Nocturnal melatonin levels in human volunteers exposed to intermittent 60 Hz magnetic fields

Bioelectromagnetics 17:263-273 (1 996)

Nocturnal Melatonin Levels in Human Volunteers Exposed to Intermittent

60 Hz Magnetic Fields

Charles Graham, Mary R. Cook, Donald W. Riffle, Mary M. Gerkovich, and Harvey D. Cohen

Midwest Research Institute, Kansas City, Missouri

Two double-blind laboratory-based studies were performed to determine whether a suppression of nocturnal melatonin similar to that observed in rodents occurs when humans are exposed to magnetic fields at night. In study 1 , 33 men were exposed to sham, 10 mG, or 200 mG intermittent, circularly polarized magnetic fields from 2300 to 0700 h under controlled environmental and exposure test conditions. Overall, exposure had no effect on melatonin levels. Men with preexisting low levels of melatonin, however, showed significantly greater suppression of melatonin when they were exposed to light and also when they were exposed to the 200 mG magnetic-field condition. Study 2 directly tested the hypothesis that low-melatonin subjects show enhanced sensitivity when exposed to light and to 200 mG magnetic fields. After preexposure screening, each of 40 men slept in the exposure facility on two nights. On one night, the men were sham exposed. On the other night, they were exposed to the 200 mG field condition used previously. Again, exposure had no overall effect on melatonin levels. The original finding of enhanced sensitivity in low-melatonin subjects was not replicated in this study. We conclude that the intermittent exposure conditions used in these two studies were not effective in altering nocturnal melatonin release patterns in human volunteers. Further research is underway with regard to exposure parameters, hormonal and immune system measures, and individual differences. 01996 Wiley-Liss, lnc.

Key words: ELF, EMF, human, neuroendocrine, cancer

INTRODUCTION of carcinogenic processes. Recent reports also indicate

Numerous epidemiological reports suggest a link between occupational or residential exposure to power- frequency magnetic fields and increased cancer risk [for recent reviews see Coleman and Beral, 1988; Savitz et al., 1989; Theriault, 19911. To date, most studies have reported only relatively modest elevation in risk. Given the extensive use of electric power, however, even small increases in risk could have substantial consequences for public health. A primary need exists for laboratory- based research to identify a biological mechanism that could account for the statistical association observed in the epidemiological studies.

Melatonin is a hormone released by the pineal gland in the brain during the dark period of the circadian cycle. Melatonin is reported to stimulate immune function [Maestroni, 19931, and it has been implicated in the control of cell proliferation, the growth of transplanted

that this hormone is a potent scavenger of free radicals [Tan et al., 19931. Beginning in the early 1980s, vari- ous studies, primarily with rodents, have suggested that night-time levels of melatonin are reduced by electric and magnetic field (EMF) exposure [Wilson et al., 1986, 1990; Reiter 1988, 1993; Lerchl et al., 1990; Kato et al., 19931. Stevens [ 19871 proposed that this relationship might provide a plausible biological mechanism to account for the epidemiological findings. A crucial link in this hypothesis, of course, is to determine whether a similar suppression of melatonin occurs when humans are exposed to magnetic fields at night.

With regard to melatonin, humans and rodents differ in several interesting and possibly important ways. For example, exposure to light is the most widely known suppressor of melatonin. In the nocturnally active rodent, brief exposure at very low light levels is sufficient

tumors, and the promotion and/or coprornotion of mam- mary tumors [Hill and Blask, 1988; Reiter, 1988; Wil- son et al., 19901. In addition to its immunomodulatory effects, melatonin also plays a key role in the regula-

Received for review May 22,1995; revision received October 19, 1995.

Address reprint requests to Charles Graham, Ph.D., Midwest Research tion of reproductive hormones implicated in a number Institute, 425 Volker Boulevard, Kansas City, MO 641 10.

0 1996 Wiley-Liss, Inc.

264 Graham et al.

to suppress melatonin [Reiter, 19891. In humans, however, the light levels required for similar suppression are reported to be orders of magnitude higher [Lewy et al., 19801. We also differ from rodents in terms of the ana- tomical location of the pineal gland. The rodent pineal gland is located peripherally in the brain, just beneath the dura and skull; in the human, the gland is positioned deep inside the brain. I t has been suggested that such differences may expose the rodent to eddy currents or to different and possibly stronger induced fields [Kato et al., 19931.

Perhaps the most striking distinction is not that humans differ from rodents but that we differ so greatly one from another. People show very large individual differences in their typical night-time peak melatonin levels, whereas rodents bred for laboratory studies do not [Reiter, 19891. In normal, healthy young men, for example, peak melatonin levels at night can range from 10 to 100 pg/ml or more [Graham et al., 1994a,b]. Little is known about how such large differences in levels of this powerful hormone impact on individual health and well being. Human melatonin levels also vary as a function of age. In general, levels are highest between 1 and 3 years old, decrease sharply until adolescence, remain fairly stable through adulthood, and finally begin a further decline after about age 50 [Waldhauser et al., 19931. Several reports also suggest that peak melatonin levels in people may vary between the sexes, being relatively lower in women at the time of ovulation [Reiter, 19891.

No controlled laboratory-based data are available concerning whether human melatonin is affected by exposure to 60 Hz magnetic fields. Wilson and his colleagues [Wilson et al., 19901, however, have examined the urinary metabolite of melatonin in people sleeping under electric blankets in their own homes. They concluded that melatonin levels were significantly decreased in some of the individuals. They suggest that, in some individuals, pineal gland function may be influenced by intermittent exposure to magnetic fields of sufficient intensity and duration. In this report, we describe the results of the first two double-blind laboratory-based studies designed to address this issue [Graham et al., 1994b,c].

STUDY 1

MATERIALS AND METHODS

Experimental design. The main goal of this study was to evaluate the effects of intermittent exposure to magnetic fields at two levels of field strength on nocturnal melatonin release patterns in human volunteers. Thirty- three healthy young men participated in two preexposure screening sessions to determine each man’s basal peak melatonin level prior to exposure and his sensitivity to

bright light, a known suppressor of melatonin. The preexposure session was followed by one all-night exposure test session.

The men were matched on basal peak melatonin level (i.e., concentration in blood at 0200 h) and were randomly assigned in triads to three experimental groups: a sham-control group; a group exposed to a 10 mG magnetic field; and a group exposed to a 200 mG magnetic field. On the exposure test night, each man slept in the human exposure test facility at Midwest Research Institute (MRI) from 2300 to 0700 h. Exposure to the circularly polarized magnetic field during the night was intermittent (1 h off/l h on, with the field cycling on and off every 15 s throughout each of the 1 h on periods). The study was performed double blind to control for the effects of expectation. Blood samples were obtained hourly through the control and test nights for determination of possible field effects on melatonin concentration. Statistical power analysis indicated that the above design was adequate to detect a 25% decrease in nocturnal melatonin with alpha set at .05 and power of .78.

Intermittent exposure was selected because previous work in our laboratory showed it to be an effective exposure condition in altering human physiology [Graham et al., 19901. The intermittency pattern used is similar to one shown in our previous work to have significant effects on measures of human brain and heart activity. A field strength of 200 mG was selected because we have repeatedly found effects on a variety of endpoints at this level [Cook et al., 1992; Graham et al., 1994al. The comparison field strength of 10 mG was selected because it is a level often found in homes and workplaces. Compari- son across field strengths is important, because i t addresses a major concern people have about exposure to power frequency magnetic fields. Subjects. The participation of volunteers in this study was reviewed and approved by the Human Subjects Com- mittee at MRI in accordance with federal guidelines and regulations [Federal Register, 199 1 1 . Volunteers were recruited by posting informational notices at local col- leges and universities. Individuals who telephoned about the study were given a complete and accurate descrip- tion of the purpose, procedures, risks, and benefits of participation and were screened to confirm that they met the specific criteria set for participation (male, ages 18- 35 years, no chronic disease or disability, no recent serious acute illness, no medications, regular sleep habits, and no night work). Prior to participation, all volunteers who met criteria came to the laboratory and toured the facilities. The study was explained again, and written informed consent obtained.

Complete data on basal peak melatonin level and light sensitivity was collected on 42 men. Of these, the

Human Melatonin Study 265

and verified using a Digital Photometer (model 516; Tektronix, Beaverton, OR) with calibration traceable to NET.

Double-blind system. The facility’s computerized double-blind system was used to prevent subjects and experimenters from knowing whether sham or exposure conditions were in effect in any given test session. This system has proven effective over multiple human exposure studies [Graham et al., 1994al. At the start of a test session, the experimenter activated the system by entering the unique identification code assigned to the subject and session into the control program. From that point on, operation of the double-blind system was completely automatic. Exposure conditions were monitored continually during test sessions by an au- tomated data-logging procedure designed to check the field status, as reported by monitoring transducers inside the facility, against the predetermined field conditions designated in the test protocol. If a discrepancy was detected, then a malfunction report was generated. In one of the exposure test sessions, a problem with the solid-state switching equipment prevented the subject from being exposed to the circularly-polarized magnetic field as planned (the horizontal field was not activated). This problem was corrected after the session. The subject was dropped from the study and replaced by another volunteer.

final test sample that completed all study requirements consisted of 33 young (mean = 23 years; range, 19-34 years), white (97%), nonsmoking (88%) men of average height (70.7 inches; range, 63-76 inches) and weight (17 1.5 pounds; range, 123-250 range) who typically slept 7 h per night. Characteristics of the screening sample of 42 men did not differ from those of the final sample. Data collection could not be completed on the remaining men for a variety of reasons (e.g., scheduling conflicts, illness, failure to keep appointments, problems during blood draws, equipment problems, etc.). Exposure facility. Characteristics and control systems of the MRI Human Exposure Test Facility have been documented by the National Institute of Standards and Technology (NIST) and are described in Cohen et al. [ 19921 and in Graham et al. [ 1994bl. The subjects were exposed to a uniform (4-7%), circularly polarized, 60 Hz magnetic field. The field was generated in the facility by six Helmholtz coils surrounding the exposure room in both the vertical and horizontal axes. The horizontal axis of the field is oriented from the door- way to the rear of the exposure room, and the vertical axis is oriented from the floor to the ceiling. Each axis of the magnetic field is independently energized from an adjustable autotransformer. The horizontal field current is shifted from the vertical field current by a phase angle of 90 degrees. Subjects slept on a cot in the facility with their bodies oriented in line with the horizontal component of the magnetic field (north to south in the facility). Recent measurements made in the facility with a three-axis Fluxgate Sen- sor (model MAG-03MC Bartington Instruments, Ltd., Oxford, U.K.) show the vertical and horizontal components of the earth’s magnetic field to be 429 mG and 101 mG, respectively.

The magnetic field was presented intermittently over the exposure test night from 2300 to 0700 h in alternating 1 h field-off and field-on periods. During field-off periods, the equipment was not energized. During field-on periods, the field cycled on and off at 15 s intervals. A software control system in the facility is programmed to switch the field generators on and off at the selected rate. Switching is accomplished by a logic control signal sent to a gate module in each generator circuit (CRYDOM model CTD24 10 Interna- tional Rectifier, El Segundo, CA). These gating circuits switch on when the AC voltage is at zero and switch off when the AC current is at zero. This type of zero- crossing procedure produces field changes relatively free of low-frequency transients. Illumination in the facility during the night was provided by incandescent lamps located above the translucent ceiling panels. Light levels were maintained at less than 10 lux during the test sessions. Lighting conditions were tested

Measures. Although additional physiological and per- formance measures were obtained in this study, the focus in this report will be on the assessment of possible field effects on melatonin. Hourly melatonin levels measured in plasma provided the primary outcome measure of the study. The melatonin concentration of the coded, frozen plasma samples was determined by using the Melatonin Direct Radioimmunoassay distrib- uted by ELIAS USA, Inc. (Osceola, WI). This assay provided for the direct quantitative determination of N-acetyl-5-methoxy-tryptamine (melatonin) in plasma without extraction. Assay methods are described in Zimmermann et al. [ 19901. The limit of quantitation for the assay was 2.5 pg/ml. Intra-assay variability was less than 10% at 15, 50, and 150 pg/ml. Inter-assay variability was 15% at 15 pglml and less than 10% at 50 and 150 pg/ml.

Subjects and experimenters also completed the Field Status Questionnaire (FSQ), a paper-and-pencil instru- ment developed in our laboratory to evaluate the effectiveness of the double-blind control procedures used in exposure studies. The FSQ was completed in the morning after each test session. The respondents answered three questions: “In your judgment, were the fields on or off? How confident are you of this judgment (on a one-to-five scale)? What are you basing this judgment on?”

266 Graham et al.

PROCEDURES

Preexposure screening. Each subject participated in two preexposure group screening sessions held within a 10 day period. The purpose of these sessions was to measure each individual's basal peak melatonin level as well as the decrease in melatonin induced by controlled exposure to bright light (sensitivity). Groups of three or four subjects were scheduled for each night session. The men slept in the dark from 2230 to 0200 h, at which time the blood draw procedure began. The men were fitted sequentially with fluoroscopic goggles (model 502300; Cone Instruments, Solon, OH) similar to those used by x-ray technicians and were taken to the phlebotomy area, where blood samples were obtained by venipuncture for melatonin assay. The phlebotomy area was lit by incandescent light at a level of 25 lux. The red-lensed goggles maintained dark-adapted vision for the subjects while limiting by 97% the photopic transmittance of the in- cident incandescent light. Thus, venipuctures could be performed safely without the ambient lighting conditions interfering with melatonin production. After the blood draw, each man was taken to the appropriate light or dark exposure area (see below), and the next subject was fitted with the goggles and taken to the phlebotomy area.

In one screening session, the subjects sat in the exposure chamber in dim light (< 10 lux) from 0200 until 0300 h. This was the light level used during the magnetic field exposure test sessions. After 1 h, a second blood sample was collected from the subjects in the order followed previously. The same basic procedures were followed in the other screening session, except that between the first and second blood draw, the subjects were taken to the light-exposure area in the laboratory, where they were seated for 60 min in front of a bank of fluorescent lighting (24,4 foot, 40 Watt Sylvania Cool White tubes) at a measured level of 5,500 lux. Half of the subjects were screened over sessions in the order dark-light; the remaining subjects were screened in the reverse order. After screening, triads of subjects matched for peak melatonin level (mean value over the two blood draws at 0200 h) were assigned randomly to the sham- control group and to the 10 mG and 200 mG intermittent field exposure groups.

Experimental testing. Subjects were instructed to eat balanced meals, to refrain from consuming alcohol for 24 h before each session, and to refrain from caffeine after 1700 h on the day of the test session. On arrival at the laboratory, the subject changed into a surgical scrub suit, vital signs were recorded, and the study nurse in- serted an indwelling catheter in an arm vein for the collection of multiple blood samples over the night. The subject got into bed in the exposure room, and the first blood sample was obtained at 2255 h. The double-blind

control system was activated at 2300 h. The subject remained i n bed until 0700 h. The nurse entered the exposure facility to collect blood samples during the night. These collections occurred between 5 min before the hour and on the hour throughout the night. The samples in EDTA tubes were immediately centrifuged. As many aliquots as possible of plasma were prepared and frozen at -20 "C, for later assay, In the morning, a final blood sample was collected, and the subject and experimenter completed the FSQ.

QUALITY ASSURANCE

A systematic protocol using test instruments traceable to NIST was followed to verify the exposure characteristics of the facility and to calibrate the recording equipment. Exposure characteristics were verified at the start and end of the study and periodically during the study. Magnetic field-generating systems were calibrated using the three-axis Fluxgate Sensor. The strength, uniformity, and phase relationship of the fields were verified. During each test session, computer logs were maintained to verify that each subject was exposed to the appropriate conditions. Chart recordings of ambient temperature and relative humidity were maintained throughout each session. All data-handling operations (e.g., manual scoring, computer entry, biochemical assay, etc.) were completed prior to breaking the double- blind code.

STATISTICAL ANALYSIS

Repeated-measures analysis of variance (ANOVA), with exposure groups (sham, 10 mG, and 200 mG) as a between-subjects factor and with the time of night as a within-subjects factor, was performed for the melatonin data. Alpha was set at P < .05. Significant main effects were followed with t tests to identify the source of significant variance. Multiple-regression analysis was used to predict 0700 h melatonin levels from age, field condition, basal peak melatonin level, and percent suppression by light. Where ANOVA was used, probability values were corrected using the Huynh-Feldt epsilon technique. Probability levels reported here are the corrected ones; the degrees of freedom are not corrected in order to indicate the actual number of data points included in the statistical tests.

RESULTS

Experimental control measures

Temperature and humidity. No differences between the three experimental groups were found for either temperature (P > .79) or humidity (P > .47), and tempera-


lustrates the classic nocturnal release pattern, with mean melatonin values rising rapidly to a peak between 0200 and 0300 h and then gradually tailing off through the remainder of the night. ANOVA performed on these data revealed no significant difference in melatonin levels across the three test groups (group main effect: F = 0.25; df, 2,30; P = .78; group by time interaction: F = 0.04; df, 14,210; P = .99). The exposure groups were not different from the sham-control group, nor were they different from each other. Thus, the overall results from this study of 33 healthy men did not support the hypothesis that exposure to 60 Hz magnetic fields at night suppresses melatonin.

The three test groups, however, were originally created by taking triads of subjects who were matched for basal peak melatonin level and then assigning them randomly to the groups. The net result of this procedure was that each of the groups contained the full range of basal peak melatonin levels represented in the study. Additional analyses were performed to take account of individual differences in preexisting levels of melatonin. All 33 men in the study were ranked from high to low in terms of their basal peak melatonin level prior to exposure. The total group was divided at the median value into a “low” group of 16 men with preexposure peak values less than 60 pg/ml and a “high” group of 17 men with preexposure peak values greater than 60 pg/ml.

The ANOVA was repeated with melatonin level included as a factor in the analysis. A significant interaction involving exposure group by time by peak melatonin level was found (F = 2.71 ; df, 14,189; P = .035)

ture and humidity did not change over the experimental session. The mean temperature was 77.3 O F , and the mean relative humidity was 34.3%.

Double-blind controls. Analysis of the FSQ rating data indicated that the subjects were unable to judge wtether the fields were on at better than chance levels (X-, 2.2; df, 1 ; P > .lo). Twenty-one of the 33 subjects reported very low confidence in their judgments (1 on a scale from 1 to 5); 9 of 11 subjects exposed to 200 mG fields had confidence ratings of 1 . Experimenters were also unable to judge field status at better than chance levels (X’, 0.59; df, 1 ; P > .3). Like the subjects, experimenters had little confidence in their judgments. These results confirm the effectiveness of the double-blind control procedures.

Melatonin concentration Preexposure screening. Across the two sessions, basal peak melatonin levels were very stable within an individual (r = 0.92), and light exposure had the expected suppressant effect. No difference was found in peak level between the two screening sessions at the 0200 h collection; at 0300 h, melatonin levels were significantly suppressed after light exposure compared to levels after dark exposure (F = 93.34; df, 3,90; P < ,001). Basal peak levels, taken as the mean of plasma melatonin concentrations at the two 0200 h blood draws, ranged from 15 to 130 pg/ml. Individuals also varied in their sensitivity to the effects of light exposure. Percent melatonin suppression (0200 h value minus 0300 h value divided by 0200 h value) ranged from 27 to 83%, with the majority of subjects showing reductions of 50- 70% in melatonin levels. Absolute values for suppression of melatonin by light exposure ranged from 9.2 to 100.7 pg/ml.

A graph of the relationship between basal peak melatonin level and suppression indicated a nonlinear relationship. Therefore, a logarithmic transformation was performed on the basal level scores, and Pearson’s r was computed between the two variables. The correlation was -0.36 (P = . O l ) , indicating that individuals with lower preexposure melatonin levels showed relatively greater suppression effects. The significance level of this correlation suggests that the relationship is not due to chance alone; however, the relatively low value of the correlation (accounting for about 13% of the variance in melatonin values after light exposure) indicates that other factors in the situation also con- tribute to the variance observed.

Experimental testing. Figure 1 summarizes the melatonin data from the 33 participants who completed all study requirements. Mean melatonin level and standard error is plotted for each experimental group at each hour from midnight to 0700 h. The curve obtained il-

‘ - - I

0 Sham 0 10 mG - 200 mG

- 3 70 m a a, Z 60

3 50

.s 40 9 9 30

20

10

0 12 1 2 3 4 5 6 7

Time (a.m.)

Fig. 1. Nocturnal melatonin levels in sham and exposed groups (n = 11 per group). Mean melatonin level and standard error are plotted for each experimental group at each hour from midnight to 0700 h. Melatonin levels in the exposure groups did not differ from each other nor did they differ from the control group.

268 Graham et al.

indicated that magnetic field exposure had differential effects on the high and low subgroups. Further ANOVAs were performed on the high and low groups separately. No significant effects involving field exposure were found for the high group. For the low group, changes in melatonin over the night were significantly different for the sham-control and two exposure groups (F = 2.54; df, 14,91; P = .043). Figure 2 illustrates the results observed. Further ANOVAs were performed to compare the 200 mG group directly to the sham group and directly to the 10 mG group. These analyses revealed that, within the low-melatonin subjects, those exposed to the 200 mG magnetic field showed significantly lower levels of nocturnal melatonin compared to either sham (F = 3.53; df, 7,56; P = .018) or 10 mG subjects (F = 3.81; df, 7,63; P = .03). Peak levels of melatonin were reduced during exposure, as were levels in the latter portion of the exposure night. The 10 mG group did not differ from the sham-control group (F = 0.155; df, 7,63; P = .89).

In a final analysis of study 1 data, multiple lin- ear-regression techniques were used to identify significant predictors of an individual's morning (0700 h) melatonin level. Age, preexposure basal peak melatonin level, and percent light suppression together predicted 44% of the variance in morning melatonin levels. Of these factors, age appeared to be the most potent pre- dictor, correlating negatively (r = -.52; P < .01) with

0 Sham c3 10 mG W 200 mG

I P

m 3 50 .g 40 8 3 30 = 20

10

n " 12 1 2 3 4 5 6 7

Time (a.m.)

Fig. 2. Exposure effects in 16 men with preexisting low melatonin levels (<60 pg/ml at 0200 h). Mean melatonin level and standard error are plotted for each experimental group at each hour from midnight to 0700 h. Melatonin levels in the 200 mG magnetic field exposure group were significantly lower than in the sham control group (P = ,018) or in the 10 mG exposure group (P = .03). Peak levels were reduced during exposure to the 200 mG field, as were levels in the later portion of the night.

melatonin level (the older the volunteer, the lower the melatonin level).

The overall results of this study on 33 healthy young men did not support the hypothesis that nocturnal exposure to 60 Hz magnetic fields suppresses melatonin. Specifically, there was no significant difference in melatonin levels across the three test groups (sham control, 10 mG, and 200 mG exposure) when the groups were matched for preexposure melatonin level. Evidence supportive of the suppression hypothesis, however, was found when analyses were performed to take account of individual differences in preexposure melatonin level. In men with preexisting low levels of melatonin, exposure to bright light at night was associated with significantly greater suppression of melatonin. For men in this category, exposure to the 200 mG magnetic field also was associated with a significant suppression of nocturnal melatonin. Peak melatonin levels were reduced, as were levels in the latter portion of the exposure night. These effects were not observed i n men with high levels of melatonin prior to exposure. The results provide partial support for the melatonin hypothesis, and they extend the focus of research from the rodent to the intact, func- tioning human being.

One possible explanation for these observations is that people with low preexposure melatonin levels have such low levels because they are particularly sensitive to environmental factors that suppress melatonin. If this is true, then it may have important implications for epidemiological research and for future studies of biological mechanisms of EMF effects. The evidence on which this observation is based, however, is drawn from analyses performed on subsets of the study volunteers. Thus, many questions remained to be answered, and there was an obvious need for further research. Study 2 was designed to examine the repro- ducibility of the above results and to directly test the hypothesis that men with preexisting low levels of melatonin show greater suppression when exposed to light and to 200 mG magnetic fields.

STUDY 2

MATERIALS AND METHODS

Experimental design. Forty male volunteers participated first in one screening session to determine preexposure basal melatonin level and light sensitivity and then in two exposure test sessions. During one of the all-night exposure test sessions, each man was sham exposed from 2300 to 0700 h; in the other test session, he was exposed for the same duration to the identical 200 mG, intermittent magnetic field condition that was used in study 1. Thus, each subject served as his own


of the same bank of fluorescent lighting that was used in study 1. A second blood sample was obtained at 0300 h after light exposure.

All subjects then participated in two all-night test sessions. In one session, each man was sham exposed; in the other, each was exposed to the 200 mG intermittent magnetic field condition used previously. Half of the subjects were tested in the order sham-field; the remaining subjects were tested in the reverse order. Procedures for sham and field sessions were identical and followed those of study 1.

control in this study. Half of the subjects were tested over sessions in the order sham-field; the remaining subjects were tested in the reverse order. The study was performed double blind in the exposure facilities at MRI. Hourly blood samples were collected during sham and field- exposure test sessions for assay of melatonin concentration. Data were statistically analyzed using ANOVA for repeat measures and multiple-regression techniques.

Subjects. Volunteers participated in this study with the approval of the Human Subjects Committee at MRI in accordance with federal guidelines and regulations. Sub- ject recruitment and informed consent procedures were the same as in study 1 . Complete data on basal peak STATISTICAL ANALYSIS melatonin level and light sensitivity were collected dn 52 Statistical analysis tested the hypothesis that men with men. Of these, the final test sample that completed all study preexisting low levels of melatonin show greater sup- requirements consisted Of 40 Young (mean = 22 Years; range, 8-32 years), white (95%), nonsmoking (90%) men Of average height (70*8 inches; range, 66-77 inches) and weight (175*8 pounds; 38-270 pounds) who typi-

pression of melatonin when exposed at night to bright light or to intermittent magnetic fields at 200 mG. Hourly melatonin values were entered into an ANOVA with order of exposure (field-sham, sham-field) and basal peak

tally slept 7-8 h per night. Characteristics of the screen- melatonin level as between-subject variables and with ing Of 52 men did not differ from those Of the field condition as a within-subjects variable. The &pen- sample. Data collection was not completed on the remaining men for a variety of reasons (e.g., scheduling conflicts, failure to keep appointments, problems during blood draws, illness, etc.).

Measures. Like study 1, hourly melatonin levels measured in plasma provided the primary outcome measure of the study; however, the Elias assay used in study 1 was no longer available for use in this study. Conse- quently, all blood samples collected were assayed blind by Professor Mark Rollag in his laboratory, using his sensitive extraction assay for melatonin [Rollag and Niswender, 19761. The results of these two assays are essentially interchangeable for melatonin in the range tested (the correlation between assays was .99 over 80 blind samples in the range from 6 to 200 pg/ml). Both subjects and experimenters, as before, independently completed the FSQ to assess the effectiveness of the double-blind control procedures.

PROCEDURES

dent variable was the melatonin level at each hourly blood sample, Alpha was set at

Data from the screening session were used to test the hypothesis that exposure to bright light results i n suppression ofn-~elatonin that is ProPortionatelY greater for men with low melatonin levels than for men with high levels. Multiple regression and Correlational techniques were used for this Purpose. Where ANOVA was used, Probability values Were corrected by using the HuYnh- Feldt epsilon technique. Probability levels reported here are the corrected ones; the degrees of freedom are not corrected in order to indicate the actual number of data Points included in the statistical tests.

RESULTS

Control measures. Analysis of the FSQ rating data indicated that the subjects were unable to judge whether the fields were on at better than chance levels (X’, 1.27; df, 1; P > .26). This was true even when subjects had strong confidence in the accuracy of their judgments. Experimenters also were uyable to judge field status at better than chance levels (X-, 0.80; df, 1; P > .37). These results confirm the effectiveness of the double-blind control procedures.

Melatonin concentration

Preexposure screening. Screening data were available for 52 subjects. Again, a broad range of individual peak melatonin levels was observed. Melatonin levels at 0200 h ranged from 15 to 122pg/ml. Light exposure had the expected suppressant effect. At 0200 h, mean melatonin level was 50.9 pg/ml; after 1 h of light exposure,

< .05.

In general, procedures followed those of study 1. Subjects in this study, however, participated in one screening session rather than two. Given the high degree of con- sistency observed earlier in individual melatonin levels, one session was deemed adequate for screening purposes. In the present study, groups of three or four subjects participated in each screening session. The men slept in the dark from 2230 h to 0200 h, at which time a blood sample was obtained following the procedures used in study 1 . After the blood draw, the men were taken to the light-exposure area, where they sat for 60 min in front

270 Graham et al.

melatonin levels were reduced to 22.3 pg/ml (t = 12.16; df, 51; P c .0001). Again, a range of light-suppression effects on melatonin was found. Percent suppression ranged from 20 to 8096, with the majority of subjects showing reductions of 40-70% in melatonin levels.

In our initial study, a significant negative correlation was found between basal peak melatonin level and percent suppression by light. To determine whether this was also true in the present study, a logarithmic transformation was performed on the basal level scores as before to correct for nonlinearity, and Pearson’s r was computed between the two variables. The correlation was .323 (P c .02), indicating that, in the present study, the relationship between peak melatonin level and the amount of melatonin suppression observed after controlled exposure to light was significant and was the opposite of that previously found. Experimental testing. Complete data were obtained for 40 men over the two test sessions. Figure 3 shows mean levels and standard errors of melatonin for these subjects every hour from midnight to 0700 h for both sham and field exposure sessions. The distributions seen under sham and field exposure were very similar, and there was a high degree of overlap in the variances. As i n the initial study, sham-control and field-exposure nights were not significantly different from one another.

To determine whether men with preexisting low levels of melatonin were significantly affected by field exposure, subjects were split at the median into low- melatonin (less than or equal to 48 pg/ml) and high- melatonin (> 48 pg/ml) groups. The low and high

groups, as expected, differed significantly in melatonin levels (F = 10.15; df, 1,35; P = .003). Melatonin values also changed over the night as expected, peak- ing at 0300 h (F = 28.26; df, 7,245; P < ,0001). A significant interaction between melatonin group and time of night was found (F = 4.67; df, 7,245; P < .009). High- melatonin subjects had a steeper rise time than low- melatonin subjects, although the groups peaked at the same time, and melatonin levels decreased at about the same rate.

We specifically tested the hypothesis that men with low basal peak melatonin would show lower melatonin levels as a function of field exposure. A statistical trend was found for a group by level interaction (F = 2.27; df, 7,245; P = ,065). Contrary to the hypothesis, the changes were found in men with high basal peak melatonin. Melatonin levels in the high group were slightly higher after 0300 h on field-exposure nights than on sham- exposure nights. When analyzed separately, however, neither the low melatonin group nor the high melatonin group showed significant effects of field exposure. The hypothesis was, therefore, not supported. Figure 4 shows melatonin levels over the night for men in the low melatonin group.

Subjects in the current study had lower peak melatonin values (median 48 pg/ml) than subjects in the initial study (median, 60 pg/ml). Therefore, we divided the present study group into those subjects with peak melatonin levels less than or equal to 60 pg/ml and those with peak melatonin greater than 60 pg/ml. The same basic results were found. Subjects were then divided into

1 0 0 ,

Field Sham t

12 1 2 3 4 5 6 7 Time (a.m.)

Fig. 3. Comparison of nocturnal melatonin levels in 40 men under sham control vs. exposure to 200 mG intermittent magnetic fields. Mean melatonin level and standard error are plotted for each hour from midnight to 0700 h. Melatonin levels on sham-control and field- exposure nights did not differ.


100

90

80

4 70 a P

60 C

50 m 2 40 C

3 30 I

20

10

0

n -

.- -

Field Sham

12 1 2 3 4 5 6 7

Time (a.m.)

Fig. 4. Lack of magnetic field effects in 20 men with preexisting low melatonin levels (<48 pg/ml at 0200 h). Mean melatonin level and standard error are plotted for control and exposure conditions at each hour from midnight to 0700 h. Melatonin levels on sham-control and field-exposure nights did not differ.

extreme groups (those with basal peak melatonin greater than 60 pg/ml and those with basal peak melatonin less than or equal to 41 pg/ml). Again, the results did not support the hypothesis. Finally, subjects were divided into two groups based on whether their melatonin suppression to light was above or below the median of 44%, and again, the results were negative.

DISCUSSION

Field-related suppression of nocturnal levels of the hormone melatonin has been hypothesized as a possible biological mechanism linking occupational and residential exposure to magnetic fields and increased cancer risk. Results of the two studies reported here do not support this hypothesis for the field conditions used; no significant differences between sham control and field exposure were found. Results also failed to support the hypothesis derived from study 1 that men with preexisting low levels of melatonin show significantly greater suppression of melatonin when exposed to light and to the 200 mG magnetic field condition.

Analysis of the screening data obtained in both studies, however, revealed large differences between people in their preexposure melatonin levels and in their responsivity to the suppressant effects of exposure to bright light at night. The idea that such differences may

be predictive of the relative degree of sensitivity an individual exhibits when exposed to power-frequency magnetic fields is intriguing and should continue to be investigated. At the least, since melatonin is involved in multiple physiological interactions, future research should determine whether individuals with high or low preexposure melatonin levels show markedly different hormonal and immune profiles. The present studies may not have provided the best situation in which to examine the interaction between melatonin base level and sensitivity. The volunteers were young, and the age distributions were restricted (80% of the men in study 2 were 18-22 years of age). This, together with differences across studies in the level and distribution of preexposure melatonin values, may have masked the possibility of observing a relationship. Ideally, future research will examine the interaction between melatonin level and sensitivity over a broader age range in both men and women.

The present studies used exposure conditions that were shown in earlier work to produce biological re- sponses [Graham et al., 19901. Given the paucity of studies in this area, however, arguments could as eas- ily have been made for the selection of other exposure conditions. To state with confidence that melatonin levels in humans are unaffected by magnetic field exposure, it is necessary to examine the effects of exposure to fields with other characteristics.

272 Graham et al.

Of importance is the examination of continuous exposure and of exposure to magnetic fields with high harmonic content and transient fields. In an industrial- ized society, magnetic field exposures from power distribution systems consist of two components: the 50 or 60 Hz fundamental power frequency that is present at all times and “transient” events that result from normal utility operations, such as switching events, opening or closing of‘ capacitor banks, etc. Transients can also result from the switching on and off of loads, whether they are residential or industrial. Recently, transients with a high-frequency content (ranging from several kilohertz to 10 megahertz) have been measured in homes [Guttman et al., 19941.

Biophysical calculations using realistic models of cells have shown that some of these transient events can induce transmembrane voltage changes in model cells that exceed thermal noise [Sastre et a]., 19941. There- fore, in principle, these events can affect biological systems. Lerchl and his colleagues [1991] suggested that transients were possibly involved in their studies of melatonin suppression in the pineal of rodents. The effects of these fields, which are relevant both to power distribution systems and to the use of common electri- cal appliances, need to be examined in humans. Simi- larly, brief daytime exposure as well as longer night-time exposure has been reported to produce alterations in nocturnal melatonin levels in rodents [Yellon, 19941. It is important to determine whether daytime exposure alters nocturnal levels of melatonin in humans.

The melatonin hypothesis [Stevens, 19871 holds that field-related suppression of melatonin acts indirectly on cancer risk by modulating the secretion of circulat- ing hormones implicated in carcinogenesis. This hypothesis is most directly relevant to breast cancer. The bulk of the epidemiological evidence, however, implies that EMF exposure may increase the risk of leukemia, lymphoma, and brain cancer. One possible mechanism with implications for this broader range of disease cat- egories stems from recent reports showing that melatonin serves as a potent scavenger of free radicals, thus reduc- ing the incidence of all types of cancer [Hardeland et al., 1993; Tan et al., 1993; Pieri et al., 1994; Reiter et al., 19951.

At physiological levels, melatonin is also known to stimulate the immune system [Maestroni, 19931, and either surgical or functional pinealectomy affects major components of the immune system. Melatonin’s effects on estrogen are well known, and estrogen itself is a potent immunomodulator with receptors expressed in many immune tissues. No laboratory studies have included women or have determined whether there are important differences in sensitivity as a function of age or sex or in women as hormonal changes occur over the

menstrual cycle. Thus, the relationship between melatonin, hormonal, and immune system activity may well be a fruitful area for further research. It is important that as much of this research as possible be carried out under controlled experimental conditions and, where fea- sible, that it involves human volunteers. Humans differ from animals in many ways, and only data from human beings are directly relevant, without extrapolation, to risk assessment.

ACKNOWLEDGMENTS

This research was supported by Electric Power Research Institute contract W04307. Preparation of this report was supported in part by National Institute of Environmental Health Sciences grant ES07053 to the senior author. Portions of this research have been presented previously at scientific meetings [Graham et al., 1993a,b, 1994al and in technical reports to the sponsor [Graham et al., 1994bl. The authors thank the members of the project team who helped make this study possible: nurses Sherri Price and Nancy E. Phelps; Deb Smith (bioassays); programmer Steve Hoffman; and data manager F. Joseph McClernon. We also acknowledge the many helpful comments and supportive actions of Robert Black and Dr. Rob Kavet, EPRI Project Officers for this research activity, and of Drs. Mark Rollag, George Brainard, and Russel Reiter.

REFERENCES

Cohen HD, Graham C, Cook MR, Phelps JP ( I 992): ELF exposure facility for human testing. Bioelectromagnetics 13: 169-182.

Coleman M, Beral V (1988): A review of epidemiological studies of the health effects of living near or working with electric- ity generation and transmission equipment. Int. J. Epidemi-

Cook MR, Graham C, Cohen HD, Gerkovich MM (1992): A repli- cation study of human exposure to 60-Hz fields: Effects on neurobehavioral measures. Bioelectromagnetics 13:26 1-285.

Federal Register (1991): Federal policy for the protection of human subjects: Notices and rules, Vol 56, No 117. June 18, 1991.

Graham C, Cook MR, Cohen HD (1990): Final Report: Immunological and Biochemical Effects of 60-Hz Electric and Magnetic Fields in Humans, Report DE90006671. Springfield, VA: National Technical Information Center.

Graham C, Cohen HD, Cook MR, Riffle DW, Hoffman SJ, Gerkovich MM ( 1993a): Melatonin suppression and nocturnal exposure to 60-Hz magnetic fields in human volunteers. Abstracts, 13th Annual Meeting, Bioelectromagnetics Society, Los Angeles, p 67.

Graham C, Cook MR, Cohen HD, Riffle DW, Hoffman SJ, McClernon FJ, Smith D, Gerkovich MM (1993b): EMF suppression of nocturnal melatonin in human volunteers. Abstracts, DOE Annual Review of EMF Research, Savannah, GA, pp 98-99.

Graham C, Cook MR, Cohen HD, Gerkovich MM (1994a): A dose response study of human exposure to 60-Hz electric and magnetic fields. Bioelectromagnetics 1 5:447-463.

ology 17:1-13.


Reiter RJ (1993): A review of neuroendocrine and neurochemical changes associated with static and extremely low frequency electromagnetic field exposure. Integ Physiol Behav Sci 2857-75.

Reiter RJ, Melchiorri D, Sewerynek E, Poeggler B, Barlow-Walden L, Chuang J-I, Ortiz GG, Acuna-Castroviejo D (1995): A review of the evidence supporting melatonin’s role as an antioxidant. J Pineal Res I 8: 1 - 1 1.

Rollag MD, Niswender GD (1976): Radioimmunoassay of serum concentrations of melatonin in sheep exposed to different lighting regimens. Endocrinology 98:482-489.

Sastre A, Kavet R, Guttman JL, Weaver JC ( I 994): Residential magnetic field transients: How do their induced transmembrane voltages compare to thermal noise? Abstracts, DOE Annual Review of EMF Research, Albuquerque, NM, pp 33-34.

Savitz DA, John EM, Kleckner RC (1989): Magnetic field exposure from electric appliances and childhood cancer. Am J Epidemiol

Stevens RG (1987): Electric power use and breast cancer: A hypothesis. Am J Epidemiol 125556-561.

Tan D, Poeggeler B, Reiter RJ, Chen L, Chen S, Manchester LC. Barlow-Walden LR (1993): The pineal hormone melatonin inhibits DNA-adduct formation induced by the chemical car- cinogen safrole in vivo. Cancer Lett 70:65-71,

Theriault G (1991): Health effects of electromagnetic radiation on workers epidemiologic studies. “Scientific Workshop on the Health Effects of Electromagnetic Radiation on Workers”, Report No. 9 1-37. Cincinnati, OH,: NIOSH.

Waldhauser F, Walentich W, Ehrhart B (1993): The physiological secretion pattern of melatonin in man. In Touitou Y, Arendt J , Pevet P (eds): “Melatonin and the Pineal Gland: From Basic Science to Clinical Application.” New York: Elsevier Science Publishers, pp 331-338.

Wilson BW, Chess EK, Anderson LE (1986): 60-Hz electric-field effects on pineal melatonin rhythms: Time course for onset and recovery. Bioelectromagnetics 7:239-242.

Wilson BW, Stevens RG, Anderson LE (eds; 1990): “Extremely Low Frequency Electromagnetic Fields: The Question of Cancer.” Columbus, OH: Battelle Press.

Yellon SM, Hilliker S (1994): The influence of acute melatonin treat- ment and light on the circadian melatonin rhythm in the Djungarian hamster. J Biol Rhythm 9:71-81.

Zimmermann RC, Schumacher M, Schroder S, Weise HC, Baars S (1990): Melatonin and the ovulatory luteinizing hormone surge. Fertil Steril 54612-618.

13 1~763-773.

Graham C, Cook MR, and Cohen HD (1994b): Investigation of the effects of magnetic field exposure on human melatonin. Electric Power Research Institute, Report TR- 104278.

Graham C, Cook MR, Cohen HD, Riffle DW (1994~): Nocturnal melatonin levels in men exposed to magnetic fields: A repli- cation study. Abstracts, DOE Annual Review of EMF Research, Albuquerque, NM, pp 5 1-52.

Guttman JL, Zaffanella LE, Johnson GB (1994): Survey measurements and experimental studies of residential transient magnetic fields. EPRI TR-104532, Dec, 1994.

Hardeland R, Reiter RJ, Poeggeler B, Tan D-X (1993): The significance of the metabolism of the neurohormone melatonin: Antioxidative protection and formation of bioactive substances. Neurosci Biobehav Rev 17:347.

Hill SM, Blask DE (1988): Effects of the pineal hormone melatonin on proliferation and morphological characteristics of human breast cancer cells (MCF-7) in culture. Cancer Res 48:612 1 - 6126.

Kato M, Honma K, Shigemitsu T, Shiga Y (1993): Effects of exposure to a circularly polarized 50-Hz magnetic field on plasma and pineal melatonin levels in rats. Bioelectromagnetics 14:97- 106.

Lerchl A, Nonaka KO, Stokkan KA, Reiter RJ (1990): Marked rapid alterations i n nocturnal pineal serotonin metabolism in mice and rats exposed to weak intermittent magnetic fields. Biochem Biophys Res Commun 169:102-108.

Lerchl A, Nonaka KO, Reiter RJ (1991): Pineal gland: Its apparent magnetosensitivity is a consequence of eddy currents. J Pi- neal Res 10:109-116.

Lewy AJ, Wehr TA, Goodwin FK, Newsome DA, Markey SP (1980): Light suppresses melatonin secretion in humans. Science 210: 1267-1269.

Maestroni GJM (1993): The immunoneuroendocrine role of melatonin. J Pineal Res 14:l-10.

Pieri C, Marra M, Moroni F, Recchioni R, Marcheselli F (1994): Melatonin: A peroxyl radical scavenger more effective than vitamin E. Life Sci 55:271.

Reiter RJ (1 988): Pineal gland, cellular proliferation and neoplastic growth: An historical account. In Gupta D (ed): “The Pineal Gland and Cancer.” Tubingen, Federal Republic of Germany: Brain Research Promotion, pp 41 -64.

Reiter RJ (1989): The pineal and its indole products: Basic aspects and clinical applications. In Cohen MP, Foley PP (eds): “The Brain as an Endocrine Organ.” Vienna: Springer, pp 96-149.

Documents

Nocturnal melatonin levels in human volunteers exposed to intermittent 60 Hz magnetic fields