International Archives of Occupational and Environmental Health

, Volume 76, Issue 6, pp 431–436

Intermittent extremely low frequency electromagnetic fields cause DNA damage in a dose-dependent way


    • Division of Occupational MedicineUniversity Hospital/AKH
    • Division of Occupational MedicineUniversity of Vienna
  • Elisabeth Diem
    • Division of Occupational MedicineUniversity of Vienna
  • Oswald Jahn
    • Division of Occupational MedicineUniversity of Vienna
  • Hugo W. Rüdiger
    • Division of Occupational MedicineUniversity of Vienna
Original Article

DOI: 10.1007/s00420-003-0446-5

Cite this article as:
Ivancsits, S., Diem, E., Jahn, O. et al. Int Arch Occup Environ Health (2003) 76: 431. doi:10.1007/s00420-003-0446-5



Epidemiological studies have reported an association between exposure to extremely low frequency electromagnetic fields (ELF-EMFs) and increased risk of cancerous diseases, albeit without dose–effect relationships. The validity of such findings can be corroborated only by demonstration of dose-dependent DNA-damaging effects of ELF-EMFs in cells of human origin in vitro.


Cultured human diploid fibroblasts were exposed to intermittent ELF electromagnetic fields. DNA damage was determined by alkaline and neutral comet assay.


ELF-EMF exposure (50 Hz, sinusoidal, 1–24 h, 20–1,000 μT, 5 min on/10 min off) induced dose-dependent and time-dependent DNA single-strand and double-strand breaks. Effects occurred at a magnetic flux density as low as 35 μT, being well below proposed International Commission of Non-Ionising Radiation Protection (ICNIRP) guidelines. After termination of exposure the induced comet tail factors returned to normal within 9 h.


The induced DNA damage is not based on thermal effects and arouses concern about environmental threshold limit values for ELF exposure.


ELF-EMF50-Hz sinusoidalIntermittent exposureComet assay


The issue of adverse health effects of extremely low frequency electromagnetic fields (ELF-EMFs) is highly controversial. Numerous contradictory results regarding the carcinogenic potential of ELF-EMFs have been reported in the literature. Some epidemiological studies indicate that exposure to ELF-EMFs may lead to an increased risk of certain types of adult and childhood cancer, including leukaemia, cancer of the central nervous system, and lymphoma (Wertheimer and Leeper 1979; Savitz et al. 1988; Feychting et al. 1997; Li et al. 1997), while others (Verkasalo et al. 1993; Tomenius 1986; Schreibner et al. 1993) did not find such an association. Interpretation of these studies is complicated, due to different and unreliable methods of exposure assessment. Therefore, in vitro studies with defined exposure conditions and with genotoxic effect markers as endpoints could provide evidence for a carcinogenic potential of ELF-EMFs.

Up to date, several comprehensive reviews regarding in vivo and in vitro laboratory studies on ELF-EMFs have been published (McCann et al. 1993, 1998; Murphy et al. 1993; Moulder 1998). Conflicting results have been reported, with genotoxic endpoints such as sister chromatid exchange (SCE), micronuclei (MN), chromosome aberrations (CA) and assessment of DNA strand breaks at exposure levels ranging from 1 μT to 10 mT. The majority of these studies, however, did not show any EMF-related genotoxic effects. Several studies with whole-body exposure of rodents to ELF-EMFs revealed DNA single-strand and double-strand breaks in the brain (Lai and Singh 1997; Singh and Lai 1998; Svedenstal et al. 1999a, 1999b). These results, in the first place, gave rise to the classification of ELF magnetic fields as being possibly carcinogenic to humans (group 2B) by the International Agency for Research on Cancer (IARC 2002).

As previously reported (Ivancsits et al. 2002a) we were able to corroborate these findings by demonstration of an increase in DNA single-strand breaks (SSBs) and double-strand breaks (DSBs) in cultured human diploid fibroblasts upon intermittent exposure to a 50-Hz magnetic field, using comet assay under alkaline (detection of SSBs + DSBs) and neutral conditions (detection of DSBs). The extent of EMF-induced DNA damage was variable in relation to the setting of on and off times, and was highest at an intermittence of 5 min on/10 min off. No effects were detected during continuous exposure.

Here we report studies on the influence of exposure time and of magnetic-flux densities on the induction of DNA strand breaks with human fibroblast cultures of three healthy donors.

Materials and methods

ELF-EMF exposure conditions and cell culture

Human diploid fibroblast strains of donors with different ages (ES1, male, 6 years old; IH9, female, 28 years old; KE1, male, 43 years old) were initiated from skin biopsies from healthy donors and maintained in culture as previously described (Ivancsits et al. 2002a). The cells were seeded into 35-mm Petri dishes at a density of 5×104 cells/3 ml, 24 h prior to ELF-EMF exposure.

The exposure system was built and provided by the Foundation for Information Technologies in Society (IT'IS foundation), Zurich, Switzerland, The set-up, which generated a vertical EMF, consisted of two four-coil systems (two coils with 56 windings, two coils with 50 windings), each of which was placed inside a μ-metal box. The currents in the bi-filar coils could be switched parallel for field exposure or non-parallel for control (sham-exposure). The residual magnetic field in the sham chamber was at least 150 times (43 dB) lower than the applied field in the exposure chamber. In addition, both chambers were not completely insulated from the earth's magnetic field, which remained at 20–50 μT. Both systems were placed inside a commercial incubator (BBD 6220, Kendro, Vienna, Austria) to ensure constant environmental conditions (37°C, 5% CO2, 95% humidity). Two fans per μ-metal box ensured atmospheric exchange of the chambers. A PC controlled and continuously monitored the exposure set-up. Data (temperature, current) were collected and stored in an encoded file. The temperature was monitored at the location of the dishes during exposure and was maintained at 36.5–37.5°C. The temperature difference between the chambers did not exceed 0.3°C. A current source based on four audio-amplifiers (Agilent Technologies, Zurich, Switzerland) allowed magnetic fields up to 2.3 mT. The field could be varied in the frequency range from DC to 1.5 kHz by a computer-controlled function generator. To enable blind exposures, the computer randomly determined which of the two chambers was exposed. This information was provided to the investigator by the IT'IS foundation in Zurich via e-mail in exchange with the transmission of comet assay results. All experiments were performed at a frequency of 50 Hz sinusoidal at intermittent exposure (5 min field on/10 min field off). Time-dependent effects were studied at a magnetic flux density of 1 mT; for dose–response effects, the magnetic flux density was varied between 20 and 1,000 μT (5 min field on/10 min field off) at a constant exposure time of 15 h. After exposure the fibroblasts were detached with trypsin and suspended in fresh culture medium. To study repair kinetics, we post-incubated fibroblasts at 37°C for 0.5–9 h. Each exposure level was tested in duplicate.

Comet assay analysis

We followed the technique described by Östling and Johanson (1984) with minor modifications by Singh et al. (1988, 1991). ELF-exposed and sham-exposed cells (10,000–30,000) were mixed with 100 μl low-melting agarose (0.5%, 37°C), and this cell suspension was pipetted onto 1.5% normal-melting agarose pre-coated slides, spread with a cover slip, and kept on a cold flat tray for approximately 10 min to solidify. After the cover slip had been removed, a third layer of 0.5% low-melting agarose was added and allowed to solidify. The slides were then immersed in freshly prepared cold lysis solution (2.5 mol/l NaCl, 100 mmol/l Na2EDTA, 10 mmol/l Tris, pH 10, 1% sodium sarcosinate, 1% Triton X-100, 10% DMSO, pH 10) and lysed for 90 min at 4°C. Subsequently, the slides were drained and placed in a horizontal gel electrophoresis tank, side by side and nearest the anode. The tank was filled with fresh electrophoresis buffer (1 mmol/l Na2EDTA, 300 mmol/l NaOH, pH>13 or pH 12.1 in the case of alkaline comet assay, and 100 mmol/l Tris, 300 mmol/l sodium acetate, 500 mmol/l sodium chloride, pH 8.5 in the case of neutral comet assay) to a level approximately 0.4 cm above the slides. For both alkaline and neutral comet assay, the slides were left in the solution for 40 min to allow equilibration and unwinding of the DNA before electrophoresis. Electrophoresis conditions (25 V, 300 mA, 4°C, 20 min, field strength 0.8 V/cm) were the same for neutral and alkaline comet assay. All steps were performed under dimmed light to prevent the occurrence of additional DNA damage. After electrophoresis the slides were washed three times with Tris buffer (0.4 mol/l Tris, pH 7.5), to be neutralized, then air-dried and stored until required for analysis. Comets were visualized by ethidium bromide staining (20 μg/ml, 30 s) and examined at 400× magnification with a fluorescence microscope (Axiophot, Zeiss, Germany). One thousand DNA spots from each sample were classified into five categories corresponding to the amount of DNA in the tail, in accordance with Anderson et al. (1994). The proposed classification system provides a fast and inexpensive method for genotoxic monitoring. Due to the classification to different groups by eye, no special imaging software is required. The technique becomes more sensitive, because many cells can be scored in a short time (1,000 cells instead of 50–100 cells with image analysing). The subsequent calculation of a "comet tail factor" allows DNA damage to be quantified as a single figure, which makes it easier for results to be compared. Due to the scoring of 1,000 cells in one experiment, which are ten times the cells processed with image analysing, standard deviations are very low. Reproducibility has been thoroughly checked.

Results were expressed as "comet tail factors", calculated in accordance with Diem, with modifications as previously described (Diem et al. 2002; Ivancsits et al. 2002a, 2002b). The same investigator performed all analyses. Figure 1 shows the five classification groups, with the group averages, and the microphotograph.
Fig. 1.

Comet assay classification groups and respective microscopic appearance (cell line ES-1)

Tail factors were calculated according to the following formula:
$$ {\rm{tailfactor}}\% = {{{\rm{A}}*{\rm{F}}_{\rm{A}} + {\rm{B}}*{\rm{F}}_{\rm{B}} + {\rm{C}}*{\rm{F}}_{\rm{C}} + {\rm{D}}*{\rm{F}}_{\rm{D}} + {\rm{E}}*{\rm{F}}_{\rm{E}} } \over {1000}} $$

A = the number of cells classified to group A and FA = the average of group A (=2.5)

B = the number of cells classified to group B and FB = the average of group B (=12.5)

C = the number of cells classified to group C and FC = the average of group C (=30)

D = the number of cells classified to group D and FD = the average of group D (=67.5)

E = the number of cells classified to group E and FE = the average of group E (=97.5)

Statistical analysis

Statistical analysis was performed with STATISTICA V. 5.0 package (Statsoft, Tulsa, USA). All data are presented as mean ± SD. The differences between exposed and sham-exposed, as well as between different exposure conditions, were tested for significance with an independent Student's t-test. A difference at P< 0.01 was considered statistically significant.


Fibroblast cultures of three healthy donors were exposed to ELF-EMFs (50 Hz sinusoidal, 1,000 μT, intermittent 5 min on/10 min off) for 1 to 24 h. Alkaline and neutral comet tail factors increased with exposure time, being largest at 15–19 h (Fig. 2). Comet assay levels declined thereafter, but did not return to basal levels. The different cell donors exhibited different basal levels, different maxima, and different end levels.
Fig. 2a, b.

Influence of exposure time on formation of DNA SSBs and DSBs in three human fibroblast strains (ES-1, IH-9, KE-1) determined with comet assay (1 mT, 5 min on/10 min off cycles). a Alkaline conditions; b neutral conditions

When exposure was terminated after 15 h the comet factor returned to basal levels after a repair time of 7 to 9 h (Fig. 3a), which comprised a fast repair rate of DNA SSBs (<1 h) and a slow repair rate of DNA DSBs (>7 h). The marked comet peak value between 12 and 17 h and the following repair kinetics could also be detected when ELF exposure was terminated after 12 h (Fig. 3b). However, it disappeared when comet assay was performed at pH 12.1 instead of pH>13, thereby eliminating the cleavage of alkali labile sites in the DNA (Fig. 4).
Fig. 3a, b.

Repair kinetics of DNA SSBs and DSBs in human fibroblasts after termination of ELF-EMF exposure (cell strain ES-1, 1 mT, 5 min on/10 min off cycles) determined with alkaline and neutral comet assay. a Repair after 15-h ELF-EMF exposure; b repair after 12-h ELF-EMF exposure

Fig. 4.

Comet assay of exposed human fibroblasts was performed at different pH (1 mT, intermittent 5 min on/10 min off)

When magnetic flux densities were varied between 20 and 1,000 μT (cell strain ES-1) we observed a dose-dependent increase of comet factor at alkaline and neutral conditions, which had already become significant at 35 μT (Fig. 5, Table 1).
Fig. 5.

Dose-dependent formation of DNA SSBs and DSBs determined with comet assay under alkaline and neutral conditions with cell strain ES-1 (exposure time 15 h, 5 min on/10 min off cycles)

Table 1.

Alkaline and neutral comet assay levels of ELF-exposed (20–1,000 μT, 15 h, 5 min on/10 min off) and non-exposed human diploid fibroblasts (ES-1)

Magnetic flux density

Alkaline comet assay

Neutral comet assay


Comet tail factor (%)

(± SD)

Comet tail factor (%)

(± SD)




































aSignificant differences (P<0.01) between exposed and non-exposed


Exposure to thermal stress may result in alterations in the integrity of DNA, comprising DNA strand breaks or apoptosis (Fairbairn et al. 1995). The induced DNA damage depends on the extent and duration of the applied heat stress. Taking these findings into account, we consider it highly unlikely that, in our experiments, the observed genotoxic damage is caused non-specifically by spots of increased temperature within the cell layer as a secondary effect of the electromagnetic field. If so, the damage would increase with prolongation of the on time during the intermittent exposure and would be largest at continuous exposure. It has previously been shown, however, that the largest effects are obtained at 5′ on/10′ off cycles, and that continuous exposure has no effect at all (Ivancsits et al. 2002a). Therefore, we conclude that the observed induction of DNA SSBs and DSBs is a direct consequence of an intermittent exposure to ELF-EMFs.

We observed an increase in DNA breaks up to 15 h of exposure and then a decline to a "steady-state level" of approximately 1.5-times the base line. This unexpected finding can be explained if the exposure activates DNA repair processes and this activation takes a time of 10 to 12 h. After this time the DNA damage is repaired at an enhanced rate, which leads to a reduction in DNA breaks, albeit not to a normalisation. This explanation is supported experimentally by the observation that the single-strand DNA breaks (alkaline conditions) are repaired after approximately 30 min, and double-strand breaks 7 to 9 h after shut down of the exposure (Fig. 3). The repair process itself also leads to a temporary increase of alkali-sensitive sites in the DNA, which are detected as a peak at hours 12 to 17 at comet assay conditions of pH>13, but not of pH 12.1, the latter not being able cleave the alkali-sensitive sites (Fig. 4).

It is well known that the repair of SSBs is a fast and almost error-free process, while the repair of more complex DNA damage (i.e. DNA DSBs) by homologous recombination, single-strand annealing or non-homologous end joining requires more time and is error prone in part (Van den Bosch et al. 2002). Therefore, DNA DSBs may affect the integrity of the genome and can lead to cell death, uncontrolled cell growth, or cancer (Van Gent at al. 2001).

In addition, we demonstrate here an increase in DNA SSBs and DSBs in relation to an increasing magnetic flux density, which becomes significant at 35 μT at 15 h of intermittent ELF-EMF exposure. This threshold is well below the guidelines of the International Commission of Non-Ionising Radiation Protection (ICNIRP 1998), which propose 500 μT per working day for occupational exposures and 100 μT per 24 h for the general population. However, no proposal with regard to intermittent exposures has been made by the ICNIRP as yet.

In conclusion, our findings strongly indicate a genotoxic potential of intermittent ELF-EMFs. The induced DNA damage was time-dependent and dose-dependent and points to the need for consideration of environmental and occupational threshold limit values for ELF-EMFs, in particular with regard to intermittent exposures.


This study was funded by the European Union under the programme "Quality of Life and Management of Living Resources", Key Action 4 "Environment and Health": QLK4-CT-01574.

Copyright information

© Springer-Verlag 2003