Content uploaded by C-K. Chou
Author content
All content in this area was uploaded by C-K. Chou on Mar 22, 2020
Content may be subject to copyright.
Bioelectromagnetics 13:469-496 (1 992)
Long-Term, Low-Level Microwave
Irradiation
of
Rats
C.-K. Chou,
A.W.
Guy,
L.L.
Kunz,
R.B.
Johnson, J.J.
Crowley,
and J.
H.
Krupp
Bioelectromagnetics Research Laboratory, Center for Bioengineering (C.
K.
C.,
A.
W.
G.,
L.
L.
K.,
R.
B.
J.),
and Department of Biostatistics
(J.
J.
C.),
University
of
Washington, Seattle;
USAF
School of Aerospace Medicine, Aerospace Medical
Division, Brooks Air Force Base, Texas
(J.
H.
K.)
Our goal was to investigate effects of long-term exposure
to
pulsed microwave radia-
tion. The major emphasis was to expose a large sample of experimental animals throughout
their lifetimes and to monitor them for effects on general health and longevity.
An exposure facility was developed that enabled 200 rats to be maintained under specific-
pathogen-free
(SPF)
conditions while housed individually in circularly-polarized
waveguides. The exposure facility consisted of two rooms, each containing 50 active
waveguides and
50
waveguides
for
sham (control) exposures. The experimental rats were
exposed to 2,450-MHz pulsed microwaves at
800
pps with a
10-ps
pulse width. The pulsed
microwaves were square-wave modulated at 8-Hz. Whole body calorimetry, thermographic
analysis, and power-meter analysis indicated that microwaves delivered at 0.144
W
to
each exposure waveguide resulted in an average specific absorption rate (SAR) that ranged
from 0.4
W/kg
for a 200-g rat to
0.15
W/kg for an
800-g
rat.
Two hundred male, Sprague-Dawley rats were assigned in equal numbers to radia-
tion-exposure and sham-exposure conditions. Exposure began at
8
weeks of age and con-
tinued daily, 21.5 h/day, for 25 months. Animals were bled at regular intervals and blood
samples were analyzed for serum chemistries, hematological values, protein electrophoretic
patterns, thyroxine, and plasma corticosterone levels. In addition to daily measures of
body mass, food and water consumption by
all
animals,
O2
consumption and C02 pro-
duction were periodically measured in a sub-sample (N=18) of each group. Activity was
assessed in an open-field apparatus at regular intervals throughout the study. After I3
months,
10
rats from each group were euthanatized
to
test for immunological competence
and to permit whole-body analysis, as well as gross and histopathological examinations.
At the end of 25 months, the survivors
(1
1
sham-exposed and
12
radiation-exposed rats)
were euthanatized for similar analyses. The other
157
animals were examined
histopathologically when they died spontaneously or were terminated
in
extrernis.
Received for review November
IS,
1991; revision received September 29, 1992.
Dr.
Chou’s
present address is Department
of
Radiation Research, City
of
Hope National Medical Center,
Duarte, CA 91010. Address reprint requests there.
L.L. Kunz’s present address is NeoRx Corporation, 410 West Harrison, Seattle, WA 981 19.
J.H.
Krupp’s present address
is
Systems Research Laboratories,
P.O.
Box
35505,
Brooks Air Force
Base,
TX
78235.
0
1992 Wiley-Liss,
Inc.
470 Chou
et
al.
Statistical analyses by parametric and non-parametric tests
of
155
parameters were
negative overall
for
effects on general health, longevity, cause of death, or lesions asso-
ciated with aging and benign neoplasia. Positive findings of effects
on
corticosterone level
and immune system
at
13
months exposure were
not
confirmed in a follow-up study
of
20
exposed and
20
control rats. Differences in
0,
consumption and C0,production were
found in young rats.
A
statistically significant increase of primary malignancies
in
ex-
posed rats vs. incidence in controls is a provocative finding, but the biological signifi-
cance
of
this effect in the absence
of
truncated longevity is conjectural. The positive findings
need independent experimental evaluation. Overall, the results indicate that there were
no definitive biological effects in rats chronically exposed to
RF
radiation at
2,450
MHz.
0
1992
Wiley-Liss.
Inc.
Key words:
SAR,
longevity, health,
tumor
incidence
INTRODUCTION
Advances in dosimetry, and a better understanding of energy absorption by
biological tissues, have eliminated many concerns regarding effects of radio-fre-
quency electromagnetic radiation [Tyler, 1975; Elder and Cahill,
1984;
NCRP,
1986;
Polk
and Postow, 1986; Lin,
1989;
Gandhi,
19901.
Despite this lessening of con-
cern for low-level, acute exposures, lack of data
on
long-term, low-level radiation
has fueled public and scientific concerns. In this context, officials
of
the United
States Air Force sought to support research in this area to provide data for use in
the development of environmental impact studies for present and planned Air Force
systems.
The goal of the project was to investigate effects
on
health of long-term ex-
posure to low-level, pulsed microwave radiation. The approach was
to
expose a large
population of experimental animals to microwave radiation throughout most of their
lifetimes and to monitor them for effects on general health and longevity.
Although the initial impetus for the study was the question of environmental
impact of the Air Force
PAVE
PAWS system, early
on
it was decided not to study
a replica of the PAVE PAWS emissions, but to create a generalized level of radia-
tion that would provide whole-body exposure based on the maximum of permis-
sible absorption
[ANSI
C95.1-
1982,
1983;
IEEE
C95.1-1991,1992] at the resonant
frequency in human beings
(0.4
W/kg),
as scaled to the proportions of the experi-
mental animal
of
choice.
Following a period of pilot studies and training of technicians, exposures to
microwaves commenced on September
I,
1980,
and concluded September 27,1982.
The
100
experimental and 100 sham-exposed animals underwent the longest near-
continuous exposure ever completed. The findings were reported in a series of
9
Air
Force technical reports, which are available through the National Technical
Information Service (Springfield, Virginia). Interested readers should refer to the
technical reports for details [Guy et al., 1983a,
b,
1985;
Chou et al, 1983; Johnson
et
al., 1983, 1984; Kunz
et
al., 1983,
1984,
19851.
METHODOLOGY
Experimental Design
Exposure criteria.
Much
of
the past work
on
chronic exposure of large numbers
of text animals has been based on anechoic chambers, metal capacitor plates, or
Long-Term
Low-Level
RF Exposure
471
resonant cavities. With these methods, the energy coupled to each animal is a function
of the group size, group orientation, and the orientation of each animal within the
group, as well as of the presence and location of water and food dispensers. Be-
cause estimates of energy absorption are uncertain, quantitative extrapolation
of
biological results from laboratory animals to human beings is virtually impossible.
In addition, the cost in time and resources of even simple experiments involving
chronic exposures of animal populations in large anechoic chambers is prohibitive.
For this study, we chose a system of cylindrical, wire-mesh waveguides to expose
a large number of animals to a common source while independently maintaining
relatively constant and quantifiable coupling of electromagnetic energy to each animal
regardless of position, posture, or movement [Guy and Chou, 1977; Guy et al., 19791.
The system, which consists of a number of independent waveguides, allows indi-
vidual animals to be continuously exposed under normal laboratory conditions while
living unrestrained and with continuous access to food and water.
A
frequency of 2450 MHz was selected
so
each rat would have approximately
the same size-to-wavelength ratio as a human being exposed at 450
MHz,
the fre-
quency near which PAVE PAWS operates. The initial consideration was to produce
the same average SAR in test animals as predicted for man exposed to a l-mW/
cm2
,
450-MHz RF field.
To
simulate radar exposure, pulse modulation was used
(10
ps
pulse, 800 pps). In addition
to
the pulse modulation, we decided to square-
wave modulate the microwaves. The inclusion of square-wave modulation was
prompted by the evidence of altered movement of Ca++ ions in chicken and cat brains
exposed to ELF-modulated RF fields [Adey, 198 11. Because the demonstrated effects
are most pronounced when the modulation frequencies correspond to the dominant
EEG frequency, we selected a modulation frequency of
8
Hz because it is at the
peak of the rat’s hippocampal theta rhythm [Coenen,
19751.
Rationale
of
biological assessment. Not only were reported biological ef-
fects from low-level microwaves selected as end points (e.g., alterations of hemato-
poietic, immunologic, and specific blood chemistry indices), but assays for effects
on general health, metabolism, and life span were also included (references listed
in later sections). In addition, end points were considered that could be assessed
without seriously compromising the health of the animal, the value of concurrent
measurements, or the power of the statistical evaluations on the chosen end points.
Only male rats were used to minimize statistical variation, i.e., to avoid the hor-
monal variations characteristic of female rats. Use of female rats would have re-
quired a substantial increase in the number of animals. A total of 155 parameters
was studied. The end points selected are shown in Table
1.
Due to space limitations,
details
of
rationale and methods of biological assessments cannot be provided here.
The original
NTIS
reports should be consulted.
Statistical considerations. For any failure-time end point, such as time to death,
time to cancer diagnosis, or time to some specified change in animal mass or blood
chemistry, an initial sample size of 100 in each group was calculated to be suffi-
cient for detection, at the .05 significance level, of a 50% increase (or
33%
reduc-
tion) in instantaneous failure rate with a probability (power)
of
90%.
For any normally
distributed end point (including transformations on failure-time variables), a sample
size of
100
per group permits the detection, at the .05 level of significance, of a
difference between groups of 40% of one standard deviation, with a power
of
90%.
Adjustment for a differential effect due to the altered experimental procedure for
472
Chou
et
al.
TABLE
1.
Endpoints Selected to Study the Effects of Long-Term, Low-Level
RF
Exposure on
Rats
~~ ~ ~ ~ ~~ ~~ ~
Category Parameters No.
Behavior
Corticosterone
Immunology
Hematology
Blood chemistry
Protein
electrophoresis
Thyroxine
Urinalysis
Metabolism
Total body
analysis
Organ mass
Histopathology
Longevity
Total
Open field behavior (activity, quadrant change, urination, defecation)
Serum corticosterone
Mitogen stimulation (PHA,LPS,ConA,PWM,PPD), B-cell, T-cell,
%CRPC, total CRPC, plaque
WBC, RBC, HCT, Hgb, MCV, MCH, MCHC, neutrophils,
lymphocytes, eosinophil, monocytes
Glucose,
BUN,
creatinine, Na,
K,
CI, CO,, uric acid, total bilirubin,
direct bilirubin, Ca, phosphorus, alkaline phosphatase, LDH,
SGOT,
SGPT, cholesterol, triglycerides, total protein, albumin, globulin
Albumin fraction, alpha-l and
2
fractions, beta fraction, gamma
fraction
Thyroxine
Urine observation
Body mass, food consumption, water consumption,
0,
consumption,
CO, production, respiratory quotient, metabolism quotient
Body mass, moisture, protein nitrogen, crude fat, nonprotein nitrogen,
total ash, mineral contents (aluminum, antimony, arsenic, barium,
beryllium, bismuth, boron, cadmium, calcium, chromium, cobalt,
copper, iron, lead, magnesium, manganese, molybdenum, nickel,
phosphate, potassium, selenium, silver, sodium, strontium,
tin,
titanium, vanadium), fatty acids (palmitic, palmitoleic, stearic,
oleic, linoleic, linolenic)
Heart, brain, liver, kidneys, testicles, adrenals
All tissues and organs
Survival days
4
1
1
0
11
21
4
1
I
7
39
9
46
I
155
the
subset
of
36
rats subjected to metabolic rate measurements had very little ef-
fect on the power calculations made, nor did adjustment for an interim euthanasia
of
20
animals.
Differences between the two groups on single measurements were assessed
by Student’s
t
tests, in some cases after transformation to improve the normality
of the data. Reported
P
values must be considered in the light
of
the multiple end
points analyzed. Logical groupings of variables were compared across groups of
the multivariate Hotelling’s
T?
statistic.
Differences in tumor prevalence or incidence were assessed with time-adjusted
analyses. The occurrence of neoplastic and non-neoplastic lesions was recorded along
with the age of the animal and whether the animal had died spontaneously or was
euthanatized. Survival curves of the exposed and sham-exposed animals were es-
timated by product-limit estimates [Kaplan and Meier,
19581
and compared by the
log-rank statistic [Mantel,
19661.
The histopathological data were grouped with respect
to age, at 6-month intervals, and the data were divided into neoplastic and non-
neoplastic diagnoses. The incidence of neoplastic
or
non-neoplastic lesions was given
as the proportion of the number
of
animals bearing such lesions at a specific ana-
tomic site (numerator)
to
the number
of
animals examined pathologically (denomi-
nator). For tissues that required gross observation for detection of lesions (i.e., skin
or subcutaneous tumors), for lesions that appeared at several sites (i.e., multiple
Long-Term Low-Level
RF
Exposure
473
lymphomas), or for tissues that were examined histologically only when lesions were
detected grossly, the denominator consisted of the number of animals necropsied
in that experimental group.
The analysis of the lesions involved a 4-way table with factors of age at death,
treatment condition, mode of death (terminated or spontaneous), and organ. The
tables were then collapsed with respect to individual organs.
From
these tables, the
Mantel-Haenszel estimate of the odds ratio was computed, and the chi-square sta-
tistic was used to test whether the odds ratio was significantly different from unity
[Mantel and Haenszel,
19591.
This statistic reflects the difference in prevalence of
lesions, over time, between the exposed and sham-exposed animals, and is appro-
priate if the lesions are “incidental” (do not affect the animal’s survival). The time
to a malignant lesion was also analyzed with survival-analysis techniques, as would
be appropriate if lesions were fatal. If an animal had malignant lesions, its time-
to-tumor was taken as its survival time. If there were no malignant lesions present,
the time-to-tumor was considered censored (i.e., the time
to
appearance of a tumor
is assumed
to
be longer than the time to death). The log-rank statistic was used to
compare the times to tumor of the exposed animals with those of the sham-exposed
animals [McKnight and Crowley,
19841.
Final protocol.
During the first year of the study, the rats were bled from the
orbital artery every
6
weeks, with the first bleeding during the 7th week of expo-
sure. In addition to the hematological and serum-chemistry evaluation of blood
collected during the first bleeding, corticosterone levels were determined in all samples
having adequate amounts of serum. In subsequent bleedings, corticosterone and
thyroxine levels were determined only quarterly, whereas the hematology and se-
rum chemistry were evaluated for each sample (every
6
weeks). This frequency of
bleeding was considered sufficient
to
detect the onset of most degenerative or disease
states that would occur during the lives of the individual rats without unduly stressing
the animals. Every
3
months a urinalysis was done on all rats, the first during the
fourth week of exposure. This frequency of biochemical evaluations increased the
opportunity to detect subclinical abnormalities and to follow their pathophysiological
course. Open-field assessment was conducted every
6
weeks.
During the second year of the study, the frequency
of
bleeding was reduced
to 12-week intervals, and the corticosterone analysis was eliminated except just prior
to euthanasia of remaining animals at the end of the 2 years; urinalysis was done
every
2
weeks, and open-field analysis was conducted quarterly.
Facilities
Animal facility.
To
maintain the colony
of
rats used in this study in the healthiest
possible state, free of chronic disease and other problems common to rats, two specific-
pathogen-free (SPF) rooms in the Division of Animal Medicine were acquired (Fig.
1).
Access to the clean hall is via a shower room, through which all personnel must
pass to shower and don autoclaved garments.
A
walk-in autoclave connected the
cage-washing facility with the clean hallway
so
that, once washed, all materials
entering the clean hall must have passed through autoclaving before being returned
to the animal rooms. All soiled cages and waste collectors left the clean rooms via
the dirty hallway and were then taken to the cage-washing facility.
Each alcove housed
20
waveguides mounted on four horizontal shelves, five
waveguides per shelf. The exposure and sham-exposure waveguides were randomly
474 Chou
et
al.
Dirty
U-#I
rIUII
4
Clean
Ha//
Power Distriknm
for
SPF
Facilities
Fir
I.
Tor,
view
of
exnosure
rooms
(T-I31
and
T-137)
with
alcove
designations and associated
ex-
-
pure-cell
identitication
system.
arranged, except that only sham-exposure waveguides could be
in
the center posi-
tion because of a sliding-glass-door operation. One
of
the alcoves in each room was
equipped as a metabolism alcove, in which 0,consumption and
CO,
production were
measured. The sixth alcove in each room was partitioned off as
a
procedures area
that was used for bleeding and as housing for the main data-collection computer
and miscellaneous supplies.
In
the
SPF rooms the airflow rate was programmed for 22 exchanges each hour,
to maintain positive-pressure flow. Over the course of the project, ambient tempera-
tures were balanced between the workspaces and alcoves to maintain
a
fairly con-
stant
21kl"C
environment in the facility. Humidity was in the range of 30-70%.
Sound-pressure measurements indicated an average level
in
the central workspace
of approximately 60 dBA (relative to 20 ,uN/cm') and alcove levels that were 6 to
10 dBA lower, depending on position within the alcove. Light-intensity measurc-
ments during the light cycle (0700-1900) indicated a 13-lux average workspace
level
and 6-lux average alcove level.
Microwave
exposure
system.
As
shown in Figure
2,
when an animal housed
in
a plastic cage was exposed in the circular waveguide to microwaves fed into the
terminal (PJ, some energy
(PA)
was absorbed by the animal, some (P,) was ab-
sorbed by the walls of the chamber, some was reflected
in
the form of both right-
hand (P,,) and left-hand (PRL) circularly polarized waves that couple back to the
probes on the feed section
of
the waveguide, and some (P,,and PTJ were absorbed
at the termination terminals. The reflected component P,, was measured as
CP,,
at the reflecting arm of the bidirectional coupler, which was placed between the
source and the input probe
(C
is the coupling coefficient of the bidirectional cou-
pler). The reflected component
p,,
was measured directly at the other terminal
of
the transmitting transducer. The power level of the incident energy launching the
right-hand circularly polarized waves was measured (as
CP,,)
at the incident-wave
terminal of the coupler. The power level of energy transmitted beyond the animal
was measured at
the
terminals
(PTA
and PT,)
of
the termination transducer. The sum
Long-Term Low-Level
RF
Exposure
475
WATER
aOTTLE4
PTA
pW
PTB
J
Fig. 2. The exposure chamber was a circularly polarized waveguide operating at 2,450
MHz.
A
rat,
housed inside
a
plastic cage, was exposed in the 20.3-cm diameter wire-mesh tube.
A
circular polarizer,
at the
left
end, converted the linearly-polarized TE,
,
mode
to
the circularly-polarized
TE,,
mode. Tuning
stubs inside the polarizer matched the impedance
of
the propagating modes. Transmitted microwaves
were terminated in thc transducer at the right side
of
the tube.
of
power levels of energy absorbed by the animal and the chamber walls can be
obtained from the equation in Figure
2.
The water bottle
in
each waveguide was electrically decoupled from the ani-
mal by two concentric 1/4-wavelength choke sections
so
that the tip
of
the water
nozzle had an extremely high impedance, virtually preventing conduction currents
between it and any contacting object. The theory
of
the waveguide operation has
been described elsewhere [Guy et al.,
19791.
Microwave generation and distribution.
Each exposure room was equipped
with two
2,450-MHz
pulsed microwave sources (Epsco, model
PGSKB,
Trenton,
NJ),
each source capable
of
providing an average output power
of
20
W
and a peak
power
of
5
kW.
These generators were controlled by a microprocessor to deliver
repetitive pulse trains
as
shown in Figure
3.
kl250ps
4
50
Pulses
r.5
I25
ms
ms
dl
Fig.
3.
Modulation characteristics
of
the microwave pulses:
8
groups per second,
50
10-ps-wide pulses
per group, with a repetition rate
of 800
pps. The period was 125
ms;
with
pulse onsets separated by
1.25-ms intervals. This is
the
equivalent
of
an
800-pps
source square-wave modulated at
8
Hz.
476 Chou
et
al.
MICROWAVE
CIRCUITRY for LOW DUTY CYCLE
HIGH
PEAK POWER
CHRONIC EXPOSURE SYSTEM
with
TYPICAL OPERATING POWER LEVELS
+k
COURER
METER
/\i35OW PEAK
I4
W
AVG
DETECTOR
PUCSED
SIGNAL
3-K]
io-1
t.
t
J.
f
v)
3
E
8
*
POWER
LEVELS
CAN BE
INCKASED Br
A
FACTOR
AS
HIGH
AS
24
WITH
TkE
ABOVE SYSTEM
Fig. 4. Schematic of the microwave-distribution system in room
T-131.
The microwave energy was
divided by means of low-loss coaxial cable via a 3-way splitter. Then the microwaves were fed through
a single-pole double-throw
(SPDT)
coaxial relay to a 2-way splitter. Microwaves from each arm
of
the 2-way splitter were fed
to
a 5-way splitter; thus, the power-level
of
microwave radiation was again
equally divided and transmitted through isolators to the two groups
of
five active exposure waveguides
in each alcovc. The distribution system
of
the second gcnerator in each room was similar except that
the microwaves were initially split in two ways to energize two alcoves. Power levels
of
forward and
reflected energy at each generator output terminal was measured and recorded through a directional
coupler and digital power meters interfaced with a microprocessor.
The microwaves from one generator in each equipment rack were transmit-
ted to three alcoves (Fig.
4).
The power levels of input, reflected, and transmitted
energy associated with one exposure waveguide per room were monitored to
ob-
tain a recording
of
the average absorption loss
of
the waveguide-rat assembly; the
average
SAR
could be calculated from the known waveguide
loss
and the mass of
the rat. Each room contained
a
total
of
nine power meters, two each for the inci-
dent and reflected energy at each generator and five for the incident, reflected, and
transmitted energy at the multiple terminals
of
the respective waveguides. The average
SAR
in the experimental animal was determined from the power meters. Through-
out the chronic study, the monitoring system was connected each day to a differ-
ent exposure waveguide
so
that every waveguide was monitored
1
day every
50
Long-Term
Low-Level
RF
Exposure
477
days over the course of the experiment. There was insignificant down time due
to
microwave power failure. Spare generators were available for this rare occurrence.
Dosimetry
Dosimetry studies conducted in preparation for this experiment were directed
toward determining the power level for each waveguide that would best simulate
with rats the exposure of man to an
RF
field.
To
determine the conditions neces-
sary for simulating such exposure, the relation between the input power and the
average and distributed
SAR
in the body of an exposed rat living in the exposure
waveguide had
to
be quantified.
A
microprocessor-controlled, twin-well calorimetry system was developed to
measure the average
SAR
in rat carcasses. The average
SARs
for live exposed rats
over the first year of exposure are shown in Figure
5.
The results show that the
SARs
calculated from the data on the live animals are very close to but slightly less than
the values calculated from measurements on rat carcasses. The results of this study
have been published by Chou et
al.
[1984]
To best simulate the exposure of human beings, from child to adult, to radiation
at the maximal levels allowed by
ANSI
C95.1-1982
[1982], the input power level for
each alcove cluster was set
so
that the average input power was 0.144
W,
which re-
sulted in an initial average
SAR
of
0.4
Wkg in young rats of 200-g body mass.
Predicted
Experimental
2450-MHt
Circular Waveguide
-
Apparent SAR
___
Actual SAR
-
Apparent SAR
Actual
SAR
___
_-__
0.2
-
0.1
7
100
200
300
400
500
600
700
MASS
(9)
Fig.
5.
Average SAR values measured for rat carcasses (Predicted) compared with average SAR val-
ues measured
for
free-roaming exposed rats (Experimental); SARs were averaged over weekly peri-
ods during first-year chronic exposure (input power
0.144
W
to the waveguide). The predicted actual
SAR was measured calorimetrically
on
rat carcasses
of
various body masses; bodies were exposed in
five orientations in the waveguide: center, far corner, side, transverse, and diagonal positions. Pre-
dicted apparent
SAR
was measured by power meters
on
rat carcasses. Experimental apparent SAR was
measured by power meters in live
rats.
Assuming that a live rat would spend equal time in each of the
five orientations, the experimental actual SAR data were calculated from the apparent SAR and cor-
rection factors. The correction factors were thc ratios between actual
SAR
and apparent SAR mea-
sured
on
rat carcasses.
478
Chou
et al.
Experimental Animals and Exposure Regimen
Two hundred male, Cesarean-derived, barrier-reared, Sprague-Dawley rats were
obtained at
3
weeks of age from Camm Laboratory (Wayne, New Jersey): the rats
were randomly assigned to exposure and control groups. Exposure began at
8
weeks
of age,
2
1
.S
hlday, 7 days a week, for
25
months. Maintenance procedures were
done between
8
A.M.
and
12
A.M.
to minimize circadian-rhythm effects. The two
and one-half hours off-time was used for cage cleaning, measurements of body mass,
food and water consumption, blood letting, and other biological procedures.
Biological Assessment
Behavior testing.
Behavior is a valuable end point for assessing neurological
effects of exposure to microwaves [Lovely et al., 1977; cf. Shandala et al., 19791.
Constraints of both design and logistics, however, made selection of appropriate
tests for this project a difficult task. Tests should not jeopardize the health of the
animals
or
the reliability of data obtained from other measures.
A
test protocol
must not entail differential treatment of an animal based on its performance
(e.g.,
shock intensity or reward magnitude) and thereby produce secondary effects as
artifacts that must be distinguished from any primary (microwave) effect.
In
addition, all testing must be performed within the
SPF
environment and in such
a manner
so
as not to interfere with the normal daily maintenance procedures
or
exposure protocols.
The risk of physical harm to the animals eliminated many standard behavioral
tests,
so
we chose a simple behavioral test based on quantification of a naturally-
occurring behavior. Open-field or exploratory behavior has long been used as a
sensitive endpoint in pharmacology and teratology, and it is accepted as a measure
of
general arousal or anxiety [Walsh and Cummins, 19761. In addition, East Euro-
pean researchers have used the open-field test extensively in biological studies of
microwaves [Shandala
et
al., 19791.
The open-field test is not the most impressive of the behavioral tests consid-
ered; however
it
is simple in nature and does not rely on elaborate
or
time-consuming
training procedures or shock-motivated performance, and it can be routinely ad-
ministered by laboratory personnel under the rigid SPF protocol.
An open-field apparatus with infrared-light-emitting sensors was used. This
apparatus provided a readout of both motion activity and the coordinates in the field.
The latter information was used to indicate an animal’s field position in one of the
possible quadrants. In addition, at the end
of
each test session the apparatus was
inspected for urination and defecation.
Evaluation
of
the immune system.
Alterations in the immune system due
to
microwave exposure have been reported and disputed in the literature [cf., e.g., Mayers
and Habeshaw, 1973; Czerski et al., 1974; Huang
et
al., 1977; and Wiktor-Jedrzejczak
et al., 19771. The conflicting results justified an assay of immunocompetence in
this study. The immune-system evaluation consisted
of
several basic tests that were
designed to detect immunological effects that might result from exposure to
RF
fields:
a. Blood lymphocyte evaluation of the numbers of
B-
and T-cell, antigen-
positive lymphocytes, and complement-receptor-bearing lymphocytes.
b. Spleen lymphocyte evaluation for response
to
the following mitogens:
phytohemagglutinin (PHA), concanavalin
A
(ConA), pokeweed mitogen (PWM),
lipopolysaccharide (LPS), and purified protein derivative of tuberculin (PPD).
Long-Term Low-Level
RF
Exposure
479
c. Direct plaque-forming cell assay (with spleen cells) and serum-antibody
titration of exposed rats immunized with the T-dependent antigen sheep red-blood
cells (SRBC).
The following immunological tests were performed at the 13-month interim
euthanasia
of
10
animals from each treatment group, and after
25
months of expo-
sure with the final euthanasia of
10
animals from each group; response of splenic
lymphocytes to various mitogens, plaque-forming ability, complement-receptor
formation, and enumeration of
B-
and T-cells.
Blood
sampling
for
corticosterone and health profile.
Pituitary-adrenal axis
activity as indexed by plasma corticosterone levels has long been interpreted as an
indicator of general arousal, i.e., alerting borne of anxiety, fear, or stress. If long-term
exposure to pulsed
RF
fields disrupts normal physiological functions or is psycho-
logically disturbing to the animal, an increased basal level of corticosterone can be
expected [Lotz and Michaelson, 19781. The endocrine system can provide evidence
of summation of multiple, otherwise subthreshold, effects. Individual corticosterone
data are of value for correlation with results from individual animals or subpopula-
tions that might exhibit abnormal indices of blood chemistry or a high incidence of
tumors, and also as a measure of a possible nonspecific microwave effect.
The research protocol required the rapid collection of blood from all test animals
in a 2-h period per day over 4 days for each blood sampling. The collection pro-
cedure was designed
to
be as rapid and atraumatic as possible.
To
prevent artifac-
tual elevation of corticosterone, blood samples for serum corticosterone were drawn
within 2 min after a rat was removed from its cage [Zimmermann and Crutchlow,
1967; Davidson et al., 19681. The animals were rapidly anesthetized by a mixture
of halothane, nitrous oxide, and oxygen; blood samples were drawn by the rela-
tively atraumatic retro-orbital technique. Alternate eyes were sampled for blood in
successive samplings so as to minimize ocular damage. A single blood sample, 1.8
to
2.0
ml, was taken at each session for all determinations.
Metabolism.
An important consideration in performing the long-term exposure
of rats is that the nominal 0.4-W/kg average SAR, initially is about
5%
of the average
metabolic rate of an active, young 200-g rat and about
10%
of its resting rate. This
SAR may be as high as
15%
of the average metabolic rate of a lethargic, old, 600-g
rat and
25%
of its resting rate. The decision was, therefore, made to use a constant
power density, which resulted in a declining SAR as the animals matured.
Exposure to microwave radiation for long periods could have different
consequences for longevity, either life-shortening or life-lengthening, depending
on the energy-budgeting option [Sacher and Duffy, 19781. Therefore, given the
importance of the metabolic versus extrinsic-budget question, the protocol provided
the following animal measurements:
a. Daily-lifetime body mass measures, i.e., growth.
b. Daily-lifetime food and water consumption.
c. 24-h cycles of oxygen consumption and carbon dioxide production,
d. Periodic assessment of thyroxine level.
e.
Periodic assessment of urine production.
f. Total-body analysis at spontaneous death or termination.
Despite the importance of direct metabolic measurements through respiratory
gas-exchange analysis, two factors precluded their application
to
all
200
animals:
measured at regular intervals throughout the life span.
480
Chou
et
al.
(1)
physical as well as financial constraints made it impossible to instrument all
200 waveguides, and (2) rotating all animals through a few instrumented waveguides
would have an associated
animal-transfer-management
risk and a subsequent
loss
of data. In addition, were such a mass rotation attempted, the need
to
allow each
animal a minimum of 2 days in the instrumented waveguide to adapt to the new
environment would have led to a rotation schedule allowing data to be obtained,
at most, twice a year from an animal, which would have been too infrequent. There-
fore, we selected a subset of the exposed and control samples for rotation through
waveguides adapted for the measurement of oxygen consumption and carbon di-
oxide production. This procedure did not result
in
loss of overall statistical power,
and it produced more frequent measures on the specific animals involved. Given
the modular arrangement
of
the rooms, 36 animals
(1
8
exposed and
18
sham-ex-
posed) were measured for respiratory gas exchange.
Histopathology.
As part of a general health screen at time
of
animal procure-
ment,
10
rats, 21 days old, received gross and histopathological examination. After
13 months,
10
exposed and
10
sham-exposed rats were randomly-selected and eutha-
natized for examination; at
25
months, the surviving 12 exposed rats and
11
sham-
exposed rats were euthanatized and examined. The other
157
animals were examined
when they died spontaneously
or
were terminated
in
extremis
during the study.
A pathologist (L.L.K.), without knowing the identity of the rats, provided
evaluative data to the technical personnel of the Bioelectromagnetics Research Labo-
ratory, who were responsible
for
computer entry and quality control. Statisticians
then evaluated the data, and the final results were reviewed by the pathologist for
appropriate interpretative comments.
The occurrence of neoplastic and non-neoplastic lesions was recorded along with
the age and the cause of death of each animal, whether the animal was euthanatized
or had died spontaneously. The data on pathology were collected to permit compari-
son of survival curves of exposed and sham-exposed animals, age-associated lesions,
and incidence
of
tumor metastases, as well as the number
of
lesions per rat.
RESULTS
Behavioral Evaluations
Figure 6 shows data from the 14 sessions of open-field assessment; except for
the first test session, 2 years of exposure to the low-level, pulsed-microwave ra-
diation did not lead
to
significant behavior alterations as measured by activity,
defecation, or urination. During the first test session, the general activity level of
the exposed animals was significantly lower
(L=
-2.24,
p
=
.026,
df
=
195),
by
approximately
9%,
than that
of
the sham-exposed animals. The open-field activity
pattern during the course of this study resembles that normally observed as a function
of
age and experience, and it apparently was not affected by a lifetime of exposure
to the low-level pulsed microwaves (Hotelling’s
T2
statistic
F
=
8.73,
P
=
.40,
df
=
8,168).
Plasma Corticosterone
Analysis of the data obtained during the five sampling periods (Fig.
7)
indi-
cates that serum corticosterone levels were not dramatically altered
in
either the
Long-Term Low-Level
RF
Exposure
481
3
300
v)
+I
I
3
250
8
200
2
100
g
150
c
2
U
w
50
1
1
I
2
I
Exposed
0
Sham
Exposed
3
4
5
6
7
8
9
1011 121314
OPEN-FIELD
SESSION
Fig.
6.
Comparison by treatment group
of
mean levels of activity throughout the
14
open-field
as-
sessment sessions.
exposed or sham-exposed rats. The multivariate statistical analyses of the data
(F
=
1.38,
P
=
.24, df
=
5,133) indicate that no overall effects of microwave radiation
were measurable by levels of serum corticosterone.
When the serum corticosterone values of exposed and sham-exposed animals
were compared for each session, a
t
test indicated that exposed animals had rela-
tively elevated serum corticosterone levels at the time of the first sampling session
(1
=
2.06,P
=
.04,
df
=
154),
and that sham-exposed animals had elevated levels at
the time of the third session
(t=
-2.25,
p
=
.026, df
=
161).
Exposed and sham-exposed
animals had comparable levels of corticosterone on all other regular sampling sessions.
The finding of elevated corticosterone was tested in a follow-up study [Chou
et al., 19861.
Two
groups of
20
animals each were exposed for 6 and
12
months,
respectively, under the same exposure parameters as in the original study. An equal
I
+I
w
FRST-YEAR
SAMPLE
PERlOD
Fig.
7.
Comparison of mean corticosterone levels from five quarterly determinations during
the
first
year
of
the project.
482
Chou
et
al.
number of sham-exposed rats served as controls. Corticosterone measured at
6
weeks,
6 months and 12 months did not show any statistically significant differences
(_P>
.05)
between 20 control and 20 exposed rats.
Immunological Competence
When compared with sham-exposed rats after
13
months of exposure (Fig.
8),
exposed animals had a significant increase in both splenic B-cells
(t
=
3.76,
P
=
.002, df
=
16)
and T-cells
(t=
3.48,
P
=
.003,
df
=
16). This apparent general
stimulation of the lymphoid system in exposed animals was not detected in the animals
after
25
months of exposure: Comparison of exposed and sham-exposed rats at
euthanasia of survivors did not reveal any significant differences in the percent-
age or total numbers of B and T cells per spleen.
No
significant differences were seen between exposed and sham-exposed rats
in
the percentage of complement-receptor-positive cells in the spleen at either the
interim or final euthanasia. These findings indicate no difference between the treatment
groups for lymphocyte maturation.
The plaque assay performed on exposed animals immunized with SRBC
in
the 13-month exposure rats exhibited a slight but statistically insignificant increase
in plaques per spleen relative to the sham-exposed. This difference reversed after
25 months when exposed animals showed a slightly lower and statistically insig-
nificant number of plaques per spleen. This assay indicated no statistically significant
alteration of the reticuloendothelial system, which first processes antibodies in the
presence of T-cells, because the SRBC antigen is T-cell dependent.
The mitogen-stimulation studies following 13 months of exposure revealed
significant differences between groups in their responses to various
B-
and T-cell
specific mitogens. The radiated animals had a nonsignificant increase in response
to
PHA
but a significant increase
in
response to
LPS
(mean of 6.06 vs. 3.67,
f
=
2.35,
P=
.032) and PWM (mean of 6.41
vs.
4.61,
t=
2.43,
p=
.027).
As
compared
with sham-exposed animals, exposed animals also had a significantly increased
response to ConA (mean
of
17.0
vs.
10.7,
t
=
2.65,
_P
=
,018)
and a decreased re-
sponse to PPD (mean of 2.74 vs. 6.98,
t=
-2.65,
p=
.018).
These results indicate
a selective effect of exposure on
the
lymphoreticular system's response to mitoge-
EB
Exposed
25
5
?I
*O
w
;
'5
;
10
2
w
n
25
i5
=o
BCells
1-Cells
Fig.
8.
Mean percentages
of
B-cells and T-cells
within
culture population
of
splenic lymphocytes
foi
exposed and sham-exposed
groups.
Long-Term Low-Level
RF
Exposure
483
nic stimulation. Mitogen-response data were not available from the 25-month ex-
posure studies because the lymphocyte cultures failed to grow.
In
a follow-up study [Chou et al.,
19861,
no
significant differences between
20
exposed and
20
sham-exposed rats were observed in the proliferation of thy-
mocytes to ConA, PHA, and PWM after
6-
and 12-months of
RF
exposure. The
same lack of differences was found for splenocytes stimulated
by
LPS,
PHA,
PPD,
ConA, and PWM. Flow cytometry revealed
no
group alterations in the number
and frequency of B- and T-cells. However, after
12
months
of
exposure, a reduc-
tion in cell surface expression of Thy 1.1 (T-cell related) surface antigen, and a
reduction in the mean cell-surface density
of
s-Ig
(B-cell related) on small lym-
phocytes in spleen were observed. The stimulatory effect observed in the origi-
nal study was not confirmed.
General
Health
Profile
In an attempt to detect and document any effects
on
the general health of the
exposed animals, the following biochemical and hematological parameters were
monitored: serum chemistry components, hematological constituents, protein elec-
trophoretic patterns and fractions, and thyroxine levels. Multivariate analyses with
Hotelling's T2 statistic on a truncated data set (outliers removed) indicated no overall
=Exposed
ashorn
Exposod
20aT
B
1
eedi
ng
Sees
1
on
Fig.
9.
Comparison of serum glucose
for
exposed and sham-exposed animals
for
IS
sampling sessions.
=Exposed
Shorn
Expooczd
2001
TI
a
2
4-
6
8
18
12
14
Wlreding
Serei-n
Fig.
10.
Comparison
of
serum cholesterol levels of exposed and sham-exposed animals from
15
sam-
pling sessions.
484
Chou
et
al.
differences among all parameters between exposed and sham-exposed samples.
Figures
9
and
10
present two representative examples of glucose and cholesterol
levels from
36
sets of data. Individual
f
tests of all parameters across all
15
sam-
pling sessions indicated a significant reduction in the absolute eosinophil counts
of exposed rats during session
2,
and marginally significant reductions in absolute
neutrophil count during sessions
2
and
3.
None of the other comparisons was sig-
nificant. Therefore, these findings indicate that after the 25-month exposure no
consistent effects were produced in bone-marrow erythropoietic cells or in the
juxtaglomerular apparatus of the kidney and its production of erythropoietins.
Twenty-one serum chemical constituents were measured in serum samples
collected during all
15
sampling sessions. The serum-chemistry tests were sensi-
tive enough to detect population changes due to aging. Statistical analysis of the
data by Student’s
t
tests did not indicate any differences between exposed and sham-
exposed animals.
Electrophoresis of the serum proteins revealed no significant changes in the
electrophoretic patterns and absolute protein fractions between the population groups.
Both groups showed a gradual decrease in the albumin/globulin ratio with increasing
age, and the overall level of globulin fractions observed in these barrier-sustained
animals was lower than that reported in conventional-colony animals. The micro-
wave exposure had no apparent effect on the functioning of various organ systems
that contributed to serum-protein concentrations.
Thyroxine levels did not differ significantly between exposed and sham-ex-
posed animals (Fig.
11).
Thus, exposure had no effect on the hypothalamic-pitu-
itary-thyroid feedback mechanism. The absolute level of serum thyroxine developed
to a maximum in young animals and decreased gradually as they aged. The corre-
lation of this age-related decrease in thyroxine levels with increasing cholesterol
(Fig.
10)
and triglyceride levels in both test and sham groups shows
it
to be a re-
liable indicator of metabolic activity in the rat.
The major conclusion that can be reached from the evaluations
of
hematol-
ogy, serum chemistry, protein electrophoretic patterns and fractions, and thyrox-
ine levels is that any significant variations of the parameters observed during the
lifetime of the exposed animals were to be expected as a function of aging.
a
Exposed
0
Shorn
Exposed
-
-
=1
B
1
eedi
n9
Serr
i
on
Fig.
I
I.
Comparison
of
thyroxine data
for
exposed and sham-exposed animals
for
blood
sampling sessions
for
which analysis
was
made.
Long-Term
Low-Level
RF
Exposure
485
550
500
450
400
-
0
-
350
5
v)
+I
300-
z
$
250-
c
-
EXPOSED
0
SHAM EXPOSED
-
-
ttttttft
t
?
t t?
1
15
-
’
BLEEDING SESSION
t
t
2ooI
150
ii~wm
ON INTERIM KILL FINAL KILL
-1
I
1
I
I
1
I
10
20
30
40
50
60
70
80
90
100
110
WEEKS
IN
STUDY
Fig.
12.
Mean weekly body mass throughout 25-month study. Arrows indicate periodic bleeding
ses-
sions
as
well
as
other significant events during the course of the study.
Metabolism
Body
mass and consumption
of
food and water.
Growth curves for micro-
wave-exposed and sham-exposed animals throughout this study (Fig.
12)
are in general
agreement with those reported for the Sprague-Dawley rat [Berg, 1960; Masoro,
19801. The asymptotic body mass was somewhat lower than expected, possibly
because of a periodic “stunting” effect coincident with the start of the regular blood-
sampling sessions.
The average daily food intake of approximately
25
to
26
g is higher than that
usually reported for the rat [Brobeck, 1948; Hamilton, 1967; Jakubczak, 19761 and
indicated by the feed manufacturer
(12
to
15
g/day). These food-intake norms,
however, are for animals housed in a standard animal facility maintained at a higher
ambient temperature (25
“C).
The amount of food eaten by the animals in our fa-
cility, which was maintained at 21
&
1
“C,
is in agreement with that reported for
animals housed at lower ambient temperatures [Brobeck, 1948; Hamilton, 1967;
Jakubczak,
19761
and in other studies in our laboratory that had used the waveguide
apparatus [Lovely
et
al., 19771. Throughout the 25 months, no overall differences
were observed between treatment groups in either food or water consumption.
The similarity in overall patterns
of
growth, food and water consumption, and
body-mass
loss
and recovery in exposed and sham-exposed samples indicates that
no effects of microwave irradiation were apparent in these measures of long-term
energy balance.
486
Chou
et
al.
Total body analysis.
With one exception, the combined analyses of organ mass,
general carcass composition, fatty-acid profile, and mineral content provided no
evidence that metabolic processes were adversely affected in the animals exposed
for
I3
or 25 months to microwave radiation.
A
highly significant elevation of adrenal
mass was indicated by the 75% increase observed for exposed rats as compared with
sham-exposed animals. However, when the animals with benign tumors
in
the adrenal
gland were separated from those without tumors, the difference became insignifi-
cant.
For
animals with tumors,
the
adrenal mass was significantly higher in the exposed
group than in the sham group. This analysis indicated that the increase in adrenal
mass was related to the tumors and was independent of the metabolic processes in
the rats. The mean adrenal mass in exposed animals without tumors was slightly
larger, but statistically insignificant, as compared with that of the sham-exposed
rats. This increase in mass was attributed to one animal with a hyperplastic adre-
nal cortex, which was secondary to a pituitary tumor.
0,
consumption and
CO,
production.
Differences between exposed and sham-
exposed rats occurred in
0,
consumption and
CO,
production in younger rats (body
mass 300-400 g) but not in the more mature animals (17-24 months old, body mass
550-600
g). The average hourly
0,
consumption for the young rats during the
nocturnal period
(1
900-0600 lights off) was significantly different between the
treatment conditions (Hotelling T’ statistic,
F
=
2.29,
P
=
0.025,
df
= 11,44).
Al-
though individual
t
tests
of
hourly
CO,
productions of the young animals did not
show consistent significant difference between treatment groups, the Hotelling
T’
statistic was significant during the diurnal (1300-1900 lights
on)
period
(lj
=
2.73.
-
P
=
.023, df
=
6,49) and even more significant during the night time hours
(F
=
2.91,
-
P
=
.006,
df
=
11,44). The effects observed in the young animals were less pronounced
during the second round (36 days later) of measurements. On an hour-to-hour basis,
the mature animals’ metabolic measures appeared less variable than those of the
young. The young animals demonstrated more marked responses to the lights-off
condition and generally higher levels
on
each measure during the night time hours,
i.e., the active portion of the rats’ circadian cycle. This apparent synchronization
of metabolic activity with the light-dark cycle has been noted by others investigating
the variation of activity, food and water consumption, and energy balance patterns
as a function of photoperiod [Zucker, 1971; Besch and Woods, 19771.
Gross
Pathological and Histopathological Evaluation
Longevity.
Product-limit estimates and log-rank statistics were used to esti-
mate and compare survival curves of exposed and sham-exposed animals (Fig.
13).
Evaluation of the curves revealed that the median survival time was 688 days for
exposed animals and 663 days for the sham-exposed. Despite subtle differences in
the survival curves in the early and late stages
of
the study, statistical analysis indicated
no significant differences during any phase
of
the life span of the animals. Statis-
tical evaluation indicated no association between
a
specific cause of death and
treatment condition: however, for cause of death due to urinary tract blockage (9
in exposed group and 19 in sham group), there is some indication that survival times
were longer in the exposed animals.
Histopathology.
Parasitic, bacterial, mycoplasmal, and viral agents were
monitored during the 25-month period.
A
low-level(15%) infestation of
the
colony
with pinworms,
Syphacia
rnuris,
occurred but no histological lesions were attrib-
uted to these nematodes. The microflora of the animals was altered over the course
of the experiment by the sporadic occurrence of
Proteus
sp.
(mirahilis,
rettgeri,
and
Long-Term Low-Level
RF
Exposure
487
1.0
0.8
5
0.6
z
[
0.4
8
0.2
f
c
-
-
t
O
@
.4
.:yo
=%%\
0
EXPOSED
SHAM
EXPOSED
01
1
I
I
I
0
200
400
600
800
AGE
(DAYS)
Fig.
13.
Survival data
for
microwave-exposed and sham-exposed animals throughout the 25-month
study.
vulgaris), Staphylococcus epidermidis, Neisseria
sp.,
Escherichia coli,
and
Kleb-
siella
sp. These intestinal flora became opportunistic organisms in the few cases
of preputial adenitis and wound infections that occurred.
Mycoplasma
sp. was not
isolated, either by culture or serology, and serological monitoring failed
to
reveal
any significant elevations in titers of any of the common rodent viruses. There were
no underlying diseases that complicated
or
produced erroneous results in the gross
or histopathological evaluations of the experimental animals.
The histopathology data were grouped with respect to the animal's age, at
6-
month intervals, and the data were divided into neoplastic and non-neoplastic di-
agnoses. The documentation of morphological lesions showed
2,184
pathological
changes in the 200 animals examined. The non-neoplastic lesions comprised 1,992
of the observed changes, with 217 unique combinations of organs and lesions. The
neoplastic lesions accounted
for
192 of the observations, with
83
unique combina-
tions of organs and types of neoplasms.
Chronic glomerulonephropathy was the most frequent cause
of
death and one
of the most consistently encountered non-neoplastic lesions. Statistical analysis
(Mantel-Haenszel estimate and chi-square statistics) indicated that glomerulonephro-
pathy was less frequently observed in the exposed than in the sham-exposed ani-
mals
(P
=
.04,
df
=
I).
Analysis
of
the other non-neoplastic lesions did not indicate
that the specific lesions were more likely in either treatment condition. To detect
a progressive development of the chronic glomerulonephropathy, the severity
of
the lesions was also evaluated. This analysis revealed no significant differences be-
tween the treatment condition and the severity of non-neoplastic lesions.
The neoplastic lesions were identified as benign or malignant, with the ma-
lignant lesions classified as primary or metastatic.
A
summary
of
these combina-
tions is presented in Table 2, which indicates the total number of primary and
metastatic malignancies and benign lesions observed in both exposed and sham-
exposed animals. The incidence of neoplastic lesions corresponds with that normally
reported for the Sprague-Dawley rats: Only two tumors were present in rats younger
than
12
months, and tumor incidence rapidly increased after
18
months
of
age
[MacKenzie and Garner, 1973; Altman et al.,
19851.
The endocrine system had the
highest incidence
of
neoplasia in the aging rats, as is expected in this animal. The
incidence of benign pheochromocytoma of the adrenal medulla was much higher
in the exposed group than in the controls (7 out
of
100
vs.
1
out
of
100).
However,
Fisher's exact test did not show a statistically significant effect
(_P
=
.065).
488
Chou
et
al.
TABLE
2.
Neoplastic
Lesions Per Organ
System
Orean Lesions
Exuosed Sham-exuosed
B PMBPM
Adrenal
Blood vessel
Bone marrow
Brain
Cervical
Lymph node
Colon
Duodenum
Edipidymis
Heart
EY e
Kidney
Liver
Lung
Lymph node
Mesentery
Nasal cavity
Pancreas
Adenoma
Carcinoma
Cortical adenoma
Cortical carcinoma
Myelomonocytic leukemia
Malignant lymphoma
Pheochromocytoma
Hemangiosarcoma
Leu kern ia
Myelomonocytic leukemia
Malignant lymphoma
Myelomonocytic leukemia
Malignant lymphoma
Myelomonocytic leukemia
Lyinphocytic lymphoma
Malignant lymphoma
Malignant lymphoma
Myelomonocytic leukemia
Malignant lymphoma
Squamous cell carcinoma
Squamous cell carcinoma
Leukemia
Myelomonocytic leukemia
Malignant lymphoma
Neurinoma
Leukemia
Myelomonocytic leukemia
Malignant lymphoma
Nephroblastoma
Adenoma
Carcinoma
Hepatocellular adenoma
Leukemia
Myelomonocytic leukemia
Malignant lymphoma
Squamous
cell
carcinoma
Leukemia
Myelomonocytic leukemia
Malignant lymphoma
Myelomonocytic leukemia
Malignant lymphoma
Transitional cell carcinoma
Transitional cell carcinoma
Leukemia
Adenoma
Islet-cell adenoma
0
0
10
0
0
0
7
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
I
2
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
3
0
0
0
1
0
0
I
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
I
0
0
2
0
0
0
1
I
I
1
1
0
1
1
0
0
I
1
0
0
0
0
0
2
I
1
0
1
1
2
1
1
1
0
0
0
1
0
10
0
0
0
I
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
1
0
1
0
1
0
1
0
0
0
0
0
1
1
0
0
1
1
0
0
0
0
0
1
1
1
0
1
0
0
0
0
0
0
1
0
0
Continued
Long-Term
Low-Level
RF
Exposure
489
TABLE
2.
Continued.
Exvosed Sham-exvosed
Orean Lesions B PM BPM
Pancreas
Parathyroid
Parotid SG
Peritoneum
Pituitary
Preputial gland
Skeletal muscle
Skin
Spleen
Stomach
SubQ tissue
Testes
Thymus
Thyroid
Ureter
Uridbladder
Zymbal’s gland
Squamous cell carcinoma
Malignant lymphoma
Myelomonocytic leukemia
Liposarcoma
Adenoma
Carcinoma
Malignant lymphoma
Myelomonocytic leukemia
Auditory sebaceous
sq
carcinoma
Basal cell carcinoma
Basal cell tumor
Keratoacanthoma
Malignant lymphoma
Pilomatricoma
Sebaceous adenoma
Myelomonocytic leukemia
Malignant lymphoma
Malignant lymphoma
Squamous cell carcinoma
Squamous cell papilloma
Fibroma
Fibrosarcoma
Lipoma
Neurinoma
Benign interstistial cell
tumor
Squamous cell carcinoma
Myelomonocytic leukemia
Lymphocytic lymphoma
Malignant lymphoma
Adenoma C-cell
Carcinoma C-cell
Leukemia
Malignant lymphoma
Malignant lymphoma
Transitional cell carcinoma
Transitional cell papilloma
Leukemia
0
0
0
0
17
0
0
0
0
0
1
1
0
1
2
0
0
0
0
3
I
0
I
0
1
0
0
0
0
10
0
0
0
0
0
1
0
0
0
0
I
0
2
0
0
1
1
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
1
1
0
0
2
0
0
0
1
0
0
1
1
1
0
0
0
1
1
0
0
0
0
1
0
0
1
1
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
21
0
0
0
0
0
0
I
0
0
0
0
0
0
0
4
0
0
0
1
0
0
0
0
0
9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
-
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
Total
62
18
36
53
5
18
This table lists neoplastic lesions found per organ system. These lesions may be benign (B), a pri-
mary malignancy (P),
or
a metastatic malignancy (M) arising from a primary malignancy in another
organ system (i.e., a malignant neoplasm may occur as a metastatic malignancy in many organs
of
a
single animal, but as a primary malignancy in only one organ system
of
an animal).
490
Chou
et
al.
The low incidence of neoplasia with no significant increase in any specific
organ or tissue required the data to be collapsed and evaluated with respect to
occurrence per se of neoplasms, with no attention given to the site or organ of
occurrence. For benign lesions, as shown in Table
3,
the Mantel-Haenszel
(M-H)
estimate of the odds ratio was 1.04. The chi-square statistic, which tests whether
the relative risk is
1,
was .001
(p=
,97, df
=
1); therefore, we found no evidence
that either group had an excess of benign lesions. For total neoplastic incidence
including benign and malignant lesions, statistical evaluation revealed no signifi-
cant difference between the exposed and sham-exposed groups
(x’
=
0.32,
P
>
.05).
A
similar set of tables was prepared for primary malignant neoplastic lesions and
is
presented in Table
4.
When all age categories for the primary malignant lesions were
considered, the
M-H
estimate of the odds ratio was 4.27 and the chi-square statistic
was 7.66
(E
=
,006, df
=
1).
With the first three age categories combined and the analysis
repeated, the
M-H
statistic was 4.38 and the chi-square statistic was 7.9
(P
=
.005,
df
=
1).
When the first four age categories were collapsed (leaving two categories:
1-24
and
25-30
mo), the
M-H
statistic was 4.47 and the chi-square was
6.97
(P
=
.OOS,
df
=
1).
When age at death was ignored completely, the
M-H
estimate of the relative risk
was 4.46 and the chi-square was 8.00
(P
=
.005,
df
=
1).
It
is interesting that the es-
timate of the odds ratio and the chi-square statistic are both insensitive to the way the
data were grouped with respect to age at death.
A
survival-type analysis also was done with time of death as a surrogate for
time to tumor development if a primary malignant lesion were present. If no pri-
mary malignant lesions were found, time
to
tumor was considered censored at the
time to death. From that analysis, the log-rank statistic is 7.63 with a
p
value of
.006.
This analysis indicates that the primary tumors occurred earlier in exposed
rats than in sham-exposed animals.
DISCUSSION
We investigated the effects on health of long-term exposure to low-level, pulsed,
microwave radiation. Among the
155
parameters studied, most of them showed no
TABLE
3.
Incidence
of
Benign
NeoDlasms
at Death
Ape Benign
No.
of
animals
neoolasms Exoosed Sham
Age considered
(rno)
1-6
Yes
No
7-12
Yes
NO
13-18 Yes
No
19-24
Yes
No
25-30 Yes
No
Age not considered Yes
NO
0
3
0
5
1
24
16
19
22
10
39
61
0
3
3
5
5
18
11
24
19
12
38
62
Long-Term Low-Level RF Exposure
491
TABLE
4.
Incidence
of
Primarv
Malienant Lesions
at
Death
Primary
malignant
No.
of
animals
Age lesions Exposed Sham
Age considered (mo)
1-6
Yes
0
0
No
3
3
7-12
Yes
0 0
No
5
8
13-18
Yes
2
2
No
23
21
19-24
Yes
9
I
No
26
34
25-30
Yes
7
2
No
75
29
Age
not
considered
Yes 18
5
No
82
95
significant differences associated with exposure during the 25-month period. However,
a few parameters showed positive effects. There was a statistically significant in-
crease in the mean of the serum corticosterone level in exposed rats at the time
of
the first blood sampling, and there was a significantly lower level at the third ses-
sion of measurement as compared with sham-exposed animals. The other signifi-
cant effects involved the immune response of the rats at
13
months of exposure and
the
O,/CO,
metabolism in young rats.
The early finding of elevated corticosterone levels was not found
in
the later
sessions of the 25-months study. The failure to repeat may be due to maturational
differences, to the decreasing SAR as animals grew, or to a combination of the two.
The lack
of
a significant difference
in
the total number of
B
and T cells in the terminal-
euthanasia animals of the original study may also be the result
of
aging, the onset
of immunosenescence, or the declining
SARs.
The role of a decreased
SAR
in animals
across time should be considered if a similar study is conducted. One could avoid
this problem by increasing the power level to keep the SAR constant.
A follow-up study was conducted to confirm both corticosterone and immune
system effects [Chou et al.,
19861.
Neither effect was confirmed in two groups of
20
animals each exposed for
6
and
12
months, respectively, under the same expo-
sure condition as the original study.
An
equal number of sham-exposed animals served
as controls. The sample size of 20 animals per group was chosen to have good
statistical power
(80%)
to detect the same magnitude of differences observed in the
original study. The failure to confirm indicates that the original findings are not robust.
The lack of discernible differences in
0,
consumption and C0,production in
the mature animals at this level of microwave exposure is
in
agreement with the
results of Phillips et al.
119751.
They exposed male adult rats to various intensities
of
2,450-MHz microwaves. Animals receiving
27
cal/min
(-
2W)
showed
no
dif-
ference from controls. The microwave exposure in our study resulted in an energy
deposition
of
1.5-2.0
cal/min
(144
mW)
throughout the lifetime
of
the animal,
492
Chou
et
al.
well below levels employed by Phillips et al. Under the ambient environmental con-
ditions of temperature, humidity, and airflow, the rate of energy deposition used
in our study was not sufficient to produce robust changes in the metabolism of the
mature rat exposed to microwave irradiation. Changes
in
the
0,
consumption and
CO,
production were observed in young, exposed animals-and these changes were
more pronounced during the first round of the measurements-are consistent with
the fact that the rate of energy absorption in our waveguide apparatus decreases with
increasing body mass. Due to the fast growth rate of the rats (Fig.
12),
the animals
were subjected to higher SARs only during the first month.
The incidence of benign pheochromocytome of the adrenal medulla was higher
but not statistically significantly
so
in the exposed group. However, we note that
the incidence of this tumor in the exposed group does not exceed the incidence of
tumors reported in the literature for this strain of rat housed under specific patho-
gen-free conditions [Anver et al.,
19821.
Strict comparisons
of
these data with those
from other laboratories cannot be made, however, because the animals were not
subjected to parallel conditions. A reference control-large numbers of untreated
rats except for observation of longevity and post-mortem analysis-would be de-
sirable in future studies.
The finding of a near fourfold increase of primary malignancies in the exposed
animals
is
provocative. These data cannot be considered as an artifact because different
statistical analyses led to similar results. Although the overall difference in num-
bers of primary malignancies is statistically significant, the biological significance
of this difference
is
open to question. First, detection of this difference required
the collapsing of sparse data without regard for the specific type of malignancy or
tissue of origin. Also, when the incidence of the specific primary malignancies in
exposed animals was compared with specific tumor incidence reported in the lit-
erature, the exposed animals had an incidence similar to that of untreated control
rats of the same strain maintained under similar
SPF
conditions. It is important to
note that no single type of primary malignancy was enhanced
in
the exposed ani-
mals. From the standpoint of carcinogenesis and under the assumption that the
initiation process is similar for both benign and malignant tumors, benign neoplasms
have considerable significance. That treatment groups showed no difference in
incidence of benign tumors is an important element in defining the promotion and
induction potential of microwave radiation for carcinogenesis.
Morphologically, carcinogenesis proceeds through transitory or progressive
states of growth, including hyperplasia and/or dysplasia, benign neoplasia, and finally
overt malignant neoplasia. This morphological continuum, which often, but not always
occurs, is the basis for grading systems and staging systems in common usage in
medical pathology. Although the exact cause of cancer remains illusive, there is con-
siderable morphological and biochemical evidence that neoplasms in humans and
animals progress through a series of stages and ultimately become completely
autonomous, invade surrounding tissue, and metastasize widely. Although there are
readily recognizable histopathologic differences between the cancer cell and the
normal cell, the biochemical differences, especially relating to the molecular biol-
ogy of
DNA
and RNA synthesis, protein and polypeptide synthesis, enzyme activity,
and membrane receptions to ultrastructural and cellular components is far from being
completely understood [Busch,
1974, 19791.
Long-Term Low-Level
RF
Exposure
493
The incidence of benign pheochromocytomas of the adrenal medulla was higher
in the exposed group than in the controls; however, no other single type of tumor
was significantly increased by the treatment, even though the primary malignan-
cies of all types is significantly elevated in the exposed group. In considering this
issue, one perspective to keep in mind is that, with the induction of cancer by a
carcinogen, tissue-specific effects are usually induced,
so
that an agent is not
usu-
ally considered carcinogenic unless it induces a significant response in any one tissue.
The
U.S.
Environmental Protection Agency Guidelines for carcinogenicity risk
assessment states,
“A
statistically significant excess of tumors of all types in the
aggregate, in the absence of a statistically significant increase
of
any individual tumor
type, should be regarded as minimal evidence of carcinogenic action unless there
are persuasive reasons to the contrary”
[U.S.
EPA,
19861.
The combining of malignant tumors from all sites for statistical comparison
of incidence in the exposed and control groups is questionable as to its biological
relevance.
A
major factor that one must consider is the different response found in
this study from what is expected in a chemical carcinogenesis study. There was no
discernible induction of benign tumors in the organs that were apparently devel-
oping malignant neoplasms. Considering that the majority of the
155
parameters
evaluated showed no differences, and especially that longevity was not affected,
the biological significance of the increased primary malignancies is unknown. Chance
variations may be the reason for difference in numbers of malignancies [Ward, 19831.
Scientists of the Georgia Institute
of
Technology have performed a comple-
mentary study, also supported by the Air Force;
200
rats were exposed to 435-MHz
fields in circular, parallel-plate waveguides,
22
hr/day for
6
months.
No
significant
differences in blood-borne end points were found [Toler et al.,
19881.
To explore
the possibility of RF-induced tumor initiation or promotion, the Georgia Tech group
exposed a large population (200 exposed and
200
sham-exposed) of mammary-tumor-
prone mice to 435
MHz
fields for
21
months. This study was specifically designed
to examine the effects of low-level, pulsed RF fields on cell growth and differen-
tiation, unlike
our
project which was designed to study effects on general health
and longevity. Their experiment is completed and the data are being analyzed. It
will be interesting to compare their results with ours.
CONCLUSIONS
Microwave exposure of
100
male rats (and
100
sham-exposed controls) at SARs
of 0.4 to
0.2
W/kg (pulsed, 2,450-MHz circularly-polarized microwaves at
21.5
h/
day, for
25
months) showed no biologically significant effects on general health,
serum chemistry, hematological profiles, longevity, cause of death, and lesions
associated with aging and benign neoplasia. Statistically significant effects were
found in corticosterone levels and immunological parameters at 13 months expo-
sure, but these findings were not confirmed in a follow-up study.
0,
consumption
and C0,production were lower in exposed young rats. These effects were not observed
in mature rats. The findings of an excess of primary malignancies in exposed ani-
mals is provocative. However, when this single finding is considered in light of other
parameters, it is conjectural whether the statistical difference reflects a true bio-
logical influence. The overall results indicate that there are no defintive, biologi-
494
Chou
et
al.
cally significant effects on rats chronically exposed to this form of microwave
irradiation. Positive findings need further independent experimental evaluation.
ACKNOWLEDGMENTS
Supported by the USAF School of Aerospace Medicine, Air Force Systems
Command, United States Air Force, Brooks Air Force Base, Texas, under contracts
F33615-78-C0631 and F33615-80-C-0612. Also supported in part by the National
Cancer Institute Grant CA
33752.
A project of this size required many dedicated
collaborators and staff members. Their contributions are deeply appreciated. In
particular, we thank Desmond Thompson, Darrel Spackman, Karl Hellstrom, Ingegerd
Hellstrom and
H.J.
Garriques. Significant contributions to protocol development
and data interpretation were made by Leo Bustad, Edward Masoro, and the late George
Sacher, who served as consultants during the study.
REFERENCES
Adey WR (1981): Tissue interaction with non-ionizing electromagnetic fields. Physiol Rev 61:435-5 14.
Altman PL, McLane
K,
Brassedu
J
(1985): “Pathology of Laboratory Animals.” Elmsford, NY: Pergamon
Press.
ANSI C95.1-1982 (1982): “American National Standard Safety Levels With Respect
to
Human Ex-
posure
to
Radiofrequency Electromagnetic Fields, 300 kHz to 100
GHz.”
New York: IEEE, pp
1-24.
Anver MR, Cohen BJ, Lattuada CP, Foster
SJ
(1982): Age-associated lesions in barrier-reared male
Sprague-Dawley rats: a comparison between Hap:
(SD)
and CRL:COBSIRI CDIR1 (SD) stocks.
Exp Aging Res 8:3-22.
Berg BN (1960): Nutrition and longevity
in
the rat.
1.
Food intake in relation to size, health and fer-
tility.
J
Futr 71:242-254.
Besch EL, Woods JE (1977): Heat dissipation biorhythms of laboratory animals. Lab Anim Sci 27:54-
59.
Brobcck JR (1948): Food intake as a mechanism of temperature regulation. Yale
J
Biol Med 20:545-
552.
Busch H (ed) (1979): “Methods in Cancer Research.” Volume 18. New York: Academic Press.
Busch H (1974): “Molecular Biology of Cancer.” New York: Academic Press.
Chou CK, Guy AW, Johnson RB (1983): “Volume
3.
SAR in Rats Exposed in 2450-MHz Circularly
Polarized Waveguide.” USAFSAM-TR-83- 19, October: NTIS publication AD-A 135376.
Chou CK, Guy AW, Johnson RB (1984): SAR in rats exposed in 2450-MHz circularly polarized
waveguide. Bioelectromagnetics 5(4):389-398.
Chou CK, Clagett JA, Kunz, LL, Guy AW (1986): “Effects of Long-Term Radiofrequency Radiation
on Immunological Competence and Metabolism.” USAFSAM-TR-85- 105, May, Brooks AFB,
TX
78235. NTIS publication AD-A 169064.
Coenen AML (1975): Frequency analysis of rat hippocampal electrical activity. Physiol Behav 14:39
I-
394.
Czerski P, Ostrowski
K,
Shore ML, Silverman C, Suess MJ, Waldeskog B (1974): “Biological Effects
and Health Hazards
of
Microwave Radiation.” Proceedings
of
an International Symposium,
Warsaw, Poland, Oct
15-1
8, 1973. Warsaw: Polish Medical Publishers.
Davidson JM, Jones LE, Levine
S
(I
968): Feedback regulation
of
adrenocorticotropin secretion in “Basal”
and “Stress” conditions: Acute and chronic effects
of
intrahypothalamic corticoid implantation.
Endocrinology 82:655-673.
Elder JA, Cahill DF
(
1984): “Biological Effects of Radiofrequency Radiation.” Report EPA-600/8-
83-026F, Health Effects Research Laboratory, Office of Research and Development, USEPA,
Research Triangle Park, NC 277
I
1.
Gandhi OP
(1
990): “Biological Effects and Medical Applications
of
Electromagnetic Energy.” Englewood
Cliffs, New Jersey: Prentice Hall.
Long-Term Low-Level
RF
Exposure
495
Guy AW, Chou CK
(I
977): System for quantitative chronic exposure of a population
of
rodents
to
UHF
fields. In Johnson CC, Shore ML (eds): “Biological Effects of Electromagnetic Waves.”
Washington DC 20402: U.S. Government Printing Office, HEW Publication (FDA) 77-801
I,
2:389-422.
Guy
AW, Wallace J, McDougall JA (1979): Circularly polarized 2450-MHz waveguide system for chronic
exposure of small animals to microwaves. Radio Sci 34(68):63-74.
Guy AW, Chou CK, Johnson RB, Kunz LL (1983a): “Volume 1. Design, Facilities, and Procedures.”
USAFSAM-TR-83- 17, September: NTIS publication AD-AI 34079.
Guy AW, Chou CK, Neuhaus B (1983b): “Volume 2. Average SAR and SAR Distribution in Man Exposed
to 450-MHz RFR.” USAFSAM-TR-83-18, September; NTIS publication AD-A 135455.
Guy AW, Chou CK, Kunz LL, Crowley J, Krupp JH (1985): “Volume 9. Summary.” USAFSAM-TR-
85-64, August; NTIS publication AD-A15951 2.
Hamilton CL (1967): Food and temperature. In “Handbook of Physiology, sec. 6: Alimentary Canal.
Vol.
1:
Control
of
Food and Water Intake.” Washington, DC: American Physiology Society,
pp 303-3 17.
Huang AT, Engle ME, Elder JA, Kinn JB, Ward TR (1977): The effect of microwave radiation (2450
MHz) on the morphology and chromosomes of lymphocytes. Radio Sci 12(6S): 173-177.
IEEE C95.1-1991 (1992): “Safety Levels With Respect to Human Exposure to Radio Frequency Elec-
tromagnetic Fields,
3
kHz
to
300
GHz.”
Piscataway, N.J.: IEEE.
Jakubczak LF (1976): Food and water intakes
of
rats as a function
of
strain, age, temperature and body
weight. Physiol Behav 17(2):25 1-258.
Johnson RB, Spackman D, Crowley J, Thompson
D,
Chou CK, Kunz LL, Guy AW (1983): “Volume
4: Open-field Behavior and Corticosterone.” USAFSAM-TR-83-42, December; NTIS publica-
tion AD-A137743.
Johnson RB, Kunz LL, Thompson D, Crowley J, Chou CK, Guy
AW
(1984): “Volume 7. Metabolism,
Growth, and Development.” USAFSAM-TR-84-3
I,
September; NTIS publication AD-A 150829.
Kaplan EL, Meier P
(1958):
Nonparametric estimation from incomplete observations. J Am Stat Assoc
Kunz
LL,
Hellstrom KE, Thompson D, Chou CK, Guy AW (1983): “Volume
5.
Evaluation
of
the Immune
System’s Response.” USAFSAM-TR-83-50, December; NTIS publication AD-A 138535.
Kunz LL, Johnson RB, Thompson
D,
Crowley J, Chou CK, Guy AW (1984): “Volume
6.
Hematological,
Serum Chemistry, Thyroxine, and Protein Electrophoresis Evaluation.” USAFSAM-TR-84-2,
March; NTIS publication AD-A141 124.
Kunz LL, Johnson RB, Thompson D, Crowley
J,
Chou CK, Guy AW (1985): “Volume
8.
Evaluation
of Longevity, Cause
of
Death, and Histopathological Findings.” USAFSAM-TR-85-
1
I,
April;
NTIS publication AD-A154283.
53:457-48
I.
Lin JC (1989): “Electromagnetic Interaction With Biological Systems.” New York: Plenum Press.
Lotz GW, Michaelson SM (1978): Temperature and corticosterone relationships in microwave-exposed
rats. J Appl Physiol 44:438-445.
Lovely RH, Myers DE, Guy AW (1977): Irradiation
of
rats by 918 MHz microwaves at 2.5 mW/cm’:
Delineating the dose-response relationship. Radio Sci 12(6S): 139-146.
MacKenzie WF, Garner
FM
(1
973): Comparison
of
neoplasms in six sources of rats.
J
Natl Cancer
lnst
SO:
1243- 1257.
Mantel N
(1
966):
Evaluation of minimal data and two new rank order statistics arising in its consid-
eration. Cancer Chemother Rep 50: 163-170.
Mantel N, Haenszel W (1959): Statistical aspects of the analysis of data from retrospective studies
of
disease.
J
Natl Cancer Inst 22:719-748.
Masoro EJ (1980): Mortality and growth characteristics of rat strains commonly used in aging research.
Exp Aging Res 6:219-233.
Mayers CP, Habeshaw JA
(1973):
Depression of phagocytosis: A non-thermal effect of microwave
radiation as a potential hazard to health. Int J Radiat Biol Phys 24:449-461.
McKnight
B,
Crowley
J
(1984): Tests for differences in tumor incidences based on animal carcinogenesis
experiments.
J
Am Stat Assoc 79(387):639-648.
NCRP
(
1986): Biological effects and exposure criteria for radio frequency electromagnetic fields. Report
86. Bethesda, Maryland: National Council
of
Radiation Protection and Measurements.
Phillips RD, Hunt EL, Castro RD, King NW
(I
975): Thermoregulatory, metabolic, and cardiovascu-
lar response of rats to microwaves.
J
Appl Physiol 38(4):630-635.
496
Chou
et
al.
Polk
C,
Postow
E
(1986): “Handbook
of
Biological Effects
of
Electromagnetic Fields.” Boca Raton,
Florida: CRC Press.
Sacher GA, Duffy PH (1978): Age changes in rhythms of energy metabolism, activity and body tem-
perature
in
Mus and Pcromyscus. In Samis
HV,
Jr.,
Capobianco
S
(eds): “Biological Rhythms.”
New York: Plcnum Publishing Corp, pp 105-124.
Shandala MG, Dumanski
UD,
Rudnev MI, Ershova
LK,
Los
IP
(1979): Study
of
nonionizing micro-
wave radiation effects upon the central nervous system and behavior reactions. Environ Health
Perspect
30:115-121.
Toler
J,
Popovic
V,
Bonasera
S,
Popovic P, Honeycutt
C,
Sgoutas
D
(1988):
Long-term study
of
435
MHz
radio-frequency radiation
on
blood-borne end points in cannulated rats, Part
I:
Engineer-
ing consideration, and Part
11:
Methods, results and summary.
J
Microw Power EM Energy
23(2):95-136.
Tyler PE (ed)
(1975):
Biological effects
of
nonionizing radiation. Ann
NY
Acad Sci 247:
5-545.
U.S.
EPA (1986): Guidelines
for
carcinogenicity risk assessment. Fed Register
5
1:33992-34003.
Walsh RN, Cummins RA (1976): The open-field
test:
A critical review. Psycho1 Rep 83(3):482-504.
Ward
JM
(1983):
Background data and variations in tumor rates of control rats and mice.
Prog
Exp
Tumor Res 26:241-258.
Wiktor-Jedrzejczak W, Ahmed
A,
Czerski P, Leach
WM,
Sell KW (1977): Immune response
of
mice
to
2450
MHz microwave radiation: Overview
of
immunology and empirical studies
of
lymphoid
splenic cells. Radio Sci 12(6S):209-219.
Zimrnerrnann
E,
Crutchlow
V
(1967): Effects
of
diurnal variations in plasma corticosterone levels on
adrenocortical response to stress. Proc SOC Ext Biol Med I25:658-663.
Zucker
L
(1971):
Light-dark rhythms in rat eating and drinking behavior. Physiol Behav 6: 115-126.
