About Us
Contacts
Company Tour
Products
Non thermal studies
Our Research
Magnetotherapy
Electro Pollution
Endogenous Fields
Media File
The Coghill Challenge
The Natural World
Origins of Electricity
EMF Video Archives
TETRA
Photon Platinum
Review of 2003 NRPB Consultation Document
BIMT Prospectus
How to find us
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| A CASE-CONTROL STUDY MEASURING EXTREMELY LOW FREQUENCY ELECTROMAGNETIC FIELD EXPOSURE AT BEDPLACES | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Abstract Most residential epidemiological studies of ELF EM field exposures have been confined to cancer-related ill health, particularly childhood leukaemia, of which the origin is presently unknown, though a few studies have also implicated asthenia, headache, depressive illness, and other non-cancer states. Myalgic encephalomyelitis (ME) is an increasingly prevalent syndrome of unknown aetiology, with symptoms also including asthenia, persistent headache, and depression. To see whether ME is associated with chronic weak ELF EM field exposure, electric and magnetic field measurements were made for a 12 hour period between 2000-0800hrs. in the bedplaces of 61 patients self- or independently diagnosed as suffering from ME in Troon, Prestwick, and Ayr (three adjacent towns in South West Scotland with populations totalling 76,330), and compared with similar measurements taken in 56 control homes of comparable socio-economic status living in the same towns. In the ME case homes mean magnetic field strengths (87nT; SD 46) did not differ significantly from controls (99nT; SD 265). Mean electric fields however were almost twice as high (22.7V/m; SD 39.2) compared with controls (13.0V/m; SD 15.5), and this difference was significant (p=0.035). Unusually, 57 percent of cases (35/61) lived in homes on or next to corners, near the ends of cul-de-sac streets, or in end-terraces where the electric cable feed entered the block. These were also the sites of highest electric fields. The significance of this finding is not clear. Stratification into exposure bands indicated a dose-response relationship between ME and electric field exposure, with mean exposures above 20V/m yielding an odds ratio of 2.6. This was not the case for magnetic fields(>120nT:1.1OR). Using a cut-off point of 10V/m, the relative risk of ME above this level was 1.71 (95%CI: 1.49-5.87). The sample data was not normally distributed, exhibiting skew, both in control and case groups. The significant difference observed between these means by t-test may therefore not be an appropriate statistical test. Nonparametric testing (Mann & Whitney, which allows for skew) nevertheless also indicated that case electric field exposures were significantly different from controls (p= 0.038), but not magnetic field exposures . These results, though from a relatively small sample, suggest that ME may be associated with weak chronic exposure to ELF electric fields. The study gives support to recent occupational studies also finding over-representation of adult cancer subtypes where occupational electric field exposure was elevated. The curious correlation between corner homes and cul de sac homes with ME cases in this sample, (which could not have arisen through selection bias), needs further investigation, and a larger study is needed to confirm these results. INTRODUCTION Myalgic encephalomyelitis (known as Yuppie Flu, ME, chronic fatigue syndrome or epidemic neurasthenia) is a sometimes seriously debilitating syndrome of variable near-disease symptoms. Several studies have described these (e.g. Poskanzer, 1970: Ramsay, 1989) which can broadly be divided into two groups: effects of impaired oxygen transport (debilitating fatigue, myalgia, giddiness, poor circulation, etc.), and mild immunodeficiency (proneness to viral and other infections, gastrointestinal flora imbalance, lowered lymphocyte count etc.). Results of a questionnaire provided to the cases in this study (see Appendix 6.1) suggest that symptoms appear to be aggravated during periods of low barometric pressure, during October/November and March/April, at high tides, at clock time changes, during winter months, in the evening after dark, (both periods of higher electricity usage), or in meteorological conditions of high positive ion balance such as before storms or at the full moon. Ramsay and Poskanzer both indicate that ME appears to favour women rather than men, with studies typically reporting 80 percent female patients between 20 and 40 years of age. The first reported US outbreak was in Los Angeles in 1934 (Gilliam, 1938), and in the UK in London (Acheson, 1952) and Coventry (1953), though some 60 epidemics, often in residential hostels and mainly affecting younger people, have since occurred in parts of the world as diverse as Iceland and Melbourne (Behan & Behan, 1980). Three major London outbreaks (Royal Free Hospital, 1955; Middlesex Hospital, 1952; and Gt. Ormond St. Hospital, 1970) occurred within 25 metres of major subterranean aquifers (Fleet River conduit, Northumberland St. (West Branch) Main Storm Relief Sewer, and Lamb's Conduit) which are also important generators of charged particles. These outbreaks were reported by Ramsay, 1989, Acheson 1954, and Dillon, Marshall et al., 1974 respectively. In Ayrshire an outbreak in West Kilbride was described (Fegan, Behan et al., 1983) affecting 22 patients between January 1980 and June 1983, mainly female. The most characteristic symptom was extreme exhaustion, particularly after exercise. No clear aetiology or diagnostic marker for ME has yet been isolated. Though according to Ramsay (1989) the disorder is often associated with Coxsackie B viral infection (which was found in 18 of the 22 West Kilbride patients), it is not clear whether the viral insult is post-opportunistic, since no clear infectious pathway has been isolated. Prevalence appears to be increasing, and may now have reached between 1.8 and 6.0 per 1000 of population in industrialized countries according to the UK ME Asssociation. Reporting another Scottish outbreak in Balfron, Stirlingshire (Keighley and Bell, 1983) the authors concluded: "If our results are considered valid, and the figure of incidence in a practice comprising 2500 adults can be extrapolated to the general population, then there are a large number of ill and unhappy patients in the community and it is suspected that many of these are to be found among returning attenders at medical and psychological clinics". Another Scottish study (Caldor and Warnock, 1984) investigated 81 cases between 1978 and 1983 and suggested that in view of the small population involved in their study the disorder may now be primarily endemic, subject to periodic outbreaks of epidemic prevalence. No studies todate have, however, mapped the incidence and prevalence for the UK by area or region. It is possible that a number of the outbreaks (eg Los Angeles, 1934, Coventry 1952, London 1953, 1955, and 1970) also coincide with increased RF transmissions from nearby radio, TV and radar installations, though this is difficult to verify. Certainly the Los Angeles outbreak of 1934 and the London outbreaks of 1952 and 1955 occurred contemporaneously or shortly after important increases in RF traffic in their localities. A syndrome known as radiowave sickness from weak long term exposure to microwaves (described in Silverman, 1980; Serdiuk & Serdiuk, 1989 and elsewhere) bears some similarity to ME symptomatology. Silverman noted that "With few exceptions functional disturbances of the central nervous system have been described as a typical kind of radiowave sickness, the neurasthenic or asthenic syndrome. The symptoms and signs include headache, fatigability, irritability, loss of appetite, sleepiness, sweating, thyroid gland enlargement, difficulties in concentration or memory, depression, and emotional instability. This clinical syndrome is reversible if exposure is discontinued". Six of these symptoms feature in the top dozen reported by our cases (Table 6.8), and all but two of the remainder in at least a quarter of cases. It may be significant that 43 of 60 reported ME outbreaks occurred between 1950 and 1977, when broadcast TV transmissions were becoming established in most industrial countries (Behan & Behan, 1980). No study todate however has investigated this possible association by region or country, and there are many other possible environmental confounders (chemicals, diet, traffic density) to consider. Against this background this study primarily aimed to investigate the residential electromagnetic environment of cases and compare them with controls to test the hypothesis that ELF EM fields are implicated in ME. METHOD AND MATERIALS Case/control selection Local media publicity first announced the commencement of the study in May 1991 (front page, Ayrshire Post) and again in early August 1991 (Ayrshire Post, Advertiser Times Herald) together with posters placed during June and July in health shops and self help groups local to the study area. This was defined as Troon (population 13,329), Ayr (49,409), and Prestwick (13,592), three adjacent Ayrshire towns with a combined population of 76,330). ME sufferers were invited to contact the study organiser and were sent a simple preliminary questionnaire together with a factsheet of practical suggestions for alleviation strategies based on sufferers' experience. No indication of the hypothesis being tested was advised to the respondents at this time. Some 137 respondents returned the initial questionnaire (approximately 1.8 per 1000 of population), though 33 of these respondents lived outside the immediate vicinity of Troon, Prestwick or Ayr, and were therefore excluded. If prevalence estimates of upto 6 per thousand are correct, this would imply that the response was at least 30 percent of sufferers in the reference area. Respondent acquisition was supplemented in February 1992 by Scottish TV coverage which indicated that the research direction would embrace electromagnetic fields and would involve residential measurements of EM field strength. Though the 63 cases finally agreeing to take part had no obvious means of pre-ascertaining the EM fields levels ambient in their homes, they were nevertheless aware of the aim of the study prior to measurement, but the possibility of selection bias on this ground was not considered high, because nearly all had volunteered prior to the February 1992 announcement. Measurements for 2 volunteers' homes did not span the reference 12 hour period through battery failure, and their data has been omitted, leaving 61 cases in the study. The homes of 61 ME cases (85 percent of the 76 still agreeing to take part out of the 107 net telephone respondents) were therefore measured. This sample size has the power to detect a 8.4V/m difference at the 95 percent probability level. Table 1: Analysis of Homes Measured
(in 2 cases results were discarded because of battery failure, leaving 61) a) moved away b) of whom one has moved away, and one was too ill. c) of whom 5 were telephone replies only, and 5 live >10 miles from Troon Controls were selected by the respondents, who were asked to find healthy persons living in the same area and of the same approximate age and sex. They were not specifically asked to supply controls with extremely similar properties (e.g corner homes). Disappointingly not all respondents were able to find controls and some controls included also indicated minor ill health symptoms in their questionnaire . Accordingly some 56 controls were identified by the study organisers for inclusion. Their average results were however similar to other reported residential fields (see discussion). Respondents agreeing to take part in the measurement study were asked to complete a detailed questionnaire identifying 56 symptoms known to be associated with ME and asking for details of place and duration of onset, date, time, and origin of diagnosis, familial health histories etc. These data were analysed to produce a profile of ME symptoms and related details. Data acquisition and dosimetry EM fields and related data was captured using data logging equipment (Delta-T Devices, Burwell, Cambs) to which were fed sensors collecting the three orthogonal components of the magnetic field, the vertical electric field, temperature, humidity and light intensity in lux. The electric field sensor consisted of a di-pole whose ends were some 20 cm apart. The magnetic sensors were wirewound coils of 12 cm diameter built in accordance with IEEE specifications and placed at 90 degree angles to each other. The signal acquisition and amplifying circuitry has been described (Philips, 1992). Several studies report that the best indicator of residential ELF EM fields is the bedroom (e.g. Lee, Yost, 1990, Preece et al.,1996). The sensing systems were therefore placed in the bedroom as near to the bedhead of the case/control as possible, and programmed to take readings every 30 secs. for approximately a 24 hour period. From this data a 12 hrs period was obtained between 2000hrs and 0800hrs., which was used as the basis for comparing exposure means and ranges, and for non-parametric statistical analysis. Appliances, electric blankets, and related EM sources were not altered from their normal use condition. To reduce the amount of data to be analysed the readings taken every 30s. were averaged into 5 mins. means by the loggers. Resulting data were analysed using spreadsheet software (Supercalc5). Because the health state of the respondent was often clearly apparent to the visiting data collectors, no attempt was made to blind the data. Readings were taken during March to June and October to December 1992, being largely discontinued during the summer months due to illness of one of the organisers, bereavement, and holidays. Moreover during the summer electromagnetic fields are seasonally low (Renew, Male, 1990). Controlling for machine error The eight instruments deployed were cross-calibrated before and after the data capture exercise, and any defective sensors were noted on individual datasheets. Tests for machine error showed that the instruments were in good self-agreement: the mean of an 80 percent machine sample of cases, when adjusted for minor machine variances in readings, was 15.97 V/m, SD 3.97 which compared with unadjusted figures of 17.59V/m; SD 7.67 (a difference of 9.2%), and data collected in controls' homes agreed very well with established ranges and norms (see discussion). Tests against independent calibrated instruments (Mersmann 3201 Fieldmeters) also showed agreement within 10 percent (see Appendix 6.4). Controlling for operator bias Since cases were not blinded to the data-collecting personnel, there was a possibility of operator bias in placing the machines advantageously within the bedroom, e.g. near an obvious electric appliance. However in many cases initial recorded levels were not indicative of the overall 24hr. means. Moreover operators were not aware whether or not the home was where onset had occurred, since questionnaires were only collected at the end of data capture. Moreover the requirement to place instruments close to the bed was a restriction on choice of placement. Geographical location From local libraries and other sources large-scale maps were obtained which identified the type of home, its location, and proximity to any important electrical installations. Homes were then categorized into mid-street, end-terrace, cul-de sac end, corner house. Corner house was defined as the house actually on the corner (code C1) or one home in from the corner (code C2), semidetached houses being treated as two homes. Homes more than one away from the beginning or end of any cul de sac were treated as mid-street homes, not as cul-de-sac homes. By ends of cul-de-sac are meant homes at the furthest interior of the cul-de-sac. RESULTS Home types of ME onset 35 cases (57 percent) lived either on corners of streets or one in from the corner, at the furthest interior of cul de sac roads (mainly one home down from the furthest interior), or at the end of terraces, which was an over-representation. There was no way in which the respondents or the study organisers could have pre-selected for over-representation of these home types. 13 cases (21 percent) had moved home since onset, but for 4 of these onset had also occurred in a corner home. The remaining 23 cases (37 percent) lived in mid street. 3 homes (5 percent) were unclassified because large scale maps were not obtainable. Table 2: Home Types
A test for the general frequency of such home types in Troon, Prestwick and Ayr was carried out by inspecting four large scale maps at random and classifying the houses . The home-types obtained in this way were a total of 458 houses of which 298 (65.1%) were mid street, 63 (13.7%) were directly on corners ( designated C1), 69 (15.1%) were one home in from any corner (designated C2) , 7 (1.5%) were end-terrace and 21 (4.6%) were cul de sac homes.. Mid-terrace homes were categorised as mid-street. This suggests that the number of ME cases living in corner- and related home types was 1.6 times the expected level, but the finding is difficult to interpret. The ratio was a little higher when home of onset was also taken into account: since respondents were asked to indicate the onset address (where not the current address), it was noted that onset had in 63 percent of cases originally occurred in corner or cul de sac homes. This elevation of incidence was not carried through into the average number of symptoms, however (see Table 6.3), and those cases in mid street actually reported a higher average symptom number. Again this finding is difficult to interpret. Table 3: Home Type Symptom Frequency
A clue to explain the finding may be that for whatever reason, electric (but not magnetic) fields were on average higher in the home types which were over-represented among cases. (Table 6.4) Table 4: Mean Field Strength found at different hometypes
a) V/m 19.47 SD: 4.90 b) V/m16.72; SD: 4.68 Note: a) and b) exclude one very high reading (285V/m; SD 83) There were no significant differences between cases and controls in respect of temperature and humidity measures: Table 5: Temperature and Humidity
The principal aim of the study was to see if there were any clear differences between field strengths in cases' homes compared with the controls. These results are set out below: Table 6: Measured Field Data a) Measured Magnetic fields (12hrs. measured between 2000-0800hrs)
Measured magnetic field resultants of the three orthogonal components showed no statistically significant difference between cases and controls. b) Measured Electric Fields (12hrs. measured between 2000-0800hrs)
Applying Student's t-test to the above figures for electric fields gives: t (n1 +n2-2) = 22.75 - 13.03 = 17.53/5.43 = 3.23 --------------- (25.2+4.29) 0.5 which with 115 degrees of freedom is significant (p=0.001). Non parametric testing Since the both case and control sample data is skewed, Student's t-test may not be the most appropriate test for significance in this case. A non-parametric test, however, is not affected by skewed distribution. Accordingly a non-parametric test (Mann and Whitney) was used to confirm whether the cases and controls were significantly different. For ranked electric field means the U-statistic was calculated to be 2042, compared with the control sample rank sum for the electric field means of 2970 which is higher, and therefore the difference between cases and control groups is significant. By contrast, the rank sums for magnetic field means were 3379 and 3273 for cases and controls respectively, which are not significantly different (See Appendix 6.5 for table of ranks). Stratification To see if there was any dose-response relationship, cases and controls were stratified into broadly equal groups: Table 7: Relative Risk for Differing Exposure Levels a) ELF 50Hz. Electric fields (Field Strength (V/m, vert.)
These figures suggest a weak dose-response relationship. Taking a cut-off point of 10V/m the relative risk was 1.71 (95%CI: 1.49-5.87). b) ELF 50Hz. Magnetic fields (Field Strength nT,rms)
These figures do not suggest any dose-response relationship. In summary, measured electric fields were significantly higher (by all tests applied) for cases than for controls. Syndrome Onset elsewhere The results above do not take into account that in 13 cases their initial onset of symptoms arose in previous residences. The means and SD for these 13 cases were
which is lower than for either the case or control homes as a whole. Excluding these 13 cases because of the possibility that if ELF EM fields were implicated in the aetiology of the syndrome, then measuring the present residence may be inappropriate, the means and SD for only case homes where the onset originated (n=48) were higher than the overall average Table 8: Fields in ME Case Homes of Onset
It is clear that the means for this subset are higher than the overall case field strengths shown in Table 6.7, which is supportive that electric fields are associated with ME onset. Rank sum test Because comparison of means may not be the appropriate statistical test in view of skewed distribution, the case and control groups were ranked again, this time excluding the 13 cases with homes of onset elsewhere, and the same rank sum test (Mann & Whitney) applied. The result of these rank sums were: 48 Cases: 2925.5 56 Controls: 2534.5 The U-statistic for this dataset was 1329.5 (assuming a normal distribution; on the stricter criterion used for smaller samples the figure is 2248.5), so the difference is also significant when only homes of onset are taken into account (Campbell, 1967). Quality of the healthy controls A further difficulty encountered was that when completing the questionnaire identical to that asked of the cases, some controls also reported ill health symptoms. In order to obtain controls which were "clean" of all ill health symptoms, these controls were then also excluded from the dataset. The mean, median and SD of the "clean" controls were:
These means are a little lower than the values for the overall controls, but not sufficiently so to require a separate analysis. ME symptom analysis Since ME is not a recognised health disorder and its symptomatology somewhat quixotic, cases were asked to indicate from which of a list of 56 symptoms previously identified by ME self-help groups they were suffering at the time of the survey measurement week. The most prevalent of these are listed below: Table 10: Symptom Frequency
Numbers in brackets are from the table of ranked symptoms by Poskanzer, Henderson et al., (1957) for comparison. These compare with 21 symptom types observed in an outbreak (in the Spring of 1956) of 62 cases in Punta Gorda Florida among a 2500 population (Poskanzer, Henderson et al, 1957). They do not however occur in the same order, perhaps because Poskanzer reported symptoms both at onset and in a review five months later. Seasonality, duration, and onset Out of 63 cases 35 (55 percent) reported that their symptoms were least noticeable during the summer months, and when they were away from the home on vacation. Duration of the disorder varied from twenty years to four months, with an average of 4.5 years. Onset was reported as sudden in just under half the cases, gradual in the remainder. There was no correlation between EM field strength and length of time in residence prior to onset, or state of recovery. Furthermore there was no obvious indication that moving to a low field environment had a beneficial effect once onset had occurred. Sex, age bracket, and time since onset are not given here. DISCUSSION Epidemiology Concerns that chronic weak exposure to ELF EM fields may cause ill health, particularly cancer, have been addressed by a number of epidemiological studies. Few of these however have carried out any measurements of the electric component of the EM field, partly through historic accident. The first epidemiological study addressing weak EM field effects (Wertheimer and Leeper, 1979) concentrated only on the magnetic component. The authors reasoned that the electric field from external sources such as transmission or distribution lines was a fairly constant feature and did not penetrate home interiors, whereas the magnetic field (to which buildings and organic tissues are transparent) varied with load, so was more likely to show a dose-response relationship. This view was at variance with eastern bloc thinking, which argued that whereas the magnetic field from powerlines was of the same order of magnitude as that of the earth, to which organisms have become acclimatised by evolution, the electric field from powerlines could reach several thousand V/m, compared with a terrestrial electric field of only 130V/m approximately (Shandala 1988). Only two residential epidemiological studies to date have measured the electric field. Of these one (Savitz, Wachtel et al., 1988) took the simple arithmetic mean of measurements in three locations in the home, the front door, the centre of the child's bedroom, and the parents' bedroom. Spot daytime measurements were taken. Savitz abandoned E-field measurements a third way through (after 129 out of 356 eligible childhood cancer cases, of which 25 had electric fields in excess of 12V/m), because no correlations were emerging between cases and controls in high power use conditions. Nevertheless having measured 40 out of 119 homes where the children had other cancers, he found an elevated odds ratio (1.65; 95% CI: 0.78-3.51). High power use was defined as with all electric appliances on simultaneously. However, Savitz failed to appreciate that alternating electric fields near appliances are present in the associated wiring even though the machine is not operating. The use of spot measurements is also highly questionable. Data collected during this survey shows a dramatic variability, such that a measurement for several minutes may bear no relation to the overall 24 hour exposure. Finally there is also a large variation in field strength even within one room, so that centre room measurements may bear no relationship to those at the bedplace where the child sleeps. Furthermore, Savitz's magnetic field data showed the highest risk ratios not in high but in low use conditions, and extrapolating for this suggested an elevated likelihood of risk from electric field exposures. Moreover Savitz took measurements only at room centres, and deliberately not near beds or appliances, arguably the principal sites of domestic exposure (Lee, Yost et al, 1990), despite a large variation in field strength not only between different rooms but also within rooms (Mader, Barrow et al., 1990). Savitz conceded that the possible ill health effects of long term electric fields were not addressed by his study design. This study therefore made little if any contribution to knowledge concerning electric field exposure effects. In a later study, London, Thomas et al, (1991) took spot measurements in the centre of 136 leukaemic childrens' bedrooms but only for brief time periods (5-10 minutes: mean 7.48V/m; SD 9.84), and in the bedrooms of 108 controls (mean 7.98 V/m; SD 12.27). Their study was therefore prone to the same fundamental flaws as that of Savitz. Moreover the authors switched from IREQ to EMDEX machines during the study, which prevented them from collecting further electric field data, since EMDEX machines do not capture electric field data without a special probe which had not at that time been developed. In the UK the prevalence of cheap nighttime electricity would invalidate such spot measurements. Overall then, the entire Western epidemiological evidence of residential hazard from electric field effects is based on measurements in the order of 80 hrs in total from two studies. (By contrast, the electric field data in this study collected for 12 hrs from 61 sites totals 732hrs approximately, and the leukaemia study adds 672hrs to this figure). A few occupational studies have investigated the effects of electric fields, but again not in any intensive way. Deadman, Camus et al, (1988) found that a control group of office workers were exposed on average to electric fields of about 4.9V/m., measured for one week, and that mean daytime occupational exposures ranged between 62 and 419 V/m. This study did not however consider health issues, so could say nothing about the effects of exposure. A more recent and very comprehensive study (Theriault, Goldberg et al.,1996) jointly sponsored by Ontario Hydro, Electricite de France (EDF) and Quebec Hydro (QH), under the direction of McGill University reported that electric fields indicated relative risk of leukaemia as elevated as 11.2 times, and that current transients (which also give rise to high electric fields) were also associated with elevated incidence of cancers, particularly leukaemias. Unfortunately the sponsors have disallowed any further publication or research from this dataset, and only a summary has been published (Amer. J. Epidemiol., July 1996). Nevertheless the original study (93pp. plus Appendices) reported an increased cancer risk among electricity linesmen at QH (but not at EDF): odds ratio 3.29 (CI: 1.23-8.79) though this was based on only 11 cases and 17 controls. The Odds ratio for QH workers exposed above a mean of 12V/m was 2.44 (CI: 1.01-5.89) based on 20 cases and 46 controls, and was just significant. These results compare well with the results reported here, except that the McGill cases' exposure duration was much longer. In response to such concerns over possible ELF health hazards several authorities (eg SWEDAC, NY Purchasing Authority) have adopted a policy of prudent avoidance, setting guidelines for ELF electric fields low as 25 V/m for exposure from VDUs, though with little supporting epidemiological evidence of hazard. Recent studies (Mader, Barrow, et al. 1990; Barnes, Wachtel et al., 1989) have also begun to characterize residential EM fields more adequately, and reveal that an average exposure (low configuration code) might be in the order of 9.31V/m (SD: 5.14) in 60 Hz AC systems but could reach well over 20 V/m in higher exposure conditions, and thus in excess of the SWEDAC and NY recommended limits. These findings are much in line with those of this study. Laboratory Evidence Much laboratory evidence implicates electric fields as the active parameter in affecting biological processes, particularly affecting erythrocyte orientation and morphology (Takashima & Asakura 1983; Murayama, 1984; etc) and calcium efflux (Bawin, Adey et al. 1976; Blackman et al., 1979; Blackman, House et al., 1988). A recent meta-analysis (Morris, Kimball, et al., 1989) found adverse haematological effects following electric field exposure in a number of endpoints from 40 animal experiments. These included a decrease in lymphocytes, and altered haemoglobin indices. More specifically Lyle, Ayotte et al. reported a 25 percent inhibition of cytotoxicity in murine T-lymphocytes exposed to powerline frequency electric fields for only 48 hours. Liburdy in 1992 reported influx of calcium in direct response to electric field exposure of lymphocytes in distinction to magnetic fields, suggesting an effect on circulating NK lymphocyte motility, indispensable for tumour immunity. In ME a diminution of these cell types is often observed: a subset of 11 acute ME cases in a sample of 50 (Behan, Behan et al., 1985) all showed a reduction in the number of suppressor/cytotoxic T lymphocytes. An attempt to relate these and related findings to the symptomatology reported by ME sufferers is made below. Possible mechanisms of interaction between electric fields and ME cases There is some evidence that electric fields may influence blood cell morphology. Several studies report that ME patients' erythrocytes exhibit significant conformational changes compared with controls, with elevated stomatocyte-spherocyte levels and consequently poor rheology (e.g. Simpson, Shand et al., 1986; Mukherjee, Smith et al.,1987; Simpson., 1989), though this has not always been replicated (Lloyd, Wakefield et al.,1989). Stomatocytic and echinocytic transformations are known to be dose-dependent and reversible, and can be caused by a broad spectrum of cationic or positively-charged drugs (eg colchicine, which is effective on lower back pain and rheumatoid conditions), or extracellular media with low pH values (Sheetz, Painter et al, 1976). By contrast echinocytic shapes are known to be induced by extra-cellular media with high pH values, or by negatively charged drugs (eg TNP). Altered pH values, or the application of opposing drug groups, normalizes erythrocytes into their usual discocytic shape (Reinhart & Chien, 1986). It has been observed that erythrocyte morphology is finally determined by the relative difference in potential between the intra- and extra-membrane potential (Glaser, 1989;Glaser & Donath, 1992) and therefore influenced by exogenous electric fields. Nevertheless, this kind of study is very prone to difficulties, since even small changes in pH or temperature can dramatically affect the results. Such dysmorphologies might explain the poor blood rheology and impaired oxygen transport seen in ME cases, but not the continued post-onset duration despite moving to subsequent low field environments (see Table 6.8), nor the mild immune incompetence, where other epidemiological studies suggest an elevated incidence of childhood leukaemia among populations chronically exposed to ELF EM fields above 200nT or surrogate measures of that value such as wire codes (Savitz, Wachtel et al., 1988; Feychting, Ahlbom et al, 1992). Moreover non-migraine headache and emotional lability (depression) are two symptoms associated with EM field exposure also common in ME cases (Haysom, Dowson, 1990; Poole, Kavet et al., 1992), the latter of which at least is difficult to explain by a dysmorphology hypothesis. 6.48 Continued duration of symptoms despite moving home seen in Table 6.8 implies that if biological effects from ELF EM field exposure are associated with ME, then the effects are not reversed with cessation after chronic exposure. One speculative possibility is that the natural blood dia- and para-magnetism observed in venous and arterial blood (Pauling and Coryell 1936) may have been permanently affected by chronic weak exposure. This hypothesis is given support by Neumann and Katchalsky, 1972, who showed that long-lived conformational changes could be induced by electric impulses (20kV/cm for several microseconds) in bio-polymers. Similarly the ferrohaemic assemblies in haemoglobin may be regarded as isolated polyions each surrounded by its own ionic atmosphere of four (or five) unpaired electrons. Altering their ionic state may in turn inhibit their affinity for haem-haem interaction in accepting oxygen molecules, hence their efficiency in oxygen transport. As the authors concluded: "Any polyvalently charged system exposed to these powerful impulses could undergo far-reaching and long-lived conformational changes similar to those discussed". Lowered oxygen bio-availablity with its attendant effects on ATP synthesis would be a plausible expalnation covering most symptoms reported in ME. Even so, the fields measured in the present study were not anything like so high, and there was no facility for measuring transients. 6.49 Electric fields do not penetrate organic tissues directly. However, the conformational response of erythrocytes to positively-charged drugs may possibly be induced by influx into the bloodstream of positively-charged ions arising from external EM field exposures by inhalation or intake of water-borne charged particles. Alternatively moving magnetic fields can induce an electric field which may in turn be implicated in the mechanism of interaction, though there is little support for this possibility from the results presented here. Sources of residential fields The source of ELF EM fields is not confined to external powerlines but implicates the internal wiring configuration and number of electric appliances in use within the home, and the method by which the earthing is achieved. London, Thomas et al., for example found significant odds ratios for childhood leukaemia with electric blankets, hairdryers, curling irons, electric dial clocks, and monochrome TV use, and that cases were more likely than controls to report use of several appliances. Barnes, Wachtel et al., (1989) commented that whereas powerline fields might fall off as the square (or greater) of the distance, ground loop currents would attenuate only with the simple reciprocal. Finally, since all these influences are in the far field, no precise calculations may be made. Barnes, Wachtel et al., (1989) pointed out that a relatively small ground loop may have the same effect as a much larger distribution line current. Moreover, in reality cases are more complicated by the geometry and the fact that the current from the unbalanced load of several homes may flow down the street and back through the plumbing of the house with the lowest resistance connection to the powerline neutral. The authors also indicate that seasonality and time of day can each independently double the field measurements. All these points underline the need for a careful assessment of the fields actually obtaining at the bedplace, where the chronic exposure arguably occurs in residential exposures. National Grid Plc (a privatized UK electricity transmission utility) admit that fields of several mT can develop from unbalanced ground return currents near homes (Renew, Male et al., 1990). Such ground return currents are more likely to return via waterpipes than through the higher resistance of the soil. Wertheimer & Leeper (1979), Kaune & Stevens (1987), Mader, Barrow et al., (1990) and others describe how such loops could be a major domestic magnetic field source, but say nothing of any implications for the electric field. Limitations of this study It proved difficult within the time and cost framework available to collect an adequate number of "clean" controls, though ideally a ratio of three controls for each case would have been preferred. This is even more disappointing in view of the large difference in means between the 12hrs exposure figures for ME controls 13V/m (Scotland) and the leukaemia controls 7.3V/m (South England) which raises a number of questions: for example are there important regional differences in average exposure to ELF EM fields? The overall means for the two control groups for the 24hrs period, though not reported in full here, were however not very dissimilar (7.3V/m and 7.5V/m), and it may be that night-time exposure in Scotland (a generally cooler climate) is higher than the national average. A second limitation was the possible lack of precision in placing the instrumentation in the bedplace. With hindsight a much more careful approach should have been adopted, but the surprisingly different electric field strength possible, even within one room, was not recognised at the outset in the study design. Future studies should take care to place the measuring instruments in exactly the same location (e.g. on the pillow). If electric fields turn out to be an important factor governing health, then in many cases simply moving the bedplace would avoid sleeping in abnormally high fields. Another further difficulty arises because myalgic encephalomyelitis is not a wholly-accepted medical ill health condition, and its symptomatology is quixotic, leading to possible misclassification. This is not so important as might at first appear: what is being investigated here is whether electric ormagnetic fields have adverse health effects, not whether they are responsible for a specific disorder, and if so what might be the plausible underlying mechanisms of interaction. Agreement with other studies Despite the inadequacies indicated above regarding controls, the field strengths reported here from control homes agree well with Barnes, Wachtel et al., (1989) allowing for the difference between 50Hz and 60Hz frequency: Table 11: Residential Measured Electric Fields (from Barnes, Wachtel et al.,1989)
*moved since onset The lower magnetic fields observed may reflect comparatively lower use of domestic appliances in the UK compared with the US. Dlugosz, Byers et al., (1989) measured ambient 60 Hz magnetic fields on the corners of a Buffalo, NY urban neighbourhood and found the range depended on whether powerlines or service drops were present. They did not measure electric fields. Table 12: Magnetic fields near corners of streets. (from Dlugosz, Byers, et al., 1989)
In the absence of influences from transmission lines or service drops the B-field range fell to 23-46nT (n=11-30; SD 21-28) suggesting that such configurations may significantly elevate field levels in corner homes, and this view is supported by two other studies. Gauger, 1985 and Caola, Deno et al. (1990) reported a range of 9-248nT in the ambient magnetic field in 24 Toronto homes, and a range of 8-70nT between rooms in one home, but noted that readings over a 7-day period showed little ambient field variation. Taken together, these studies seem in fair agreement with our measurements of controls and cases, allowing for the additional burden of internal loads from home appliances. Barnes, Wachtel et al. (1989) commented that the electric field was sensitive to changes in the location of people, furniture, and the operation of appliances. Unlike the magnetic field it is also present whether the appliance is in use or not, so a home with many appliances plugged into the circuitry, even if not in use, could generate a higher mean E-field strength than a home with few appliances or circuits, because (unlike the higher frequencies) there is no fixed relationship between the electric and the magnetic field at ELF frequencies. (In parenthesis, ME is sometimes described in the UK as Yuppie Flu, - a "Yuppie" being someone of young and upwardly mobile carreer prospects - because it sometimes appears to affect younger people with higher incomes. Such people tend to own more electric and electronic appliances). CONCLUSIONS Against a background of much uncertainty as to the active biological affector in EM field insult, and a paucity of data characterizing mean electric fields to be expected in homes, these results confirm other recent epidemiological indications (e.g. Theriault, Goldberg et al., 1996) that weak electric (as opposed to magnetic) fields correlate with reported ill health effects. The dose-response relationship reported here gives some support to the statistically significant differences between cases and controls leading to this conclusion, but a bigger sample, especially a larger number of controls, which is a weakness of this study, would have improved its statistical power, and the results need to be confirmed in a larger study. The observed over-representation of ME in corner houses may be worthy of further study, provided that careful attention is made to eliminating potential confounders such as traffic density, wind patterns, and other features altered by such locations Acknowledgements I am indebted to Graham Trantor, John Shankland and Gus Guthrie for data collection, and to SEARCH ME who provided financial assistance and administrative help during its compilation. Kyle and Carrick sewage and planning departments kindly provided geographical details. Roger Coghill, MA, 1996 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
