number of tumors per animal, and tumor volume and weight.
This incidences of tumors were 43/60, 42/60, and 42/60 for the sham-exposed, 250 µT and 500 µT 50 Hz groups, respectively. The numbers of tumors and tumors per animal were also similar between all the groups, with 111 tumors in the DMBA controls and 102 at 250 µT and 90 at 500 µT in the exposed groups. The rate of tumor appearance and tumor volume were the same in all groups. The authors concluded that magnetic fields had no effects in this study. [The interpretation of these results is made uncertain by the lack of histopathological data.]
Skin tumor promotion after initiation with DMBA was examined in groups of 32 female mice exposed to a 2000 µT, 60 Hz continuous magnetic field for 6 h/d five days per week for up to 21 weeks (McLean et al., 1991). Mice were initiated with DMBA (10 nmol in 200 l of acetone) on the dorsal skin and were then exposed to the magnetic field with or without TPA promotion (1 µg per week for 21 weeks). As none of the field-exposed or sham-exposed mice developed papillomas in the absence of TPA, the authors concluded that magnetic fields did not act as a tumor promoter. A slight, nonsignificant decrease in the time of appearance of tumors was observed in animals treated with TPA plus EMF, in comparison with animals treated with TPA and sham exposed. [The number of animals with tumors at 21 weeks was extremely high (> 90% in both exposed and control groups) and essentially precluded a judgment as to whether EMF could affect the incidence of skin papillomas.]
In a second study, two groups of 48 mice were similarly initiated with DMBA and promoted with 0.3 µg (4.9 nmol) TPA for 23 weeks, with or without magnetic field exposure (Stuchly et al., 1992). The onset of tumor development occurred earlier in mice (p < 0.05 for weeks 16, 17, and 18) that were treated with TPA plus magnetic fields when compared with the TPA plus sham-exposed mice. Although a difference in the cumulative number of mice with tumors was observed during the experiment, neither the number of animals affected nor the number of tumors per animal was statistically different between the two groups at the end of the study.
In a third study, the protocol was similar to that in the study described above, except that sham or field exposure continued for 52 weeks, whereas TPA treatment was discontinued after week 23. McLean et al. (McLean et al., 1995) reported that while there was no overall increase in total tumors associated with exposure to fields, more field-exposed animals had malignant tumors (8/48) than sham-exposed animals (1/48, p < 0.03). [No indication of the number of tumors per animal was given.] A review of three independent studies involving a total of 288 SENCAR mice used to study the effects of 60 Hz magnetic fields on the growth and development of skin tumors showed mixed results: in one study, more tumors were seen in field-exposed (86) than sham-exposed mice (48), while two studies showed the opposite effect (33 magnetic field-exposed versus 50 sham-exposed and 27 magnetic field-exposed versus 86 sham-exposed; p = 0.01). The authors concluded that the results did not support a role of magnetic fields as tumor co-promoters (McLean et al., 1997).
In a lifespan study of skin carcinogenesis in NMRI/HAN mice exposed to sinusoidal magnetic fields, no evidence was found that magnetic fields promote the formation of skin tumors (Rannug et al., 1993a). Groups of 30 mice were initiated with DMBA (25.6 µg applied topically to the shaved dorsal skin of each mouse) and exposed from seven weeks of age to either 50 Hz sinusoidal magnetic fields with flux densities of 50 or 500 µT for 103 weeks for 19-21 h/d or sham conditions. There was no increase in tumor promotion with exposure to magnetic fields. [The flux densities used in this study were lower than those used by McLean et al.] (McLean et al., 1995)
As part of this study, a group of animals was exposed to fields for two years in the absence of treatment with DMBA or TPA and were then monitored and evaluated over the course of their lifetime for survival and the appearance of skin tumors; at death, they were assessed by complete necropsy and histopathological evaluation of skin tumor types. The survival of mice treated with DMBA and TPA but exposed to 500 µT was shorter than that of controls or 50 µT-treated animals (median survival, 74 versus 94.5 and 87.5 weeks, respectively). There were no skin tumors in either control or field-exposed groups, and no significant differences were observed between sham-exposed and field-exposed groups in the incidences of other neoplastic lesions or leukemia.
In a second study, this group investigated the tumor promoting effects of continuous and intermittent magnetic fields in sensitive female SENCAR mice (Rannug et al., 1994). Groups of 40 mice were treated with DMBA (2.6 µg) one week before exposure to continuous and intermittent (15 s on/off) 50 Hz horizontal, AC fields with flux densities of 50 and 500 µT for 19 or 21 h/d for 104 weeks. Untreated, DMBA-treated, and TPA-promoted control groups were included. No tumors were found in animals treated with DMBA with no co-promotion by TPA and exposed to continuous fields of 50 or 500 µT. Four tumors in four animals were found at 50 µT and 13 tumors in five animals at 500 µT given intermittently, with two tumors in DMBA-treated sham-exposed animals. These increases were not significantly different when the two intermittent exposure groups were compared with the sham-exposed group. The time to first tumor was shortened in the intermittently exposed animals when compared with the DMBA-treated controls. [There were far fewer tumors in the field-exposed groups (< 5%) than in the TPA-treated positive control animals (> 97%). No hyperplastic response was seen with the magnetic field exposures. The results did not support the hypothesis that magnetic fields promote skin carcinogenesis in SENCAR mice at flux densities of 50 and 500 µT.]±
A further study was conducted in SENCAR mice (Sasser et al., 1998) as a collaborative effort between Battelle and the M.D. Anderson Cancer Center. Mice were exposed to DMBA (10 nmol in 200 µL acetone) and then to TPA at 0.85, 1.7, or 3.4 nmol. They were then exposed to 2000 µT, 60 Hz magnetic fields for 23 weeks. Skin tumor incidence and multiplicity were monitored, and the tumors were histologically evaluated at the end of the study. Papillomas were seen in 15 of 48 field-exposed and 10 of 48 sham-exposed mice. These differences were not significantly different. The overall conclusion of the authors was that, within the sensitivity limits imposed by the animal model and the exposure parameters employed, no co-promotional effect of field could be demonstrated.
[Several studies of promotion and co-promotion have been conducted in mice to examine the effect of 50 and 60 Hz magnetic fields on the development of skin carcinomas and papillomas. In all of these studies, skin carcinomas were initiated by treating the animals with a known chemical carcinogen; they were then exposed to various intensities of magnetic field or combinations of magnetic fields with a known chemical promoter (TPA). Within the constraints of all five studies, there were no significant promotional effects of magnetic fields on skin tumor development.
In a series of experiments, Rannug et al. (Rannug et al., 1993b; Rannug et al., 1993c) investigated the possibility that magnetic fields interact with known initiators or promoters of cancer to induce preneoplastic lesions in rats. They used partially hepatectomized male Sprague-Dawley rats treated with NDEA to initiate tumor development. The animals were then exposed to magnetic fields for 12 weeks, beginning one week after initiation, to determine whether growth of enzymatically altered foci would be promoted in liver cells. In two studies, groups of nine to 10 rats were initiated with NDEA and then promoted with 50 Hz horizontal magnetic fields with flux densities of 0.5 or 50 µT (experiment 1) or 5 or 500 µT (experiment 2). A slight increase in staining for -glutamyltranspeptidase was reported in the first experiment at 50 µT (p < 0.01) but not in the second experiment.
In a third study, groups of 10 rats were exposed to magnetic fields with flux densities of 0.5 or 500 T both during initiation with NDEA and throughout co-promotion with phenobarbital (300 ppm in diet) (Rannug et al., 1993c). The magnetic field inhibited, although not significantly, the size and number of focal lesions. The authors concluded that there was no evidence of a promotional or co-promotional role of magnetic fields.
As an addition to the study of Babbitt et al. (Babbitt et al., 1998), sections of brain were examined from mice exposed to 0, 350, 475, or 600 rads of ionizing radiation in four fractionated doses with and without subsequent exposure to 1.4 mT of circularly polarized 60 Hz magnetic fields. Hematoxylin and eosin-stained sections were prepared of the brain and were reviewed for primary proliferative lesions. Seven primary brain tumors or hamartous lesions (lipomas) were found in treated animals, with no apparent correlation with exposure to either radiation or magnetic fields. The authors concluded that this study provides no evidence of an effect of magnetic fields on primary brain tumors in female C57BL/6 mice (Kharzi, submitted manuscript).
[The sensitivity of this mouse model for agents that cause brain cancer has not been established, and ionizing radiation did not affect brain tumor incidence.]
A study in which DMBA was used as an initiator was conducted in newborn male and female Swiss-Webster mice (Shen et al., 1997). Each pup received a subcutaneous injection of 35 µg DMBA within 24 h of birth. Two weeks later, the mice were separated into either a sham-exposed group or a group exposed to a 1 mT, 50 Hz magnetic field. Exposure was continued for 3 h/d, 6 d per week for 16 weeks. The percentages of animals with thymic lymphomas and lymphomatous leukemia were 30% (50/165) of field-exposed and 30% (46/155) of sham-exposed mice. The authors reported no evidence for a promotional effect of a 1 mT, 50 Hz magnetic field on lymphoma/leukemia induced by DMBA in mice.
In the second study (Anderson et al., 1997), a similar protocol, exposure, and end-points were used. Leukemia cells were inoculated at 2.2 x 106 or 2.2 x 105, and intermittent field presentation (3 min on, 3 min off) was added, with replication of the continuous field exposures. Again, no significant exposure-related differences were observed for continuous fields at either level of cell inoculum or for intermittent fields at the lower cell inoculum; however, with intermittent fields at the higher inoculum an apparent decrease in the latency to disease (from 60 to 45 days) was observed in comparison with the sham-exposed animals. The authors reported that, taken together, the results for both inoculae argue for a lack of effect of magnetic fields on the progression of leukemia in this model. There remains, however, the slight effect of intermittent exposure, particularly when in animals with a higher load of injected leukemia cells.
[Neither studies in which an initiating event such as g-irradiation or a chemical carcinogen was used nor studies of progression of leukemia after injection of viable leukemic cells showed an effect of exposure to magnetic field at a variety of intensities. The absence of an increased incidence of leukemia/lymphoma in the long-term bioassay is consistent with these results.]
| Pim1 | Male | |||||
| Pim1 | Female | |||||
| p53 | Male | |||||
| p53 | Female | |||||
Groups of approximately 100 Eµ-Pim1 transgenic mice were exposed to 50 Hz magnetic fields at intensities of 0, 1, 100, or 1000 µT for up to 18 months (Harris et al., 1998). Animals that died during the study were subjected to histological evaluation. No difference in lymphoma incidence was seen between field- and sham-exposed mice. The authors concluded that long-term exposure to 50 Hz magnetic fields had no tumorigenic effect in the lymphoma-prone mice. The authors assumed that the animals did not have leukemia.
[Transgenic mouse models have not been fully evaluated for their predictive value for testing carcinogens. The design is also unusual in that the 'healthy' animals (nearly 50% of the animals) were discarded at the end of the study without examination. It was further noted that during the in-life portion of the study more than 7% of the animals were autolysed and discarded without diagnoses.]
In several long-term bioassays, no association was found between exposure to magnetic fields and brain cancer; however, the sensitivity of rodent models for assaying brain cancer has not been well established.
Most of the investigations carried out until now have followed the pattern of the traditional testing of chemical agents suspected to be carcinogenic. While additional traditional studies are fully justified and may produce useful results, it is conceivable that investigations of the role of the factors involved in the multistep, multifactorial carcinogenesis process (perhaps including EMF) may require different approaches than those used until now.
The overall conclusion of the Working Group is that most of the studies suggest a lack of carcinogenicity, and the few with borderline positive results are inadequate to conclude that exposure to magnetic fields at the magnitude and field configurations at which they were investigated increases the incidence of cancer in rodents.
There is inadequate evidence in experimental animals for carcinogenicity from exposure to extremely low frequency electromagnetic fields.
[This conclusion was supported by 19 members of the Working Group; there were 8 votes for 'lack' of carcinogenicity, 1 abstention and 1 absent.]
[There was a minority report written on this opinion; this report is in Appendix B]
| (NTP, 1998b) | F344 rats B6C3F1 mice | 100 males, 100 females/species/group | 2, 200, 1000 µT continuously(60 Hz) & 1 mT intermittently(60 Hz, 1 hour on, 1 hour off) | 2 years, GLP (18.5 h/d) | No increase in carcinogenesis (brain, mammary gland, or leukemia) in male or female rats or mice |
| (Mandeville et al., 1997) | F344/N rats | 50 females/group | 2, 20, 200, 2000 µT (60 Hz) | 2 year, GLP (20 h/d) | No increase in carcinogenesis |
| (Yasui et al., 1997) | F344 rats | 48 females/group | 0.5, 5 mT (50 Hz) | 2 year chronic,(22.6 h/d) | No increase in carcinogenesis |
| (Margonato et al., 1995) | Sprague-Dawley albino rats | 256 males/group | 5 µT (50 Hz) | 32 weeks (22 h/d) | No increase in carcinogenesis |
| (Beniashvili et al., 1991) | rats | 50 females/group | MNU | 20 µT (50 Hz); | lifelong(0.5, 3 h/d) | Increase incidence, decreased latency of mammary tumors in rats exposed for 3 h/d; more malignant tumors |
| (Mevissen et al., 1993) | Sprague- Dawley rats | 18 females/group | DMBA | 30 mT (50 Hz) (homogeneous field) | 13 weeks (24 h/d) | Increase tumor number per animal (66 vs. 61%); not reproduced upon repeat |
| (Löscher et al., 1994) | Sprague- Dawley rats | 36 females/group | DMBA | 0.3 - 1.0 µT (50 Hz) (gradient field) | 13 weeks (24 h/d) | No significant differences in histopathology; nocturnal melatonin significantly lower in exposed animals |
| (Löscher et al., 1993) | Sprague- Dawley rats | 99 females/group | DMBA | 0.1 mT (50 Hz) (homogeneous field) | 13 weeks (24 h/d) | 50% increase in mammary tumor incidence (51% exposed vs. 34% in sham-exposed) |
| (Baum et al., 1995) | Sprague- Dawley rats | 99 females/group | DMBA | 0.1 mT (50 Hz) | 13 weeks (24 h/d) | No effect on incidence of mammary tumors; significant increase in malignant tumor size in exposed |
| (Mevissen et al., 1996a) | Sprague- Dawley rats | 99 females/group | DMBA | 10 µT (50 Hz) (homogeneous field) | 13 weeks (24 h/d) | Negative (61% tumor incidence in sham vs. 66% in exposed) |
| (Mevissen et al., 1996b) | Sprague- Dawley rats | 99 females/group | DMBA | 50, µT (50 Hz) (homogeneous field) | 13 weeks (24 h/d) | Significant increase in tumors in exposed animals (apparent dose-response reported) |
| (Mevissen et al., 1998a) | Sprague-Dawley rats | 99 females/group | DMBA | 0.1 µT (50 Hz) (homogeneous field) | 13 weeks (24 h/d) | Significant increase in tumors in exposed animals; no change in tumor size |
| (NTP, 1998a) | Sprague- Dawley rats | 100 females/group | DMBA | 0.1- 0.5 mT (50 Hz)
0.1- 0.5 mT (50 Hz), 0.1 mT (60 Hz) | 13 weeks (18.5 h/d) 26 weeks (18.5 h/d) | Negative (43% tumor incidence in sham vs. 48% in 0.1 mT exposed and 38% in 0.5 mT exposed)Negative (96% tumor incidence in sham vs 90% in 0.1 mT (50 Hz) exposed and 95% in 0.5 mT exposed) |
| (Ekström et al., 1998) | Sprague- Dawley rats | 60 females/group | DMBA | 0.25 - 0.5 mT (50 Hz) (intermittent, 15 s, on/off) | 25 weeks (19-21 h/d) | Negative (71% tumor incidence in sham vs. 70% in exposed; exposure 1 week after DMBA) |
| (McLean et al., 1991) | SENCAR mice | 32 females/group | Negative at 22 weeks | |||
| (Stuchly et al., 1992) | SENCAR mice | 48 females/group | Decreased tumor latency, increased tumor incidence | |||
| (McLean et al., 1995) | SENCAR mice | 48 mice/group | Increased fraction with malignant tumors in exposed; non-significant increase in overall tumor incidence | |||
| (Rannug et al., 1993a) | NMR/HAN mice | 30 females/group | Negative | |||
| (Rannug et al., 1994) | SENCAR mice | 50 females/group | Negative for continuous; dose trend of increased tumors/tumor-bearing animal in intermittently exposed | |||
| (Sasser et al., 1998) | SENCAR mice | 100 females/group | Negative |
| (Rannug et al., 1993b) | Sprague-Dawley rats | 10 males/group | NDEA | 0.5, 5, 50, 500 µT (50 Hz) | 12 weeks (19 h/d) | Negative |
| (Rannug et al., 1993c) | Sprague-Dawley rats | 10 males/group | NDEA phenobarbital | 0.5 and 500 µT | 13 weeks (19 h/d) | Exposed and a decrease in no. of foci. (0.5 mT), mean focus area and volume of foci |
| (Babbitt et al., 1998) | C57Bl/6J mice | 195-450 males or females/group | X-irradiation (0, 350, 475, 600 R) | 1.4 mT circularly polar. 1 mT horizontal and vertical (60 Hz) | 2 year chronic (18 h/d) | Negative |
| (Shen et al., 1997) | Swiss-Webster mice | 165 exposed 155 sham- exposed | DMBA within 24 h of birth | 1 mT (50 Hz) | 32 weeks (3 h/d, 6 d/week) | No difference in lymphoma/leukemia between groups; increased liver infiltration in exposed |
| (Sasser et al., 1996) | Fischer 344 rats | 72 males/group | Leukemic cells injected (2.2 x 107) | 1 mT (60 Hz) continuous | 18 weeks (20 h/d) | Negative |
| (Anderson et al., 1997) | Fischer 344 rats | 72 males/group | leukemic cells injected (2.2 x 107 or 2.2 x 106) | 1 mT (60 Hz) continuous and intermittent | 18 weeks continuous and intermittent | Negative for continuous, decrease in latency for intermittent exposure in 107 cell inoculum |