Review | Published:

“Dirty electricity”: what, where, and should we care?

Journal of Exposure Science and Environmental Epidemiology volume 20, pages 399405 (2010) | Download Citation

Abstract

Environmental exposure to high-frequency voltage transients (HFVT), also termed dirty electricity, has been advocated among electro(hyper)sensitive interest groups as an important biological active component of standard electromagnetic pollution. A literature search was conducted in PubMed, in which only seven articles were identified. Exposure to HFVT was associated with increased cancer risks, while preferential removal of 4–100 kHz HFVT from 50–60 Hz ELF circuits was linked to a variety of improvements in health (plasma glucose levels in diabetic patients, symptoms of multiple sclerosis, asthma and other respiratory illnesses, and insomnia), well-being (tiredness, frustration, general health, irritation, sense of satisfaction, mood), and student behavior. However, all these published studies were subject to significant methodological flaws in the design of the studies, the assessment of exposure, and the statistical analysis, which prevented valid assessment of a causal link between this exposure metric and adverse effects. Environmental exposure to HFVT is an interesting EMF exposure metric, which might explain the spurious results from epidemiological studies using ‘standard’ ELF and RF exposure metrics. However, at present, methodological problems in published studies prohibit the valid assessment of its biological activity.

Introduction

Recent decades have seen a huge increase in the use of electronic devices as well as the birth and subsequent explosion in the use of wireless technology, resulting in increased exposure of nearly everyone to wide ranges of frequencies of non-ionizing radiation in all aspects of daily life. Besides the benefits, this has also resulted in increased public concern about potential adverse health effects of these exposures (Aldrich and Easterly, 1987; Ahlbom and Feychting, 2003; Genuis, 2008; Hardell and Sage, 2008; BioInitiative Report: A Rationale for a Biologically-based Public Exposure Standard for Electromagnetic Fields (ELF and RF) Online (2009)). The literature assessing potential adverse effects on human health from exposure to electromagnetic fields (EMF) in the extremely-low frequency (ELF) to radiofrequent (RF) frequency ranges encountered in the general environment is enormous and is increasing rapidly (for example Ahlbom et al., 2004). However, despite of all these data, a conclusion on whether environmental exposure to EMF causes adverse health effects in human population remains to be answered (Hardell and Sage, 2008; Roosli, 2008; IARC, 2009; Schuz et al., 2009).

Data suggest that somewhere between 3–35% of the population (Havas, 2006), claim to be sensitive to electromagnetic field exposure in their environment, although this number is more likely closer to 3% than to 35% (Leitgeb and Schrottner, 2003). Susceptibility to electromagnetic fields is generally referred to as “electrosensibility” or “electromagnetic hypersensitivity (EHS)”, depending on the severity of the effects (Leitgeb and Schrottner, 2003). EHS patients commonly report a variety of adverse effects when exposed to EMF, including somatic (such as skin and muscular skeletal problems and gastrointestinal disturbances) and neurasthenic (including fatigue, headaches, concentration problems and sleep disturbances) symptoms (Levallois, 2002), of which some can be measured on a cellular level (Johansson, 2006). There seems to be geographical variation in both the type of symptoms reported and the attributed source association (Irvine, 2005). However, EHS individuals have not been able to recognize exposed from sham conditions under controlled experiment conditions (Rubin et al., 2005; Roosli, 2008), and thus appear not to fulfill the fundamental criteria for proof of causality (Ward, 2009). In contrast, experimental data showed that people from the general population are able to detect 50 Hz electric currents well below an established perception threshold limit of 500 μA (IEC, 1994), with on average women being more susceptible than men (Leitgeb and Schroettner, 2002).

Nonetheless, although the exact magnitude of effects of EMF on population health cannot be assessed, largely due to misclassification and bias in the assessment of exposure (Kundi, 2009), most studies suggest that risks, if present, are likely to be small to moderate. Given that most people nowadays are routinely exposed to EMF, this would however imply that potentially large numbers of people could suffer from adverse health effects or diminished wellbeing related to EMF exposure. Large research consortia, such as for example the INTERPHONE study (Cardis et al., 2007), have been initiated to address public concern regarding effects of EMF. In their shadow, smaller studies are being conducted, some of which providing a novel approach to the question of how exposure to EMF could affect human beings and why results from most environmental studies remain spurious. This review discusses one such concept, which describes a novel exposure metric to which it has been argued biological systems are susceptible. The use of this exposure metric has recently grown in popularity among EHS groups, and is known by them as “dirty electricity”.

Dirty electricity

The EMF frequency bands that are generated in connection with the production, transmission, distribution and use of electrical power usually have a frequency of 50–60 Hz (Ahlbom and Feychting, 2003; Leitgeb and Schrottner, 2003). The term dirty electricity refers to the electromagnetic energy flowing along a conductor that deviates from a pure 50/60 Hz sine wave and has both harmonic and transient properties. The name ‘dirty electricity’ originates from the term ‘dirty power’ used in industry for the high frequency voltage transients that are caused by interruptions in the electrical current flow from connected electrical equipment and that can damage equipment (Milham and Morgan, 2008). These deviations from the 50/60 Hz sine wave are generally in the lower radiofrequent (RF) spectrum, and instead of the misplaced term ‘dirty electricity’ should more correctly be referred to as ‘electromagnetic fields generated by RF transients’ or ‘high frequency voltage transients’.

In industry, these high frequency voltage transients (HFVT) can be removed to protect equipment from power surges by installing large capacitors. Similarly, in the home or office environment filter units have been developed (Graham/Stetzer (G/S) filters (Graham, 2000, 2002), although similar units are produced by other companies as well), which utilize the same principle as the industrial capacitors to remove or reduce the amount of HFVT on electrical circuits with an optimal filtering capacity between 4 kHz and 100 kHz. This implies that exposure to HFVT is not similar to RF exposure, which has been defined as being within the frequency range of 100 kHz to 300 GHz (Ahlbom et al., 2004), but rather within a new EMF spectrum ranging from low frequency EMF (Feychting et al., 2005) up to the very low spectrum of RF.

HFVT levels on the main circuits can be measured using standard RF and low frequency EMF measurement equipment. However, a dedicated microsurge meter has also been developed which measures the magnitude of the rate of change of voltage (as R.M.S.) as a function of time, preferentially in the higher frequency transients between 4 and 150 kHz (Milham and Morgan, 2008). This microsurge meter expresses the higher frequency transients in somewhat obscure G/S units, which range from 1 to 1999 (Graham, 2003). Although no objective criteria have been developed, it has been argued that values below 30 G/S units are desirable and that the environment for individuals claiming to be extremely sensitive to electromagnetic fields should be below 20 G/S units (Havas, 2006).

In general, a positive correlation exists between the level of 50/60 Hz electromagnetic fields in environments and HFVT levels, and as such ELF exposure measures used in many epidemiological studies (Wood, 1993; Roy and Martin, 2007; Schuz et al., 2009) may in fact be a surrogate for HFVT exposures (Milham and Morgan, 2008).

Materials and methods

A literature search was performed for peer-reviewed publications, specifically assessing dirty electricity or high frequency voltage transients as the metric of exposure rather than the conventional ELF/RF exposure measures, in PubMed (up to July 31, 2009). This search resulted in only five publications (Havas, 2002, 2006, 2008; Havas and Olstad, 2008; Milham and Morgan, 2008). References from the literature lists of these papers retrieved one additional journal paper in which a case study was described (Genuis, 2008) and one conference proceeding (Havas et al., 2004). In this review, these studies will be discussed in terms of the methods used and the assessment of exposure used to quantify exposure-response associations.

Results

Health Effects

In the few published papers, HFVT carried on the 50/60 Hz sinewave have been linked to a surprisingly diverse variety of adverse effects.

Milham and Morgan (2008) in their study among school teachers describe a cancer cluster of 18 cancer cases (n=16 teachers) among 137 teachers in a middle school, which included cases of malignant melanoma (n=4), Burkitt's lymphoma (n=1), polycythemia vera (n=1), multiple myeloma (n=1), leiomysosarcoma (n=1), and cancers of the thyroid (n=2), uterus (n=2), colon (n=1), pancreas (n=1), ovary (n=1), larynx (n=1), and female breast (n=2). Excess cancer risk was compared with expected number of cancer cases in the general population and additional analyses included associations with years of employment at the school and estimated exposure levels in the classrooms they generally worked. However, interpretation of a causal association from these data is fraught with problems given an apparent bias in selection of the general population sample. These data suggest that even at the lowest exposure levels (comparable to the general population) almost twofold increased cancer risks compared to the general population exists, which does further increase with cumulative EMF exposure, which in this study can also be regarded as employment duration. Furthermore, in a subsequent letter to the editor, Morgan discussed that contrary to what the authors reported the cancer cases were not verified by the cancer registry and inaccuracies were found in the reported case definitions (Morgan, 2009).

Havas published two publications on chronic health effects of HFVT (Havas, 2006, 2008). Both publications discuss case studies in which removal of high frequency voltage transients from the environment has led to symptom improvements for patients. She describes (Havas, 2006) strong associations between blood sugar levels in two self-reported EHS diabetic patients (one type 2 and one type 1) and HFVT levels. In addition, two case-studies of patients diagnosed with multiple sclerosis were described; one teacher reported a variety of neurasthenic symptoms while the second case reported somatic conditions. After installment of HFVT filters at the school (case 1) or home (case 2) symptoms disappeared completely within days, with additional further improvements in mobility after one week in the second case. She further introduced (Havas, 2008) a third type of diabetes mellitus, referring to diabetic patients whose blood sugar levels respond to electromagnetic exposure. Genuis (Genuis, 2008) described four case studies in which cases’ health symptoms disappeared after minimizing EMF exposure. Only one of these case descriptions specifically discussed “dirty electricity” as the relevant exposure metric and described a case of chronic insomnia that disappeared within a week after power surges were reduced from 1,600 G/S units to below 30. Although surprisingly strong health benefits are described associated with a reduction in the level of HFVT exposure, these studies all suffer from methodological problems preventing the establishment of a causal link between the specific exposure and the reported health effects. Most notably, none of these case studies were blinded to the exposure. Generally, this tends to increase treatment effects compared to blinded studies (Schulz et al., 1995; Day and Altman, 2000), and is especially problematic in these studies where the cases were self-reported EHS patients. Studies have shown that placebo or nocebo effects can modulate perception and biological processes inducing measurable physiological changes (Enck et al., 2008; Zubieta and Stohler, 2009), and indeed changes in blood glucose levels in these studies can alternatively plausibly be ascribed to expectation and anxiety (Surwit et al., 1992; Surwit and Schneider, 1993; Park et al., 2008) rather than to the reduction in exposure. Furthermore, both exposure and biological response are self reported and have been measured at random moments throughout the day. Given the natural variation in blood glucose levels throughout the day (Malherbe et al., 1969; Molnar et al., 1972), the strong correlation with the level of exposure and little remaining (natural) variability reported in these studies is very surprising.

The effects of the removal of HFVT exposure on reported adverse effects on health and wellbeing have also been assessed specifically in schools. A study by Havas and colleagues in a school for children with learning disabilities (Havas et al., 2004), describes an intervention study in which HFVT-filters were installed throughout the school for a 2-week period followed by a 1-week post-filter reference period. During both periods teachers (n=18) completed one questionnaire twice daily on student behavior and a second questionnaire on their own physical wellbeing and performance at the end of the day. Teachers reported that they felt less tired and less frustrated, healthier, less irritable, had a greater sense of satisfaction and had an improved mood in the weeks the filters were installed compared to the weeks without the filters. Teachers further reported improved student behavior when the filters were installed, although this depended on the student grade. This was attributed to (a) higher grades changed classrooms more often than lower grades resulting in more variable exposure, and (b) a negative correlation between susceptibility to EMF exposure and age of the students. Again, this study also suffered from problems in the blinding of exposure to the study individuals. Although, teachers were blinded to the nature of the research, they were not blinded to the exposure since G/S filters were only installed in the 2-week measurement period (although the authors commented most would be unnoticed). Furthermore, even though exposure was not randomized, no information was collected on potential confounding variables. Differences in student behavior and tiredness among teachers (which, since both are reported by the teachers, will be linked in a virtuous circle) could similarly be caused by difference in coping load related to other factors, such as for example different courses in that period, different general stress level, exams, or longer hours (Bauer et al., 2006; Hjern et al., 2008).

A second study (Havas and Olstad, 2008) was conducted among three schools. Although comparable in design to the previous study, the design of the study was improved by including sham exposure periods using dummy filters. The range of improvements in wellbeing reported by the teachers during the “filter periods” was larger than those reported in the previous study and included improvements in the incidence of headaches, general weakness, dry eye or mouth, facial flushing, depression, mood, dizziness, pain, skin irritation, clarity of thought and more energy. In addition to the first study, removal of HFVT in this study was also further associated to decreased incidence of asthmatic symptoms and other respiratory symptoms among teachers. Although marginal improvements in student behavior in the elementary and middle school were noted, these were absent in the high school. The authors suggested that the absence of an effect at high school level might be contributed to other sources of RF radiation, such as cell phone use and wireless computing, frequent changes in class rooms, or in concordance with the previous study, that students with attention deficit disorder (ADD) or attention deficit hyperactivity (ADHD) might be more sensitive to EMF energy. Although utilizing an improved design, again exposure was not randomized and no information on potential confounding variables was collected.

In summary, there are serious methodological problems in the design of all of these studies and they cannot be used to support the existence of a causal link between HFVT exposure and adverse health effects. These flaws prevent a valid assessment of potential adverse effects associated with high frequency voltage transient exposure in the environment.

Exposure Assessment

All studies described in this review made use of G/S filters to filter out the high frequency voltage transients. However, exposure itself was measured using a variety of different equipment.

In the study by Milham and Morgan (2008), HFVT were measured using the G/S microsurge II meter while magnetic fields were measured using a FW Bell model 4080 tri-axial Gaussmeter (range 25–1,000 Hz) and a Dexsil 310 Gaussmeter (range 30–300 Hz). Measurements were conducted on three occasions: ones by the researchers using the Bell meter and G/S microsurge meter (n=7), later that year by the teachers with the same equipment (n=35) and approximately one year later by electrical consultants using the G/S microsurge meter and a Dexsil 320 Gauss meter on multiple outlets per room (n=51). However, only the last measurements were used for the exposure assessment in the analyzis of the cohort. Reference measurements were obtained from another elementary school, an office building and from 125 private residences, using the G/S microsurge II meter only. A total of 631 rooms were surveyed. However, no field exposures or personal exposure measurements were collected in this study as the researchers were denied access to the school.

The other studies conducted in schools (Havas et al., 2004; Havas, 2008) also had very limited exposure assessment in which personal exposure measurements or reference measurements were not collected. Instead, exposure assessment was solely based on weekly power quality measurements in rooms where teachers or students spent their days using a Fluke 79 III meter (up to 20 kHz) connected to a Graham ubiquitous filter to remove the 50/60 Hz signal (Havas et al., 2004) or using a Microsurge meter, in the weekend when the G/S filters were installed with the light switched on but with unknown information about other electrical equipment (4–100 kHz). (Havas and Olstad, 2008). Only limited magnetic field exposure data were collected from spot measurements in randomly selected class rooms using a trifield meter (n=2 per school).

Exposure assessment in the case reports was, if mentioned at all, marginal and limited to either power quality measurements using a Protek 506 Digital Multimeter connected to a G/S filter measured by the case himself (Havas, 2006, 2008) or using the Microsurge meter without any additional information (Genuis, 2008). For one case reported in (Havas, 2006), only self-reported movement into and from “dirty” environments was used as a metric for exposure.

In summary, the exposure assessment strategy is all these studies was poor. In fact, besides one study in which a few magnetic field spot measurements were collected, “real” exposure measurements in the environment, or preferably on the individuals themselves as a measure for personal exposure, were not conducted. As such, no data are available about what effects the installment of the G/S filters has on personal ELF and RF exposure levels. Neither is any data available on changes in the spatial variability of ELF and RF exposure levels in the rooms where the filters were installed. Furthermore, as no data has been collected on temporal variability of exposure, it remains unclear whether the short-time spot measurements or average exposure levels used in these studies have any relevance to longer term time-integrated exposure. Finally, by not including personal or field exposure measurements the influence of other sources of exposure, such as mobile phone base stations (Hutter et al., 2006; Roosli, 2008; Kundi and Hutter, 2009) which, although to a large extent in different frequency ranges, have been shown to be important determinants to total EMF exposure (Frei et al., 2009), is not taken into account.

Statistical Analyses

A major similarity between all studies is that the sample sizes are all relatively small. The studies conducted in schools described in this review included respectively 18 cancer cases among 137 teachers (Milham and Morgan 2008), or self-reported effects from 18 (Havas et al., 2004) and 44 (Havas and Olstad, 2008) teachers. The other studies described case studies (Havas, 2006, 2008) only.

Furthermore, formal statistical assessment of the data was only conducted in the study by Milham and Morgan (2008). Studies by Havas et al. involving more than one case instead relied on the absolute number of individuals reporting improved or decreased subjective measures of wellbeing, without any formal statistical testing of the differences (Havas et al., 2004; Havas and Olstad, 2008). However, regardless of whether the studies would have been larger in size and properly analyzed, the flaws in their designs, as described above, would have still prevented establishment of any causal links.

Discussion

It has been argued that the current approach in which a scientifically sound conclusion on health risk of a specific exposure is based on review of all available studies and if deemed necessary, subsequent amendment of protection strategies renders this debate open to criticism and controversy, which is routinely exploited by interest groups to prevent introduction of measures to reduce exposure levels (Carpenter and Sage, 2008; Genuis, 2008). This has been especially persistent in EMF research due to the intransigent disbelief of some vocal scientists and by delayed EMF legislation by injection of confusion and doubt into the scientific debate by focusing on uncertainties and deflecting attention from harm potential by vested interests (The International Commission for Electromagnetic Safety, 2006; Genuis, 2008). It has been argued that, whether causality has been proven or not, the associations between EMF exposure and health effects are sufficiently strong to warrant actions to reduce exposure levels in the general population (Carpenter and Sage, 2008; Genuis, 2008). However, it is precisely the discussion and subsequent scientific testing of these uncertainties that are of the utmost importance for the correct interpretation of data on potential health effects of electromagnetic field exposure. As hardly any populations in the world can be found without exposure to electromagnetic fields, even if the true risk of EMF exposure is only small to moderate and the effects on the population attributable risk could be enormous (Carpenter and Sage, 2008; Genuis, 2008). Yet, precautionary measures needed to reduce exposure on a population level will be draconian and without a valid scientific basis for exposure limits or even a biological mechanism for long-term effects to act upon will be unfeasible to introduce, if a causal relation does exist at all. The studies described in this review are stark examples of how problems in the design of studies and in the interpretation of the data can suggest important exposure-response associations where a closer evaluation of the data shows these cannot be used to argue against or in favor of the null hypothesis and as such infuse the scientific debate with further confusion. Rather than implementing new guidelines and exposure-reduction methods the available data ask for new, properly conducted and well-powered studies to assess whether high frequency voltage transient exposure might in fact be the biologically active exposure metric.

Whether the spurious and often contradictory results in epidemiological studies assessing adverse effects related to 50/60 Hz electromagnetic field exposure could be due to ELF exposure being in fact a proxy for another exposure metric — in this case high frequency voltage transients (“dirty electricity”) — is an interesting scientific hypothesis worthy of additional research. However, at present, the number of peer-reviewed publications discussing potential health effects of HFVT is limited, while those that have been peer-reviewed and published are all subject to methodological problems casting severe doubt on the reported causality of the observed associations between the exposure and the wide variety of adverse effects. Apparently, the body of evidence is more robust since references are made to other patients that have been ‘cured’ using similar methods as those described in this review (Havas, 2006, 2008). As such, claims made in these peer-reviewed papers apparently also refer to others, although none have been discussed outside of the “grey” or informal circuit. Nonetheless, in peer-reviewed publications references to these “informal circuit” cases are exploited as the basis for the effectively of the removal of “dirty electricity” by HFVT filters. High frequency voltage transients as a biological active exposure metric are at present merely speculation. This should however, not be a reason to dismiss it as a potential biologically active exposure, but rather be an incentive for conducting new studies with improved study designs and exposure assessment.

It is important to realize that, although at present no scientific data on which to base the existence of EMF-susceptible diabetic subgroup, attenuation of insulin secretion through distortion of the influx of calcium by EMF exposure has been reported under controlled conditions in vitro (Sakurai et al., 2004). Furthermore, data from laboratory studies have suggested that a small fraction of self-reported electrohypersensitive patients do have signs of thyroid dysfunction, liver dysfunction and chronic symptoms that, in combination with somatic conditions, might explain the symptoms reported by EHS (Dahmen et al., 2009). However, although more work is needed current scientific evidence suggests that cognitive behavioral therapy is effective for (some) EHS patients (Rubin et al., 2006).

If HFVT are indeed a biologically active exposure metric missed by the majority of the international scientific community, with 50/60 Hz harmonic sine wave exposure merely being its proxy, this would have important complications for the exposure assessment in future epidemiological studies. Because of its dependency on local factors, the exposure estimates will need to be obtained on an individual level rather than relying on aggregated measures such as distance to transmission lines and transformers (Breckenkamp et al., 2008; Duyan et al., 2008; Maslanyj et al., 2009). Further complications in exposure estimation will be introduced as , even though some data suggest biological systems might be especially susceptible to non-linear ‘resonant’ frequency-amplitude combinations (Zhadin, 2001; Binhi and Savin, 2002), biological effects from HFVT might be caused by the intermittent nature of HFVT compared to the LF-RF frequency harmonic sine waves that carry the transients (Ivancsits et al., 2002; Mathur, 2008). These factors will most likely only further increase the potential for exposure misclassification in studies relating “standard” EMF exposure metrics to health effects (Kheifets et al., 2009; Schuz et al., 2009).

In conclusion, although the studies described in this review are subject to severe problems in the methodology and subsequent analysis and interpretation of the data, they are used as the scientific basis exploited by interest groups to advocate their cause and for commercial enterprises to offer equipment aimed at reducing HFVT ‘dirty electricity’ exposure in the environment. However, whether this is a valid claim and whether these filters are, or are not effective cannot be based on the currently available data. As such, blinded, properly powered and controlled studies with optimized study design and assessment of exposure to a wide EMF frequency range are required before any conclusions can be drawn on high frequency voltage transient exposure as a biological active exposure metric.

References

  1. , and Electromagnetic radiation. Br Med Bull 2003: 68: 157–165.

  2. , , , , and Epidemiology of health effects of radiofrequency exposure. Environ Health Perspect 2004: 112(17): 1741–1754.

  3. , and Electromagnetic fields and public health. Environ Health Perspect 1987: 75: 159–171.

  4. , , , , , , and Correlation between burnout syndrome and psychological and psychosomatic symptoms among teachers. Int Arch Occup Environ Health 2006: 79(3): 199–204.

  5. , and Molecular gyroscopes and biological effects of weak extremely low-frequency magnetic fields. Phys Rev E Stat Nonlinear Soft Matter Phys 2002: 65(5 Part 1): 051912.

  6. BioInitiative Report: A Rationale for a Biologically-based Public Exposure Standard for Electromagnetic Fields (ELF and RF) Online. 2009 May (cited 2009 August 6). Available from : .

  7. , , , and Applicability of an exposure model for the determination of emissions from mobile phone base stations. Radiat Prot Dosimetry 2008: 131(4): 474–481.

  8. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , and The INTERPHONE study: design, epidemiological methods, and description of the study population. Eur J Epidemiol 2007: 22(9): 647–664.

  9. , and Setting prudent public health policy for electromagnetic field exposures. Rev Environ Health 2008: 23(2): 91–117.

  10. , , and Blood laboratory findings in patients suffering from self-perceived electromagnetic hypersensitivity (EHS). Bioelectromagnetics 2009: 30(4): 299–306.

  11. , and Statistics notes: blinding in clinical trials and other studies. BMJ 2000: 321(7259): 504.

  12. , , , , , and Modeling of electromagnetic environment of transmission lines for studying effect of ELF-EMF. Conf Proc IEEE Eng Med Biol Soc 2008: 2008: 1331–1334.

  13. , , and New insights into the placebo and nocebo responses. Neuron 2008: 59(2): 195–206.

  14. , , and EMF and health. Annu Rev Public Health 2005: 26: 165–189.

  15. , , , , , , , , , and Temporal and spatial variability of personal exposure to radio frequency electromagnetic fields. Environ Res 2009: 109(6): 779–785.

  16. Fielding a current idea: exploring the public health impact of electromagnetic radiation. Public Health 2008: 122(2): 113–124.

  17. A Ubiquitous Pollutant. Electronics Research Laboratory, College of Engineering, University of California, Berkeley, Memorandum No. UCB/ERL M00/55. 2000.

  18. Mitigation of Electrical Pollution in the Home. Electronics Research Laboratory, College of Engineering, University of California, Berkeley, Memorandum No. ECB/ERL M02/18. 2002.

  19. A Microsurge Meter for Electrical Pollution Research. Electronics Research Laboratory, College Engineering, University of California, Berkeley, Memorandum No. UCB/ERL M03/03. 2003.

  20. , and Biological effects from electromagnetic field exposure and public exposure standards. Biomed Pharmacother 2008: 62(2): 104–109.

  21. Intensity of electric and magnetic fields from power lines within the business district of 60 Ontario communities. Sci Total Environ 2002: 298(1–3): 183–206.

  22. Electromagnetic hypersensitivity: biological effects of dirty electricity with emphasis on diabetes and multiple sclerosis. Electromagn Biol Med 2006: 25(4): 259–268.

  23. Dirty electricity elevates blood sugar among electrically sensitive diabetics and may explain brittle diabetes. Electromagn Biol Med 2008: 27(2): 135–146.

  24. , , and Teacher and student response to the removal of dirty electricity by the Graham/Stetzer filter at Willow Wood School in Toronto, Canada. In Biological Effects of EMF's Third International Workshop Kos, Greece, 2004, 311–317.

  25. , and Power quality affects teacher wellbeing and student behavior in three Minnesota Schools. Sci Total Environ 2008: 402(2–3): 157–162.

  26. , , and School stressors, psychological complaints and psychosomatic pain. Acta Paediatr 2008: 97(1): 112–117.

  27. , , , and Subjective symptoms, sleeping problems, and cognitive performance in subjects living near mobile phone base stations. Occup Environ Med 2006: 63(5): 307–313.

  28. IARC. Non-Ionizing Radiation, Part 1: Static and Extremely Low-Frequency (ELF) Electric and Magnetic Fields. IARC Press, Lyon, 2009.

  29. IEC. IEC 479-1, Effects of Current on Human Beings and Livestock. Part 1: General Aspects. International Electrotechnical Commission, Geneva, Switzerland, 1994.

  30. Definition, Epidemiology and Management of Electrical Sensitivity. Radiation Protection Division of the Health Protection Agency, Chilton, Didcot, Oxfordshire, HPA-RPD-010, 2005.

  31. , , , , and Induction of DNA strand breaks by intermittent exposure to extremely-low-frequency electromagnetic fields in human diploid fibroblasts. Mutat Res 2002: 519(1–2): 1–13.

  32. Electrohypersensitivity: state-of-the-art of a functional impairment. Electromagn Biol Med 2006: 25(4): 245–258.

  33. , , , , , , , , , and Future needs of occupational epidemiology of extremely low frequency electric and magnetic fields: review and recommendations. Occup Environ Med 2009: 66(2): 72–80.

  34. The controversy about a possible relationship between mobile phone use and cancer. Environ Health Perspect 2009: 117(3): 316–324.

  35. , and Mobile phone base stations-effects on wellbeing and health. Pathophysiology 2009: 16(2–3): 123–135.

  36. , and Electric current perception study challenges electric safety limits. J Med Eng Technol 2002: 26(4): 168–172.

  37. , and Electrosensibility and electromagnetic hypersensitivity. Bioelectromagnetics 2003: 24(6): 387–394.

  38. Hypersensitivity of human subjects to environmental electric and magnetic field exposure: a review of the literature. Environ Health Perspect 2002: 110(Suppl 4): 613–618.

  39. , , , and Circadian variations of blood sugar and plasma insulin levels in man. Diabetologia 1969: 5(6): 397–404.

  40. , , , and Power frequency magnetic fields and risk of childhood leukaemia: misclassification of exposure from the use of the “distance from power line” exposure surrogate. Bioelectromagnetics 2009: 30(3): 183–188.

  41. Effect of chronic intermittent exposure to AM radiofrequency field on responses to various types of noxious stimuli in growing rats. Electromagn Biol Med 2008: 27(3): 266–276.

  42. , and A new electromagnetic exposure metric: high frequency voltage transients associated with increased cancer incidence in teachers in a California school. Am J Ind Med 2008: 51(8): 579–586.

  43. , , and Day-to-day variation of continuously monitored glycaemia: a further measure of diabetic instability. Diabetologia 1972: 8(5): 342–348.

  44. RE: A new electromagnetic exposure metric: high frequency voltage transients associated with increased cancer incidence in teachers in a California school, May 28, 2008; 51:579–586. Am J Ind Med 2009: 52(4): 350–351.

  45. , , , and Screening for type 2 diabetes is feasible, acceptable, but associated with increased short-term anxiety: a randomised controlled trial in British general practice. BMC Public Health 2008: 8: 350.

  46. Radiofrequency electromagnetic field exposure and non-specific symptoms of ill health: a systematic review. Environ Res 2008: 107(2): 277–287.

  47. , and A comparison of important international and national standards for limiting exposure to EMF including the scientific rationale. Health Phys 2007: 92(6): 635–641.

  48. , , and Electromagnetic hypersensitivity: a systematic review of provocation studies. Psychosom Med 2005: 67(2): 224–232.

  49. , , and A systematic review of treatments for electromagnetic hypersensitivity. Psychother Psychosom 2006: 75(1): 12–18.

  50. , , , , and An extremely low frequency magnetic field attenuates insulin secretion from the insulinoma cell line RIN-m. Bioelectromagnetics 2004: 25(3): 160–166.

  51. , , , and Empirical evidence of bias. Dimensions of methodological quality associated with estimates of treatment effects in controlled trials. JAMA 1995: 273(5): 408–412.

  52. , , and Electromagnetic fields and epidemiology: an overview inspired by the fourth course at the International School of Bioelectromagnetics. Bioelectromagnetics 2009: 30(7): 511–524.

  53. , and Role of stress in the etiology and treatment of diabetes mellitus. Psychosom Med 1993: 55(4): 380–393.

  54. , , and Stress and diabetes mellitus. Diabetes Care 1992: 15(10): 1413–1422.

  55. The International Commission for Electromagnetic Safety. Benevento Resolution. In: The Precautionary EMF Approach: Rationale, Legislation and Implementation. International Conference 2006.

  56. The role of causal criteria in causal inferences: Bradford Hill's “aspects of association”. Epidemiol Perspect Innov 2009: 6: 2.

  57. Possible health effects of 50/60 Hz electric and magnetic fields: review of proposed mechanisms. Australas Phys Eng Sci Med 1993: 16(1): 1–21.

  58. Review of Russian literature on biological action of DC and low-frequency AC magnetic fields. Bioelectromagnetics 2001: 22(1): 27–45.

  59. , and Neurobiological mechanisms of placebo responses. Ann NY Acad Sci 2009: 1156: 198–210.

Download references

Acknowledgements

The author would like to thank professor Agius for his insightful suggestions and comments regarding this manuscript.

Author information

Affiliations

  1. Centre for Occupational and Environmental Health, Health Sciences Research Group, School of Community Based Medicine, University of Manchester, Oxford road, Manchester M13 9PL, UK

    • Frank de Vocht

Authors

  1. Search for Frank de Vocht in:

Competing interests

The authors declare no conflict of interest.

Corresponding author

Correspondence to Frank de Vocht.

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/jes.2010.8

Further reading