Laboratory studies on humans have certain advantages. They focus directly on the species of main public interest, thus avoiding the problems of extrapolation from data obtained in other species; furthermore, even negative results can be of immediate use in addressing public concern. Such studies can also be used to directly evaluate the effects of exposure on higher-order cognitive functions important in daily life, such as memory, attention, and information processing, as well as their underlying electrophysiological and neurochemical substrates. Controlled laboratory testing with human volunteers can help to define dose metrics and response categories for epidemiological studies, and can guide animal research to areas in which more invasive mechanistic investigations might be valuable. Like animal studies, human laboratory studies allow separate testing of the effects of electric, magnetic, and combined fields to determine whether any effects found are related to specific characteristics of the exposure situation or to their interactions.
The effects of experimental human exposure to EMF are derived from three major research initiatives and from efforts in individual laboratories. These include a long series of studies of utility workers begun in the 1960s in the former USSR (Asanova & Rakov, 1966), the human laboratory research conducted in the 1970s in Germany (Hauf & Wiesinger, 1973; Silny, 1986), and the human laboratory research program started in 1982 at the Midwest Research Institute in the USA (Graham et al., 1990). Dedicated facilities for human exposure testing have been designed and constructed in Australia (Wood et al., 1997), Canada (Blondin et al., 1996), England (Stollery, 1986), France (Selmaoui et al., 1996b), Germany (Hauf & Wiesinger, 1973), New Zealand (Podd et al., 1995), the Russian Federation (Lyskov et al., 1993b), and the USA (Cohen et al., 1992; Doynov et al., 1998). Research with human volunteers is currently under way in many of these facilities.
Graham and Cohen (Graham & Cohen, 1985) performed a laboratory-based study of the perception of 60 Hz electric fields (0-15 kV/m) and of magnetic fields (0-40 µT) by 10 men and 10 women aged 21-35. The men and women had similar sensitivity. The threshold of 90% of the group was ³ 9 kV/m. Perception improved when the field onset was abrupt and when the volunteer changed body position in the field. Perception of initial field onset ceased after about 20 min of continuous exposure but was immediately re-established by body movements within the field. No evidence was found for perception of magnetic field ² 40 µT, and the presence or absence of the magnetic field did not influence perception of the electric field. These results replicate two earlier, well-controlled laboratory studies of perception of magnetic fields by human volunteers: Schmitt and Tucker (Schmitt & Tucker, 1978) and Tucker and Schmitt (Tucker & Schmitt, 1978) tested the ability of 200 volunteers to perceive 60 Hz magnetic fields over the range 0.7-1.5 mT and reported little evidence for perception.
The magnetophosphene phenomenon is also of interest in connection with the 'melatonin hypothesis'. Exposure to light at night is known to reduce circulating concentrations of the pineal hormone melatonin. Wood et al. (Wood et al., 1997) suggested that magnetic field-induced phosphenes may play a role similar to light at night and thus represent a possible biological mechanism to account for the reported reduction in melatonin during nocturnal exposure to magnetic fields. [The authors did not evaluate this possibility, and it does not seem to be a particularly promising avenue of research.] Induction of magnetophosphenes in humans is sharply frequency dependent and requires magnetic flux densities more than 10 000 times higher than those usually encountered in residential settings. Maximum sensitivity occurs between 20 and 30 Hz; higher thresholds have been observed for both lower and higher frequencies (Lovsund et al., 1979). The minimum flux density required to induce detectable magnetophosphenes is > 3-5 mT at 20 Hz (Silny, 1986). At 60 Hz, the threshold at the retina for perception of magnetophosphenes is at least an order of magnitude greater than at 20 Hz.
In the four studies summarized in the first half of Table 4.41, EEG activity from multiple standardized scalp locations was tested in paradigms that allowed comparison of field-exposed and unexposed conditions. The raw data were submitted to fast Fourier transform spectral analysis to obtain information about possible field-induced alterations in the absolute and relative power of the dominant brain frequency, the total amount of energy being generated, and the distribution of energy in different frequency bands and over various regions of the head. Gamberale et al. (Gamberale et al., 1989), for example, performed standard clinical EEG examinations on experienced utility company linemen before and after performance of their usual duties under typical exposure conditions on one day and under simulated conditions on a second day. They found that acute exposure over one day to fields associated with a 400 kV transmission line had no negative effects on the nervous system in this group of healthy men.
Lyskov et al. (Lyskov et al., 1993a; Lyskov et al., 1993b) performed two double-blind laboratory-based studies of the effects of continuous and intermittent (1 s on, 1 s off) exposure to magnetic fields on EEG activity in 17 male and 17 female volunteers. A Helmholtz-type coil device was used to produce a horizontal magnetic field (45 Hz, 1.26 mT, 25% uniformity) around the head of each volunteer, for 60 min in study 1 and for 15 min in study 2. EEG activity measured immediately before and after exposure and sham-exposure periods indicated that effects were predominantly associated with intermittent exposure. The 60-min exposure was associated with decreases in absolute and relative power in the lower EEG frequency bands (delta, 1-4 Hz; theta, 4-7 Hz) and increases in power in the higher frequency bands (alpha, 8-12 Hz; beta, 13-20 Hz). Field exposure had significantly (p < 0.05) different effects from sham exposure in 14 of 36 measures. This result was much greater than the two significant differences to be expected purely by chance alone. Intermittent exposure for 15 min under the same test conditions also resulted in significant (p < 0.05) effects on EEG activity. The most prominent change was an increase in both absolute and relative power in the alpha band; absolute power in the beta band also showed a non-significant increase. Short-duration exposure, however, had no effect on the lower frequency bands. The authors suggested that this pattern of EEG changes is indicative of the combined effects of relaxation and activation caused by the magnetic field on different structural and functional systems of the brain.
Since 1991, Bell and Marino and their coworkers have examined the effects of magnetic fields on human EEG activity. The study of Bell et al. (Bell et al., 1992) summarized in Table 4.41 is representative of the approach taken by these investigators. A Helmholtz-type coil device is used to repeatedly expose the head and chest of neurological patients or volunteers from the general population to DC fields or to AC magnetic fields of varying frequency (² 60 Hz) and intensity (² 100 µT), either alone or in combination, in a test series of second-duration exposure epochs interspersed with no-exposure epochs. EEG activity during exposure epochs is compared with that in the preceding no-exposure epochs by non-traditional analytical techniques, in which the EEG spectrum is divided into 0.05 Hz bands and a field effect is defined as a change, either positive or negative, in the power coefficients in one or more bands. [The findings across these studies are ambiguous and not informative with regard to dose-response relationships. A number of potential sources of error and artifact are associated with this type of test and analysis, including fatigue, lapses of attention, body movements, visual-induced artifacts, and transients. The mixed use of neurological patients and normal volunteers in the same experiment creates additional ambiguity.]
The lower half of Table 4.41 summarizes the results of six recent laboratory studies designed to evaluate the effects of magnetic fields on ERP components. Graham et al. (Graham et al., 1995) measured short-latency (< 50 ms) ERP components to visual, auditory, and somatosensory stimuli in a double-blind study of exposure to a 60 Hz 10 µT, 20 µT field of 36 healthy young men and women. ERP measures of neural conduction time during exposure and under sham-exposure conditions were not statistically significantly different in any of the three sensory modalities. In a study with similar measures, described in section 4.4.3, Dowman et al. (Dowman et al., 1989) obtained similar results in monkeys. [These results suggest that exposure to combined EMF or to magnetic fields alone at levels higher than those found in most residences has no effect on neural conduction time in peripheral or central sensory afferent pathways.]
Four double-blind cross-over studies (Table 4.41) (Cook et al., 1992; Graham et al., 1987; Graham et al., 1990; Graham et al., 1994) showed that exposure of human volunteers for up to 6 h to combined EMF at intensities as high as 12 kV/m and 20 µT per axis had no effect on longer-latency components of the visual ERP. These results replicate those of an earlier report by (Silny, 1986) who used 50 Hz magnetic fields up to 5 mT. In contrast, replicable effects of whole-body exposure to combined electric and circularly polarized 60 Hz magnetic fields on the auditory ERP have been reported in three laboratory studies (Cook et al., 1992; Graham et al., 1987; Graham et al., 1994). The field levels in these studies ranged from 6 kV/m and 10 µT to 9 kV/m and 20 µT. [All of the studies conducted in this laboratory involved circularly polarized magnetic fields, and flux is reported as per axis (e.g. a 20 µT per axis circularly polarized field gives a resultant field of 28.3 µT).] Auditory ERP measures were obtained while volunteers performed a widely used neurophysiological target detection task, the 'oddball task' (Donchin, 1984; Harbin, 1985; Otto et al., 1985). In comparison with sham exposure, field exposure was associated with a significant (p < 0.05) increase in the amplitude of P300, the primary cognitive component of the auditory ERP waveform. In the context of the oddball task, increases in amplitude indicate possible interference in the normal physiological processes involved in distinguishing relevant from irrelevant stimuli (i.e. increased susceptibility to distraction). [These studies of auditory ERP activity represent within-laboratory replications.] Lyskov et al. (Lyskov et al., 1993a) reported somewhat contradictory results: the P300 component of the auditory ERP when measured in response to the presentation of simple tone stimuli was not affected by 15 min of cephalic exposure to a 50 Hz 1.26-mT horizontal magnetic field.
[The studies of the effects of exposure to EMF on human EEG activity differ in a number of important respects, e.g. exposure duration, strength, orientation, polarity, whole-body vs. cephalic exposure, physiological recording parameters, and analytical techniques. Taken as a whole, however, they suggest that EEG measures are affected by exposure to electric and/or magnetic fields; however, the reported effects do not appear to result from alterations in neural conduction time in major sensory modalities and may well be related to shifts in physiological arousal and attention during testing. The biological magnitude (< 10 %) of the changes observed is well within normal ranges for the EEG parameters assessed. The data are insufficient to suggest a health effect.]
Two recent reports (see Table 4.42) suggest that exposure to magnetic fields in the ELF range also disrupt objective EEG measures of night sleep in human volunteers. Akerstedt et al.(Akerstedt et al., 1997a; Akerstedt et al., 1997b) performed a double-blind laboratory study to evaluate the effects of all-night exposure to a 50 Hz, 1 µT horizontal magnetic field on EEG sleep parameters and endocrine or hormonal measures in 18 men and women, 24-49 years old. Each subject slept in the exposure facility for two nights. The first night was considered an adaptation session in which the subjects became used to sleeping in the laboratory while wearing physiological recording sensors. On the second night, half of the subjects at random were sham exposed, and half were continuously exposed to the magnetic field. Sleep EEGs were recorded throughout the test nights, and blood samples were obtained six times per night from an indwelling catheter. In comparison with sham-exposure conditions, field exposure was associated with a significant (p < 0.01) reduction in the duration of slow-wave sleep (sleep stages III and IV), which is often considered to be the restorative portion of night sleep. The nights of field exposure nights were also associated with reduced total sleep time, reduced sleep efficiency, and reduced rapid-eye-movement (REM) sleep; these effects did not reach statistical significance.
Graham and Cook (Graham & Cook, 1998) performed a double-blind, laboratory-based study to evaluate the effects of exposure to magnetic fields on EEG measures of nocturnal sleep patterns in 24 healthy men aged 18-35. Each man slept in the exposure facility from 23:00 to 07:00 h for three nights. The first two nights were adaptation sessions during which the subjects became used to sleeping in the laboratory. On the third night, eight men were sham exposed; seven men were continuously exposed to a 60 Hz, 20 µT per axis (resultant, 28.3 µT) circularly polarized magnetic field; and nine men were exposed intermittently (1 h on, 1 h off) to the same field. The sleep EEG was recorded throughout the night. Physiological data collection, sleep scoring, and statistical analyses were all performed in a blinded fashion. The results seen with continuous exposure did not differ from those in sham-exposed controls for any measure, while intermittent exposure resulted in a significant distortion of nocturnal sleep in six of the 10 measures evaluated. Intermittent exposure was associated with less total sleep time (p = 0.003), reduced sleep efficiency (p = 0.003), increased time in stage 2 sleep (p = 0.009), decreased REM sleep (p = 0.001), and increased REM latency (p = 0.04). Subjects exposed intermittently to the field also reported sleeping less well (p = 0.001) and feeling less rested in the morning (p = 0.03) than subjects in the other two groups.
[The exposure and test procedures were quite different in these two studies, and the different results obtained may be due to those differences. The findings of Akerstedt et al. (1997a) are relevant to levels of residential exposure. However, the possible confounding effect of serial blood sampling on EEG measures of nocturnal sleep should be examined further. The exposure in the study of Graham and Cook are more relevant to occupational exposure in the utility industry than to residential exposures, but they observed little slow-wave sleep, suggesting that the volunteers were not completely adapted to the laboratory environment. It will be important to determine whether the findings of these preliminary studies are replicable, to extend them to women and to older individuals, and to ascertain whether daytime exposure also alters sleep architecture. The available data on the effects of ELF fields on sleep are insufficient to suggest a health risk; however, poor sleep quality can have a detrimental effect on worker safety and performance, and modification of REM sleep has been associated with decrements in memory function and learning processes.]
The ability to perceive the world around us, to process information quickly and accurately, and to maintain appropriate levels of attention and arousal are all significant human functions. Even subtle changes in such functions can have important consequences. Performance depends on intact, functioning physiological systems, and changes in performance in toxicological or exposure studies can provide important information about underlying neuronal dysfunction. The most frequently used performance measures in research on EMF with human volunteers are reaction time, vigilance or sustained attention, memory function, and tasks involving time perception and information processing. Many earlier studies were reviewed by Carstensen (Carstensen, 1987).
Table 4.43 summarizes the results of 10 recent studies. Even a cursory review of the Table indicates that the results for performance have been very mixed. For example, two studies found improvements in a reaction-time task, two found decrements in reaction time, and five found no effect. A similar lack of consistency is evident in the results of studies of more complex performance, partly because different investigators seldom use the same task and the range of exposure conditions is limited. The mixed results may also be due to differences in protocol between laboratories and to variations in volunteer motivation. More standardization in this area of research is needed.
[Human performance of many types of task appears to be unaffected by exposure to relatively high electric and/or magnetic fields. When effects are observed, however, the magnitude of the alteration is generally in the order of 10% or less, and little reliable evidence exists for a consistent dose-response relationship. Thus, there is insufficient evidence at this time to indicate that daytime exposure to ELF EMF at occupational levels presents a health risk. The extent to which performance is disrupted by fields of low intensity similar to those found in offices and residences, however, has not been well evaluated. Similarly, it is not known if nocturnal field exposure has detrimental effects on daytime performance.]
Sazonova (Sazonova, 1967) examined groups of switchyard workers in the former USSR, who differed in the duration and intensity of their exposure to 50 Hz fields. The pulse rates of people in the group with an average exposure for > 5 h/d to 12-16 kV/m were lower by 2-5 beats/min at the end of the day, although they had been equivalent at the start of the day, and this difference was found even after exercise (p < 0.05). Early studies of effects on the heart rate in human volunteers were reviewed by Carstensen (Carstensen, 1987) and Cook (Cook et al., 1992). A number of the earlier studies were published as meeting abstracts or in non-English-language journals.
Table 4.44 presents the results of six recent studies of this phenomenon. Replicable field-related slowing of the heart rate was reported in four laboratory-based studies with double-blind cross-over experimental designs (Cook et al., 1992; Graham et al., 1987; Graham et al., 1990; Graham et al., 1994). Negative results have also been found, however. The biological mechanism underlying the phenomenon is unknown, and the magnitude of the observed effects across studies is small (< 10 % change from the mean).
Korpinen et al. (Korpinen et al., 1993) see table 4.44 used ambulatory recording techniques to perform an extensive series of studies of the effects of occupational exposure to EMF on heart rate. No field-related changes in mean heart rate were found as a result of exposure to the 50 Hz fields directly under power transmission lines ranging in intensity from 110 to 400 kV. [There are insufficient data to suggest a health risk.]
Heart-rate variability (HRV) results from the action of neuronal and cardiovascular reflexes, including those involved in the control of temperature, blood pressure, and respiration. Quantitative spectral analyses of alterations in HRV with digital Fourier transform provide useful indicators of beat-to-beat variations in sympathetic and parasympathetic nerve activity in vivo. This type of variability is not consciously perceived, and it should not be confused with heart-rate reactivity, which is the slowing or speeding of the heart rate in direct response to perceived situational or personal stimuli (e.g. exercise, anxiety, and relaxation). Certain alterations in HRV have recognized prognostic value for coronary artery disease (Hayano et al., 1990; Liao et al., 1997), post-infarction risk (Bigger et al., 1993; Kleiger et al., 1987), diabetic autonomic neuropathy (Bernardi et al., 1992), and systemic hypertension (Huikuri et al., 1996; Liao et al., 1996). Data from large longitudinal human studies in the USA and other countries have indicated that quantitative spectral assessment of HRV in healthy and diseased populations offers prognostic information for the risk of sudden cardiovascular death and death from all causes, beyond that provided by the evaluation of traditional risk factors. In the Framingham Study (Tsuji et al., 1996; Tsuji et al., 1994), HRV was examined in a middle-aged and elderly cohort after exclusion of subjects with cardiovascular disease or medications that could affect HRV. Reduced HRV was significantly associated with subsequent sudden cardiovascular death, even after all previous risk factors (e.g. blood pressure and cholesterol concentrations) had been taken into account. Similar results were obtained in a 25-year prospective study of a middle-aged and elderly cohort (Dekker et al., 1997); these results complement more recent concordant findings (Liao et al., 1997).
Sastre et al. (Sastre et al., 1998); see Table 4.44, described three recent double-blind studies in which sufficient data were collected to examine HRV in detail in 77 volunteers as they slept through the night (23:00-07:00 h) in the laboratory while being exposed to intermittent (1 h on, 1 h off) or continuous 60 Hz magnetic fields at intensities of 1 or 20 µT. In study 1, intermittent exposure to 20 µT reduced HRV in the spectral band associated with the neural control of thermoregulation and blood pressure (known as the 'low band') in comparisons with sham-exposure conditions and field exposure to 1 µT (p = 0.03). In study 2, each of the 23 volunteers served as his own control. They were sham exposed in one session and exposed to the 20 µT field in a second session. In comparison with sham-exposure conditions, intermittent field exposure was again associated with a reduction in low-band power (p = 0.02) and also with a significant (p = 0.008) increase in the power in the spectral band associated with natural respiration-induced alterations in heart rate (known as the 'high band'). In a third study performed to determine the effects of continuous rather than intermittent exposure, no significant effects on HRV were found. [Taken together, these findings are consistent with the hypothesis that intermittency of exposure to magnetic fields is an important parameter in the human cardiac responses that result from those exposures.]
In the studies of Sastre et al., low-band power was reduced by approximately 17% by nocturnal exposure to an intermittent magnetic field of 20 µT, whereas the biological magnitude of the reduction in low-band power typically seen in clinical studies is 20-40%. In the studies of Sastre et al., none of the volunteers reported any cardiovascular difficulties associated with the night-time exposures; secondly, the changes seen to date as a function of intermittent exposure are similar but not identical to those reported as predictive of cardiovascular morbidity and mortality. In the clinical studies, a higher cardiovascular risk was associated with a reduction in low-band power coupled with modest reductions or no change in high-band power (Barron & Lesh, 1996; Bernardi et al., 1992; Bigger et al., 1993; Hayano et al., 1990; Huikuri et al., 1996; Kleiger et al., 1987). The pattern seen with field exposure is a reduction in low-band power with an increase in high-band power. This pattern has, to our knowledge, been reported only for changes in HRV associated with stage 2 sleep (Vaughn et al., 1995).
Extrapolation of the laboratory data described by Sastre et al. (Sastre et al., 1998) in combination with the results of clinical studies of HRV lead to the biological hypothesis that chronic occupational exposure to intermittent power-frequency magnetic fields may affect specific types of cardiovascular disease by altering cardiac autonomic control. Under this hypothesis, such exposures would be linked to increased risks for arrhythmia-related deaths and deaths due to acute myocardial infarct. No increase in risk would be predicted for deaths due to cardiovascular events that are the end-product of processes that develop independently form autonomic nervous input over extended periods of time (i.e. atherosclerosis and chronic coronary heart disease). Savitz et al. (Savitz et al., 1998c) recently reported results that support both the positive and the negative predictions derived from this hypothesis, in a cohort analysis of 139 903 male electric utility workers. The study is described in detail in section 4.5.1.
[Primarily because of the small number of studies reported at this time, there is insufficient evidence of a health risk.]
Stevens et al. (Stevens et al., 1997) proposed that exposure-related suppression of nocturnal melatonin might provide a plausible biological mechanism to account for some of the epidemiological reports linking occupational or residential exposure to EMF with increased cancer risks. Much of the evidence for the melatonin hypothesis, however, is based on data for rodents (see section 4.4.5). A crucial link for this hypothesis is to determine whether melatonin is suppressed when humans are exposed to magnetic fields at night.
Humans and rodents differ in regard to melatonin. Rodents are nocturnally active, and their nocturnal melatonin concentrations are more easily suppressed by light. Differences in the geometry of the skull and the anatomical location of the pineal gland may cause stronger eddy currents in field-exposed rodents. People show large individual variation in their melatonin patterns, while inbred laboratory strains of rodents do not. Among healthy young men, for example, the peak melatonin concentrations at night can range from 10 pg/ml to over 100 pg/ml (Graham et al., 1998). Human melatonin concentrations also vary as a function of age: in general, they are highest at the age of 1-3, decrease sharply until adolescence, remain fairly stable through adulthood, and finally begin a further decline after about the age of 50 (Waldhauser et al., 1993). Little is known about the basic biology of melatonin in humans and how differences in the concentrations and patterns of this hormone affect individual health and well-being.
The effects of exposure to electric and/or magnetic fields on blood concentrations of melatonin and its major urinary metabolite 6-hydroxymelatonin sulfate (6-OHMS) in humans have been examined in 12 studies (Table 4.45), six with exposure in the laboratory and six observational studies of occupational and residential exposure. The results of five of the six laboratory studies were negative. All-night exposure of human volunteers to magnetic fields under controlled exposure and lighting conditions in the laboratory had no apparent effect on nocturnal blood concentrations of melatonin when compared with equivalent sham-exposure conditions. The exposure parameters evaluated included field frequency (50 or 60 Hz), field polarity (linear or circular), field type (continuous or intermittent), and field intensity (1-20 µT). The studies were performed under double-blind control, and four had a cross-over experimental design in which each volunteer served as his own control to further reduce error variance. When suppression of melatonin has been observed in experimental animals, it was typically a 25-30% reduction (Portier et al., 1998). Two of the three studies reported by Graham et al. (Graham et al., 1997; Graham et al., 1996) had statistical power greater than 0.80 to detect a similar degree of suppression of melatonin in humans at the p < 0.05 level of significance.
The laboratory study of Wood et al. (Wood et al., 1997) is the single one with not completely negative results. They first determined the nocturnal melatonin curve for each individual in their study and then timed presentation of the magnetic field to occur before, during, or after the time of the peak concentration of melatonin in the circulation. Exposure during the rising portion of the nocturnal melatonin curve significantly (p < 0.01) delayed peak onset in one individual, with possible trends for similar effects in several other individuals. The authors note that, while the results are suggestive, the study should be considered only preliminary and the results interpreted with caution. [Owing to the exploratory nature of this study, a number of concerns could be raised in regard to the experimental design, statistical analysis, and interpretation of the findings.]
All six of the studies of occupational and residential exposure listed in Table 4.45 provide at least some evidence of field-related suppression of 6-OHMS, the major urinary metabolite of melatonin. [As would be expected, there is wide variation in exposure conditions, the duration, precision, and type of measures obtained, the presence of possible confounders (e.g. light at night, shift work), and the general characteristics and health status of the individuals studied. Nevertheless, these studies are directly relevant to the effects of magnetic fields under present-day environmental conditions. The study of Burch et al. (1998) is perhaps the most relevant for occupational exposure in the utility environment, and that of Kaune et al. (1997) for assessing the effects of residential exposure.] Burch et al. reported that TWA magnetic field intensity, intermittency, or cumulative exposure had little influence on morning concentrations of 6-OHMS in the workplace. At home, morning concentrations of 6-OHMS were significantly (p < 0.05) associated with 24-h measures of magnetic fields if the exposure remained temporally stable over at least 3-5 min. Additional analyses suggested that morning 6-OHMS concentrations were most likely to be reduced in individuals who were exposed to temporally stable magnetic fields both at home and in the workplace.
Using 72-h measurements of magnetic fields in bedrooms made with EMDEX-2 monitors, Kaune et al. (Kaune et al., 1997) reported a significant (p < 0.05) decrease in the log 6-OHMS concentration in morning urine samples as the log mean magnetic field intensity increased (range, 0-1.5 µT). For example, a twofold increase in average magnetic field resulted in a 6-8% decrease in urinary metabolite concentration. This effect was strongest in summer and was limited to women who used medications known to reduce melatonin (e.g. beta blockers). A similar relationship was observed as a function of the proportion of night-time bedroom measurements > 0.2 µT. Additional analyses indicated that 6-OHMS concentrations were not altered as a function of wire code, short-term variation in bedroom or personal dosimetry measurements, or the proportion of light-at-night measurements > 10 lux. No relationship was observed between 6-OHMS concentrations in urine and menopausal status, current smoking, or use of an electric blanket within the previous month.
[The data are inadequate to suggest an effect of EMF on melatonin concentrations; however, additional research is needed to better understand the differences found in the laboratory and in observational studies. Obvious factors to consider include single versus long-term exposure, the type and complexity of the exposure parameters assessed, and the characteristics of the individuals studied. More needs to be known about the effects of exposure at different times of day. Inclusion of morning urine samples to assess the contribution of melatonin in studies of occupational or residential exposure might be useful.]
Five double-blind, laboratory-based studies have been conducted to evaluate the effects of exposure to power-frequency electric and/or magnetic fields on a variety of biochemical measures in humans. The volunteers in these studies were unable to detect a difference between field and sham exposure. Selmaoui et al. (Selmaoui et al., 1997) described the results of additional analyses performed on the data collected by Selmaoui et al. (Selmaoui et al., 1996b) to evaluate pituitary, thyroid, and adrenocortical hormones. No significant differences were observed between sham- and field-exposure (50 Hz, 10 µT, 8 h) in the concentrations of thyroid-stimulating hormone, follicle-stimulating hormone, luteinizing hormone, triiodothyronine, thyroxine, free triiodothyronine, thyroxine-binding globulin, cortisol, or 17-hydroxycorticosteroids. Gamberale et al. (Gamberale et al., 1989) also failed to find any field-related change in the concentrations of testosterone, thyroid-stimulating hormone, luteinizing hormone, follicle-stimulating hormone, cortisol, or prolactin in his study of the exposure of 26 linesmen working on a 400 kV transmission line. Maresh et al.(Maresh et al., 1988) reported no effect of exposure to a 9 kV/m, 20 µT field on the concentrations of cortisol, growth hormone, testosterone, and plasma lactic acid or on the hematocrit or hemoglobin content in comparison with sham-exposure conditions. Graham et al.(Graham et al., 1990) also reported no effects of multiple 6-h sessions of exposure to combined EMF (9 kV/m, 20 µT) on the concentrations of dopamine, cortisol, epinephrine, norepinephrine, 5-hydroxyindoleacetic acid, homovanillic acid, or vanillylmandelic acid, which include the traditional measures of stress (cortisol, epinephrine, norepinephrine, and dopamine).
[These studies provide no evidence for detrimental effects of exposure to ELF EMF for at least 8 h on human hypothalamic, pituitary, thyroid, or adrenal hormonal systems or on traditional biochemical measures of the human stress response; however, very little is known about the response of women to EMF or about possible effects on female reproductive hormones.]
Few laboratory studies have examined the effects of exposure to EMF on immune function in human volunteers. In perhaps the most comprehensive report to date, Selmaoui et al. (Selmaoui et al., 1996a) described the results of additional analyses of the data they collected in 1996 to evaluate hematological and immune system measures. No significant differences were observed between sham- and magnetic field-exposed (50 Hz, 10 µT, 8 h) men in hemoglobin concentration, hematocrit, or erythrocyte, platelet, total leukocyte, monocyte, lymphocyte, eosinophil, or neutrophil counts. The results of flow cytometric analysis of immunological variables (CD3, CD4, CD8, natural killer cells, and B cells) before and after exposure and control sessions were also unchanged. Earlier, (Hauf, 1982) and Graham et al. (Graham et al., 1990) also reported no reliable field-related changes in blood chemistry, leukocyte or lymphocyte counts, blood gases, lactate concentration, or circulating stress hormones.
[All of the results of laboratory studies obtained to date are based on short-term exposure of healthy young men; no published reports have addressed possible differential effects as a function of age or gender, although the observational studies of suppression of the melatonin metabolite suggest that such studies might be appropriate. The data are inadequate to suggest a health risk.]
Various authors have suggested that environmental exposure to electric and/or magnetic fields acts as a low-level stressor or is associated with an increased frequency of a variety of negative mood states, including clinical depression, suicidal tendencies, feelings of irritability, and loss of libido (Bell et al., 1991; Poole et al., 1993; Reichmanis et al., 1979; Wilson, 1988). In this context, it would be important to determine whether field exposure under controlled laboratory conditions has a negative effect on traditional self-reported measures of mood and personality or on objective biochemical indicators of the human stress response. [The use of double-blind control procedures is essential in this type of evaluation, since the primary purpose of such procedures is to allow a distinction between any effects due to field exposure per se and any effects associated with the subjective expectation about being exposed.]
Eight double-blind studies have addressed this issue (Akerstedt et al., 1997a; Akerstedt et al., 1997b; Cook et al., 1992; Graham et al., 1987; Graham et al., 1990; Graham et al., 1994; Maresh et al., 1988; Selmaoui et al., 1997; Stollery, 1986; Stollery, 1987), with overwhelmingly negative results. The range of exposure parameters assessed in these studies included field frequency (50 and 60 Hz), magnetic field flux density (1-30 µT), electric field strength (0-12 kV/m; and injection of electric current equivalent to 36 kV/m), and exposure duration (2-8 h). [Short-term exposure of healthy young volunteers to EMF has no apparent health consequences for stress levels or mood.]
Reports of individuals who claim to be sensitive to electricity first began to appear in the late 1970s with the introduction of VDTs into the modern office environment (Pearce, 1984). Thousands of employees, primarily in Sweden, have reported they are sensitive to EMF (for recent reviews, see (Bergqvist & Vogel, 1997; Liden, 1996)). Although there is substantial variation among individuals, the reported subjective symptoms and physiological reactions include sleep disturbances, general fatigue, headache, difficulty in concentrating, dizziness, eye strain, facial skin problems (e.g. dry skin, rosacea, seborrhetic eczema, and sensations of itching, burning, or stinging). Typically, the symptoms appear only intermittently at first, subsiding when the individual is away from the VDT. Over time, however, the symptoms may become more pronounced and persistent and begin to interfere with the ability to work or to stay in the general work environment (Bergqvist & Wahlberg, 1994; Sandström et al., 1997). Other electromagnetic devices and appliances (e.g. office equipment, fluorescent lights, household appliances, televisions, and mobile telephones) have also been reported to trigger these adverse reactions in afflicted individuals.
A number of 'provocation' studies have been carried out to evaluate individual sensitivity to DC fields, electrostatic fields, AC 50 Hz and 60 Hz EMF, low-MHz radio-frequency fields, and EMF of the type described by Kavet and Tell (Kavet & Tell, 1991) as associated with both plasma and cathode ray tube VDTs (Andersson et al., 1996; Arnetz, 1997; Arnetz et al., 1997; Sandström et al., 1997; Swanbeck & Bleeker, 1989). These studies have been performed in a controlled laboratory environment with double-blind control procedures. Healthy controls and volunteers suffering from electrical sensitivity were asked to detect when they were being exposed in a series of standardized test trials. The measures included detection accuracy, self-reported measures of stress and arousal, blood samples for analysis of stress biochemistry, and histological samples from facial skin.
In general, patients and volunteers in these studies were not able to reliably distinguish exposed from unexposed conditions, and neither subjective symptoms nor biochemical measures were significantly related to the actual exposure conditions. Patients did, however, report accentuated symptoms when they believed they were being exposed, regardless of the actual conditions (Andersson et al., 1996). Berg et al. (Berg et al., 1990) performed a histopathological study of 134 individuals with skin complaints. Analysis of objective biopsy data revealed no significant difference between 83 persons highly exposed to VDTs and 51 persons with low exposure to VDTs. No dose-response effect was observed between the amount of VDT exposure and objective skin signs.
Some positive results have, however, been reported. Rea (Rea et al., 1991) found that 16 patients out of 100 had symptoms and signs in response to blind exposure. This finding could not be replicated in a subsequent study (Wang et al., ). Similarly, Johansson et al. (Johansson et al., 1994) performed an uncontrolled provocation study in which two patients sat in front of a television set. Analysis of skin biopsy samples revealed total disappearance of somatostatin-positive cells after 3 h of exposure. [Few procedural or exposure details were given, and the study has not been replicated.]
Sandström et al. (Sandström et al., 1995) performed a case-control study of 163 VDT-exposed workers with skin rashes and found an increased odds ratio (3.0; 95% CI, 1.2-7.2) for symptoms in people who worked in rooms with a 50 Hz background electric field > 31 V/m in comparison with workers in rooms with background electric fields < 10 V/m. After adjustment for possible confounding factors (e.g. work duration, psychosocial climate, and job stress), the odds ratio increased to 4.0 (1.2-13). Further work (Sandström et al., 1997) indicated that the odds ratio was even higher when only females were considered in the analysis (6.6; 1.7-26). [To date, this is the only well-controlled study showing an association between exposure to ambient fields and skin rashes. It is not known, however, whether this association implies a causal relationship between exposure and the symptoms.]
Various alternative hypotheses have been advanced to account for this phenomenon. These include an imbalance in the autonomic nervous system (Portier et al., 1998), enhanced neurophysiological sensitivity to flicker from fluorescent lights (Sandström et al., 1997), general hypersensitivity to environmental stimuli (Lyskov et al., 1998) occupational stress resulting from increased work load and lack of personal autonomy and social support (Berg et al., 1992; Eriksson et al., 1997), and variations in the physical and chemical quality of the indoor environment (Arnetz et al., 1997; Nielsen & Schneider, 1998; Stenberg et al., 1995). Others have noted a similarity in the symptoms of individuals with electrical hypersenstivity and individuals suffering from 'multiple chemical hypersensitivity', 'environmental somatization syndrome', 'environmental illness', and 'twentieth century disease' (Black et al., 1990; Liden, 1996).
[Some individuals have subjective symptoms apparently related to VDT use in the office environment. The evidence is inadequate to relate such symptoms to the EMF associated with that use. Some individuals appear to show positive responses to EMF challenges, but no high-quality double-blinded challenge studies have been conducted which conclusively establish the existence of sensitivity to EMF. There is also no established mechanism for electrical hypersensitivity.]
There is weak evidence that short term human exposure to ELF EMF causes changes in heart-rate variability, sleep disturbance, or suppression of melatonin.
[The conclusion for effects on heart-rate variability was supported by 13 Working Group members; there was 1 vote for 'moderate' evidence, 2 votes for 'no' evidence, 8 abstentions, and 5 absent.]
[The conclusion for effects on sleep disturbances was supported by 15 Working Group members; there were 9 abstentions and 5 absent.]
[The conclusion for effects on melatonin was supported by 16 Working Group members; there was 1 vote for 'moderate' evidence, 2 votes for 'no' evidence, 5 abstentions, and 5 absent.]
There is no evidence that such exposure has other effects on the biological end-points studied in the laboratory.
[This conclusion was supported by 12 Working Group members; there
were 2 votes for 'weak' evidence, 11 abstentions, and 5 absent.
The tie vote was broken by the Chair.]
| Spectral analysis | ||||
| (Gamberale et al., 1989) | 26 experienced utility linesmen aged 25-52 years | Men inspected insulators on 50 Hz, 400 kV transmission line, line active on one of the two test days (0700-1700 h; personal dosimeter: average exposure 2.8 kV/m, 23.3 µT | Standard 21-lead clinical EEG exam given at start and end of each test day | No evidence of EEG abnormalities, no evidence of changes in the stability or amplitude of the EEG alpha (8-12 Hz) rhythm |
| (Lyskov et al., 1993b) | 9 male and 11 female volunteers | 1-h no-exposure sham control condition.1-h exposure of the head to a 45 Hz, 1.26 mT horizontal (ear-to-ear) magnetic field, either continuously or intermittently (1s on, 1 s off). | Spectral analysis (FFT) of the EEG recorded from 7 scalp sites before and after exposure and control conditions | Effects seen primarily with intermittent exposure. Absolute and relative power decreased in low frequency EEG bands (1-7 Hz)*, and increased in higher frequency EEG bands (8-20 Hz)* |
| (Lyskov et al., 1993a) | 8 male and 6 female volunteers | 15-min no-exposure sham control condition. 15-min intermittent (1 s on, 1 s off) exposure of the head to above magnetic field condition. | Same as above. | Absolute and relative power increased in higher frequency EEG bands (8-20 Hz)*, but 15-min exposure had no effect on lower frequency bands. |
| (Bell et al., 1992) | 10 volunteers and 10 neurological patients | The head of each subject was exposed in a series of 2 s on, 5 s off trials under four test conditions: no-exposure control; 78 µT, AC magnetic field; 78 µT, DC magnetic field; and combined 78 µT AC and DC magnetic field conditions | EEG recorded from multiple sites. Effect defined as change (±) in any 0.5 Hz EEG spectral band compared with previous control period. | No EEG effects found in control. Overall, AC magnetic field associated with EEG changes in 19 of 20 subjects* (30% responded to DC field, 70-80% responded to AC field). Combined AC/DC exposure not different from AC exposure alone |
| Event-related-potential | ||||
| (Graham & Cohen, 1985) | 18 men 18 women 18-35 years | 12 in sham-exposed control group 12 in 60 Hz, 10 µT exposure group12 in 60 Hz, 20 µT exposure group | Short-latency (< 50 msec) neural conduction components of auditory, visual and somatosensory ERPs recorded during magnetic field and control conditions | Main ERP neural conduction time measures not influenced by magnetic field in any sensory modality. One exception: amplitude decreased for somatosensory ERP mid-latency components**, an effect similar to that seen in monkeys by Dowman et al., 1989 |
| (Graham et al., 1987) | 12 men 18-35 years | All subjects participated in four, 6-h test sessions; half sham, half involving exposure to a 60 Hz 9 kV/m, 20 µT combined field.
| Long-latency (> 100 msec) ERPs recorded in standard "oddball" target-detection paradigm before and after test sessions. | No effects on visual ERP. Exposure associated with increased amplitude of P300 component of the auditory ERP* |
| (Cook et al., 1992) | 30 men 21-35 years | 18 men exposed and sham-exposed equally over four, 6-h sessions to a continuous 9 kV/m, 20 µT combined field. 12 men exposed to combined field in all sessions. | Long-latency ERPs recorded in "oddball" task before, during, and after test sessions | No effects on visual ERP. Exposure associated with increased amplitude of the P300 component of the auditory ERP* . Effects on ERP components greatest soon after field activated, and again at end of day when field switched off. |
| (Graham et al., 1994) | 54 men 18-35 years | All subjects sham exposed in one 6-h session. In a similar session: 18 exposed to a 60 Hz, 6 kV/m, 10 µT field18 exposed to a 60 Hz, 9 kV/m, 20 µT field18 exposed to a 60 Hz, 12 kV/m, 30 µT field | Same as above | Exposure to low and intermediate-strength groups associated with a timing delay (latency) in the appearance of P300 component in ERP waveform**. Effect did not occur in the high EMF group. Amplitude of the N200/P300 component complex increased in all field-exposed groups compared with sham control*. |
| (Graham et al., 1990) | 28 men 18-35 years | 14 subjects in sham control 6-h test session. In a similar session, 14 subjects exposed to a continuous 60 Hz, 12 kV/m, 30 µT field | Same as above | No effects found for visual ERP. Equipment problems resulted in ambiguous effects on latency and amplitude measures of the P300 component of the auditory ERP. |
| (Graham et al., 1998) | 24 men 18-35 years
| 60 Hz, circular polarization 8 No-exposure sham controls 7 Continuous exposure (20 µT/per axis, 2300-0700 h) 9 Intermittent exposure, 1 hr on, 1 h off (20 µT/per axis, 2300-0700 h) | 10 objective EEGsleep measures, sleep quality report measures | Continuous exposure not different from sham on any EEG measure. Intermittent exposure different from sham on 6 of 10 EEG measures** (e.g., less total sleep time, reduced sleep efficiency, increased time in stage 2 sleep, and decreased REM sleep). Sleep quality data reflected EEG disturbances*. |
| (Akerstedt et al., 1997a) | 18 men & women 24-49 years | 50 Hz, Linear polarization 9 No-exposure sham control 9 Continuous exposure 1 µT, (2300-0700 h) In dwelling catheter for serial blood draws | Objective EEG measures, self-report measures, melatonin, Cortisol, GH, ACTH, Prolactin | Exposure associated with a significant* reduction in slow wave sleep (stages 3 and 4). Exposed subjects also had less total sleep time and reduced sleep efficiency. No effect on hormones. |
* = p < 0.05
** = p < 0.01
| (Cook et al., 1992) | 30 men 21-35 years | 18 men exposed and sham exposed equally over four, 6-hr sessions to a continuous 9 KV/m, 20 µT combined field. 12 men exposed to combined field in all sessions. | Pre- and post- session multi-task test battery (e.g., RT, math, vigilance, time estimation, memory). | EMF exposure effects on performance different from sham-exposed on only one task in the test battery (exposed subjects made fewer errors when performing the choice RT task *). |
| (Gamberale et al., 1989) | 26 linesmen 25-52 years | Men inspected insulators on 50 Hz, 400 KV transmission line. Line active on one of the two test days (0700-1700 h; personal dosimeter: average exposure 2.8 kV/m, 23 µT) | RT, vigilance, memory, and perceptual speed tests before and after exposure and control days. | Performance on exposure days did not differ from performance on control days. |
| (Graham et al., 1987) | 12 men | Four 6-hr test sessions; half sham, half involving cont. exposure to a 60 Hz 9 kV/m, 20 µT field. | Pre- and post- session multi-task test battery.(RT, math, vigilance, time estimation, memory). | EMF exposure effects on performance were different from sham -exposed on only one of the battery tasks (exposed subjects made less errors when performing the choice reaction time task *). |
| (Graham et al., 1990) | 28 men 18-35 years | 14 men in sham control 6-h test session. In a similar session, 14 men exposed to continuous 60 Hz, 12 kV/m, 30 µT field. | 4 administrations of multi-task battery (RT, math, vigilance, memory, time estimation). | Performance on exposure days did not differ from performance on control days. |
| (Graham et al., 1994) | 54 men 18-35 years | All men sham exposed in one 6-h session. In a similar session: 18 men: 60 Hz, 6 kV/m, 10 µT field18 men: 60 Hz, 9 kV/m, 20 µT field18 men: 60 Hz, 12 kV/m, 30 µT field | RT, attention, and time perception tests given before, after, and 2 times during exposure and control sessions. | Only exposure at the lowest level was associated performance changes, compared with sham-exposed conditions. RT was slower (12%) and accuracy decreased (13%) on the time perception task*. |
| (Lyskov et al., 1993b) | 9 men 11 women | 1 h no exposure control condition | RT to auditory stimuli | Performance on exposure days did not differ from performance on control days |
| (Lyskov et al., 1993a) | 8 men 6 women | Same as above, except exposure duration in all conditions was 15 min | RT to auditory stimuli | Performance on exposure days did not differ from performance on control days. |
| (Stollery, 1986) | 76 men 18-65 years | All men sham exposed in one session. In 2nd session, 50 Hz, 50 µAmp current injected from 10:30-16:00 h into each of 10 electrodes on body. (Equivalent to 36 kV/m) | Memory, attention, vigilance, and reasoning tasks given 4 times a day | No EMF effects on vigilance, sustained concentration, or verbal reasoning were observed. |
| (Podd et al., 1995) | 6 men18 women 19-55 years | 5-min exposure of head to a 0.1 mT, 0.2 Hz or 43 Hz magnetic field, or to a sham-exposed condition | RT to visual stimuli | Performance on exposure days did not differ from performance on control days. |
| (Teresiak & Szuba, 1989) | 64 men 22-63 years | Study 1: 50 Hz ambient electric field (0, 4.4, 10.9, and 13 kV/m). Study 2: 50 Hz current injected into men (0, 50, 135, and 160 µamps). Equivalent to Study 1. | RT to the presentation of visual and auditory signals | Significant slowing of RT found in the presence of the 13 kV/m field*, and with current injection of 160 µA*. |
RT, reaction time
* = p < 0.05
** = p < 0.01
| (Graham et al., 1987) | 12 men 21-35 years | All men participated in four, 6-h test sessions; half sham, half involving continuous exposure to a 60 Hz 9 kV/m, 20 µT field. | Electrocardiogram (R-R interval) | HR slowing was greater after EMF exposure than in sham exposure*. |
| (Maresh et al., 1988) | 11 men 21-29 years | Sham exposure (2 h) EMF exposure (2 h, 60 Hz, 9 kV/m, 20 µT) Each preceded by 45 min of rest or exercise | Electrocardiogram (R-R interval) | HR slowing was greater during EMF exposure than in sham exposure*. |
| (Cook et al., 1992) | 30 men 21-35 years | 18 men exposed and sham exposed equally over four, 6-hr sessions to a continuous 9 kV/m, 20 µT combined field. 12 men exposed to combined field in all sessions. | Electrocardiogram(R-R interval) | HR slowing was greater during EMF exposure than in sham exposure** Also lower in subjects who were field exposed in all sessions, but the effect was not significant. |
| (Graham et al., 1994) | 54 men 18-35 years | All men sham-exposed in one 6-hr session. In a similar session: 18 men: 60 Hz, 6 kV/m, 10 µT field 18 men: 60 Hz, 9 kV/m, 20 µT field18 men: 60 Hz, 12 kV/m, 30 µT field | Electrocardiogram(R-R interval) | HR slowing greater after EMF exposure at 9 kV/m, 20 µT*, but not at lower or higher exposure conditions. |
| (Graham et al., 1990) | 28 men 18-35 years | 14 men in sham-exposed 6-hr test session. In a similar session, 14 men exposed to a continuous 60 Hz, 12 kV/m, 30 µT field. | Electrocardiogram(R-R interval) | HR was not slower at higher EMF exposure than in sham exposure. |
| (Korpinen et al., 1993) | 41 men 21-48 yrs | 26 men sat for 1-h under a 400 KV line (3-4 kV/m, 1.4 µT), and at 200 m from the line. 15 men also tested in sham conditions (0.01 kV/m, 0.01 µT). | Ambulatory monitoring of electrocardiogram | No effect on heart rate at field strengths < 6 kV/m and 10 µT. |
| (Sastre et al., 1998) | Study 1: 29 men 18-35 yrs Study 2: 22 men 18-35 yrs Study 3: 26 men 18-35 yrs |
11 men in sham-exposed group. Two intermittent Exposure groups of 9 men each (60 Hz, 1 µT and 20 µT). Each man was his own control. 2 sessions: sham and intermittent exposure to 60 Hz, 20 µT. Each man was his own control. 2 sessions: sham and continuous exposure to 60 Hz, 20µT. |
HRV analyses based on electrocardiogram R - R intervals. Same as above. Same as above. | Intermittent exposure at 20 µT reduced "low band" power compared with sham exposure or 1 µT*. Intermittent exposure reduced "low band" power*, and increased "high band"; power**, compared with sham exposure Continuous exposure at 20 µT had no effect on heart rate variability. |
HR, heart rate; HRV, heart rate variability
* = p < 0.05
** = p < 0.01
| Laboratory studies | ||||
| (Graham et al., 1996)(2 studies) | Study 1: 33 men 18-35 years
Study 2: 40 men 18-35 years | 60 Hz, circular polarization 11 men: no-exposure, sham control group 11 men: 60 Hz 1 µT intermittent
exposure 11 men: 60 Hz 20 µT intermittent exposure (Exposure duration 23:00-07:00 h) 60 Hz, circular polarization (23:00-07:00 h) All men sham exposed in one session. All men intermittently exposed to 20 µT in 2nd session. (Each man his own control). | Hourly melatonin (2300-0700) Hourly melatonin (23:00-07:00) | No overall effect; Men with low basal melatonin had greater suppression in a response to light* and
magnetic fields*. Effects of exposure no different from sham exposure.No effect in men with low basal melatonin. |
| (Graham et al., 1997) | 40 men 18-35 years | 60 Hz, circular polarization, (23:00-07:00 hrs) All men sham exposed in one session All men continuously exposed to 20 µT in 2nd session. (Each man his own control). | Hourly melatonin (23:00-07:00) | Effects of exposure not different from sham exposure. No effect in men with low basal melatonin. |
| (Selmaoui et al., 1996b) | 32 men20-30 years | 50 Hz; 10 µT; continuous and intermittent exposure linear and circular polarization; (23:00-08:00)Each subject was his own control | Hourly melatonin, 6-OHMS | Effects of exposure not different from sham exposure. |
| (Akerstedt et al., 1997b) | 18 men andwomen24-49 years | 50 Hz, 1 µT, continuous exposure,linear polarization, (23:00-07:00) | Hourly melatonin | Effects of exposure not different from sham exposure |
| (Wood et al., 1997) | 30 men 18-49 years | 50 Hz, 20 µT, circularly polarized, sinusoidal and square wave fields, 1.5 - 4.0 h exposures | Periodic melatonin (1800-0700 hrs) | No effect when sinusoidal or square wave exposure examined separately; no effect when exposure occurred at MLT peak time or later at night; early night exposure delayed MLT onset in 1 subject exposed over 5 sessions**. |
| Environmental studies | ||||
| (Wilson et al., 1990) | 32 women 10 mentested at home | Home use of standard (conventional) versus modified AC or DC CPW electric blankets for 8 weeks (0.2 - 0.6 µT exposure range). | Morning and evening 6-OHMS levels (notcreatinine corrected) | No overall effect. 7 of 28 CPW users had decreased 6-OHMS in the last 3-week test period* and showed a rebound effect*. |
| (Pfluger & Minder, 1996) | 108 male electric railway workers | 42 controls (50 Hz, 1 µT); 66 locomotive engineers (16.7 Hz, 20 µT) Continuous sampling of magnetic field (30 min - 4 h) | Urinary 6-OHMS samples collected in morning and evening (creatinine corrected) | Evening 6-OHMS decreased on work days, but not on leisure days*. No effect on morning 6-OHMS levels. No evidence for dose-response curve. |
| (Burch et al., 1998) | 142 male utility workers 20-60 years | 57 controls (60 Hz, 0.1 µT) 29 field generation workers (60 Hz, 0.2 µT) 56 field distribution workers (60 Hz, 0.1 µT) 72 h EMDEX magnetic field level and light intensity measurements taken. | 1 home baseline and 3 workday morning 6-OHMS samples (creatinine corrected) | Workplace: No effects on 6-OHMS of magnetic field intensity, intermittency, or cumulative exposure. At home: Temporally stable magnetic field exposure associated with reduced 6-OHMS* |
| (Arnetz & Berg, 1996) | 47 VDT workers who participated in a 1988 study | Compared 1 day working versus not working in front of VDT. No magnetic field measures taken. | Assay of a.m. and p.m. melatonin samples obtained in earlier study (not creatinine corrected) | Melatonin decreased on work day compared with control day*. |
| (Kaune et al., 1997) | 203 women (control group in breast cancer study)20-74 yrs | 72-h EMDEX measurement of bedroom magnetic field levels (range 0 - 1.5 µT) and light intensity. Personal exposure monitoring. | Morning 6-OHMS levels measured on 3 consecutive days, 3 or 6 months apart(creatinine corrected) | As log mean bedroom magnetic field increased, log 6-OHMS levels decreased*. Effect is strongest in summer and in women taking medications which reduce melatonin. No effect on 6-OHMS of wire code, intermittency, personal dosimetry, or light-at-night. |
6-OHMS, 6-hydroxymelatonin sulfate; CPW, continuous polymer wire; VDT, visual display terminal
* = p< 0.05
** = p< 0.01