AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 39:499±504 (2001)
Pregnancy Outcome Following Exposure
to Shortwaves Among Female
Physiotherapists in Israel
Yehuda Lerman, MD, MPH,1;3 Ruben Jacubovich, MD,1 and Manfred S. Green, MD, PhD2,3
Background The ®ndings of the few epidemiological studies on the possible association
between shortwave diathermy use by pregnant physiotherapists and adverse pregnancy
outcome are inconsistent. We investigated such an association among physiotherapists in
Israel.
Methods Individualized data on exposure to shortwaves, ultrasound, and heavy lifting
were collected by questionnaires and telephone interviews.
Results The 434 studied women included 930 pregnancies: 175 ended in spontaneous
abortions, 45 had fetal malformations, 47 were delivered prematurely, and 33 infants had
low birth weight. The remaining 630 normal pregnancies comprised the control group.
Univariate analysis showed that exposure to shortwaves was associated with a
signi®cantly increased odds ratio (O.R.) for congenital malformations (O.R. 2.24, CI
1.27±4.83, P ˆ.006) and low birth weight (O.R. 2.99, CI 1.32±6,79, P ˆ.006). This effect
increased in a dose-related manner. After controlling for potential confounding variables,
only low birth weight reached statistical signi®cance (O.R. 2.75, CI 1.07±7.04, P ˆ.03).
From the potentially confounding variables tested, febrile disease during pregnancy was
found to be signi®cantly associated with low birth weight (O.R. 3.37, CI 1.38±8.25,
P ˆ.01).
Conclusions The ®ndings of our study suggest that shortwaves have potentially harmful
effects on pregnancy outcome, speci®cally low birth weight. Am. J. Ind. Med. 39:499±
504, 2001. ß 2001 Wiley-Liss, Inc.
KEY WORDS: electromagnetic radiation; shortwaves; diathermy; pregnancy outcome; birth weight; prematurity; spontaneous abortion; congenital malformations;
physiotherapy
INTRODUCTION
Shortwave diathermy units heat tissue through absorption of energy from electromagnetic ®elds. Shortwaves are a
1
Occupational Health and Rehabilitation Institute, Ra'annana, Tel-Aviv University, Tel-Aviv,
Israel
2
The Israel Center for Disease Control, Gertner Institute,Tel-Hashomer,Tel-Aviv University,
Tel-Aviv, Israel
3
The Sackler Faculty of Medicine,Tel-Aviv University,Tel-Aviv, Israel
*Correspondence to: Prof. Yehuda Lerman, Occupational Health Center, Sherutei Beriut
Clalit, 101Arlozorov St.,Tel-Aviv 62098, Israel. E-mail: ylerman@post.tau.ac.il
Accepted 30 January 2001
ß 2001Wiley-Liss, Inc.
fraction of the high-frequency band of the non-ionizing
electromagnetic spectrum. The wave frequency is typically
27.12 MHz. These devices are widely used in physiotherapy
departments. Shortwaves have been shown to cause abortion
and abnormal fetal development in animal models [Brown
Woodman et al., 1989]. A search of the literature revealed
sparse epidemiologic studies on the possible association
between shortwave diathermy use by pregnant physiotherapists and adverse pregnancy outcome, and their conclusions
were inconsistent.
Kallen et al. [1982] showed an excess risk of congenital
malformations among the offspring of female physiotherapists who reported working with shortwave diathermy.
500
Lerman et al.
Taskinen et al. [1990] reported an excess risk of spontaneous
abortion after the 10th week of gestation among Finnish
physiotherapists who used ultrasound and shortwave
diathermy, but this disappeared when confounders were
controlled. In a parallel study, the frequency of using
shortwaves was signi®cantly associated with congenital
malformations but, interestingly, only in the lower exposure
group. Larsen [1991] and Larsen et al. [1991] found no
signi®cant association between shortwave exposure and
congenital malformations or spontaneous abortions, although they did note a non-signi®cant association between
shortwave exposure and low birth weight among male
births. Quellet-Hellstrom and Stewart [1993] found no
signi®cant association between shortwave use and spontaneous abortions. Finally, Guberan et al. [1994] found no
association between shortwave exposure and low birth
weight.
The results of the works reviewed above are inconsistent in answering the question of whether occupational
exposure to physiotherapists is correlated with different
pregnancy outcomes. Moreover, in some of them, the study
design was that of a simple case control study in which
biases are inherent. Also, each of the previous studies
analyzed only one or two of the possible adverse pregnancy
outcomes. We initiated this study in order to assess whether
exposure to shortwave radiation is associated with the four
adverse pregnancy outcomesÐspontaneous abortions, congenital malformations, prematurity, and low birth weightÐ
within the same group of Israeli physiotherapists. While we
included other variables which might affect the pregnancy
outcome, our primary objective was to test the hypothesis
that shortwave exposure is a reproductive hazard.
MATERIALS AND METHODS
A detailed questionnaire was sent by mail to all
physiotherapists registered as members of the Union of
Israeli Physiotherapists. Not all of the registered physiotherapists are actively working in the ®eld, nor is it
known what is the percentage of those who are. The target
population included Israeli female physiotherapists who had
ever been pregnant. We assembled a cohort of female
physiotherapists who answered the anonymous questionnaire and provided information on their reproductive
history. The investigation was designed as a prevalent case
control study. In addition, all physiotherapy departments
associated with the sick funds, hospitals, and private clinics
were identi®ed and interviewed by telephone in order to
achieve a greater response rate. The questionnaire and
interview included queries on the outcome of pregnancy and
the use of shortwave diathermy, ultrasound, and heavy
lifting during pregnancy. The questionnaire was also mailed
to male physiotherapists: it included questions on orthopedic problems and was anonymous in order to minimize the
selection bias associated with overrepresentation of female
physiotherapists who had abnormal delivery outcome.
Study Population
The groups were divided according to the following
pregnancy outcome:
Spontaneous abortionÐoccurring before the 28th week of
gestation (n ˆ 175);
Congenital malformationsÐaccording to the ICD codes
740±759.99 (n ˆ 45);
PrematurityÐdelivery before the 36th week of gestation
(n ˆ 47); and
Low birth weightÐless than 2500 g and delivered after the
36th week of gestation (n ˆ 33).
A mother could contribute more than one pregnancy as
a ``case''.
Controls
The control group (n ˆ 633) consisted of pregnancies of
mothers who reported no adverse reproductive outcome
(spontaneous abortion, congenital malformations, prematurity and low birth weight) in any of their pregnancies.
Normal pregnancies of mothers who had other pregnancies
that ended in an abnormal delivery were excluded.
Exposure Measurements
The duration of exposure to shortwaves and ultrasound
per week was estimated on a scale in which 0 ˆ no exposure,
1 ˆ less than 10 hours/week, and 2 ˆ more than 10 hours/
week. For exposure to heavy lifting, the scale was 0 ˆ no
exposure, 1 ˆ 5±25 times/week, 2 ˆ more than 25 times/
week.
Other Retrieved Data
The questionnaire and telephone interview also included information about conditions during pregnancy which
can act as confounders, such as febrile disease, consumption
of alcohol, use of drugs and tobacco during pregnancy, age
of the mother at each pregnancy, previous abortions and
number of pregnancies.
Statistical Methods
Crude relationships between pregnancy outcome and
exposure to shortwaves, ultrasound, and heavy lifting were
estimated by the odds ratio (O.R.) together with the exact
95% con®dence interval (95% CI). The 2 and Fisher exact
tests were used to test statistical signi®cance. Multiple
Pregnancy Outcome of Shortwave-Exposed Physiotherapists
logistic regression analysis was used to control for
confounders.
TABLE II. The Odds Ratio (O.R.) of Congenital Malformations for Individual
Occupational Variables Among Israeli Female Physiotherapists
RESULTS
There was a total of 774 questionnaires which
contributed 1416 pregnancies. The 434 normal pregnancies
from mothers who had reproductive hazards in one or more
of their pregnancies were excluded from the study (nor were
they included in the control group). Another 49 pregnancies
from 37 women were excluded because the mother's age
was not available. The ®nal study population was comprised
of 434 women who contributed 933 pregnancies: 175
pregnancies in the spontaneous abortion group; 45 in the
congenital malformation group; 47 in the prematurity
group; and 33 in the low birth weight group. The remaining
633 pregnancies from women who reported no adverse
outcome in any pregnancy served as the control group.
Spontaneous Abortion
The O.R. between spontaneous abortion and exposure
to shortwaves, ultrasound and heavy lifting are presented in
Table I. Shortwaves did not increase the risk of spontaneous
abortion (O.R. ˆ 0.9). The O.R. was decreased for handling
ultrasound equipment and for heavy lifting. The O.R. of the
confounding factors showed no signi®cant association. In
the multivariate analysis, there was no change in the O.R.
for shortwaves.
TABLE I. The Odds Ratio (O.R.) of Spontaneous Abortion for Individual Occupational Variable Among Israeli Female Physiotherapists
Spontaneous abortion
Exposure index
Shortwaves
No exposure
Exposed
1^9 hours/week
> 10 hours/week
Heavy lifting
No exposure
Exposed
5^25 times/week
> 25 times/week
Ultrasound
No exposure
Exposed
1^9 hours/week
> 10 hours/week
Cases
(n)
Controls
(n)
O.R.
95% CI
P
92
80
60
20
323
310
222
88
0.90
0.94
0.79
0.64^1.27
0.65^1.37
0.46^1.36
0.56
0.78
113
59
38
21
306
327
216
111
0.48
0.47
0.51
0.34^0.69
0.31^0.71
0.30^0.85
0.00
0.00
0.01
92
80
36
44
230
403
225
178
0.49
0.40
0.60
0.35^0.69
0.26^0.61
0.41^0.93
0.00
0.00
0.02
501
Congenital
malformations
Exposure index
Shortwaves
No exposure
Exposed
1^9 hours/week
> 10 hours/week
Heavy lifting
No exposure
Exposed
5^25 times/week
> 25 times/week
Ultrasound
No exposure
Exposed
1^9 hours/week
> 10 hours/week
Cases
(n)
Controls
(n)
O.R.
95% CI
P
13
28
16
12
323
310
222
88
2.24
1.79
3.40
1.27^4.83
1.04^4.33
1.50^7.40
0.006
0.038
0.002
20
21
15
6
306
327
216
111
0.98
1.06
0.82
0.60^2.07
0.65^2.46
0.32^2.11
0.71
0.47
0.69
6
34
15
19
230
403
225
178
3.23
2.55
4.09
1.50^8.68
1.26^8.25
1.60^10.45
0.002
0.01
0.002
Congenital Malformations
The O.R. for congenital malformations for individual
occupational variables are presented in Table II. Increased
exposure to ultrasound and shortwaves increased the O.R.
signi®cantly. The O.R. of heavy lifting and the confounding
factors showed no signi®cant association with congenital
malformations. In the multivariate analysis, only ultrasound
exposure showed a signi®cant association with congenital
malformations (O.R. 4.19, 95% CI 1.58±11.13, P ˆ 0.004).
The odds ratio of exposure to shortwaves decreased to 1.33
(95% CI 0.68±2.75, P ˆ 0.44).
Prematurity
No signi®cant association was found between prematurity and exposure to shortwaves, ultrasound, and heavy
lifting (Table III).
Low Birth Weight
Exposure to shortwaves increased the O.R. signi®cantly, and it increased with increasing exposure (Table IV).
From the potential confounding factors, febrile disease
during pregnancy increased the O.R signi®cantly (O.R. ˆ
3.37, 95% CI 1.38±8.25, P ˆ 0.01). In the multivariate
analysis, exposure to shortwaves showed a signi®cant
502
Lerman et al.
TABLE III. The Odds Ratio (O.R.) of Prematurity for Individual Occupational
Variables Among Israeli Female Physiotherapists
TABLE IV. The Odds Ratio (O.R.) of Low Birth Weight for Individual Occupational Variables Among Israeli Female Physiotherapists
Prematurity
Exposure index
Shortwaves
No exposure
Exposed
1^9 hours/week
> 10 hours/week
Heavy lifting
No exposure
Exposed
5^25 times/week
> 25 times/week
Ultrasound
No exposure
Exposed
1^9 hours/week
> 10 hours/week
Cases
(n)
Low birth weight
Controls
(n)
O.R.
95% CI
P
25
21
18
3
323
310
222
88
0.87
1.04
0.84
0.48^1.59
0.55^1.96
0.13^1.49
0.66
0.88
0.17
28
18
13
5
306
327
216
111
0.60
0.65
0.49
0.32^1.11
0.33^1.29
0.18^1.30
0.10
0.22
0.14
22
24
14
10
230
403
225
178
0.60
0.60
0.60
0.58
0.31^1.10
0.33^1.35
0.27^1.27
0.10
0.20
0.17
association with low birth weight (O.R. ˆ 2.75, 95% CI
1.07±7.04, P ˆ 0.03). The O.R. for febrile disease was
increased but with no statistical signi®cance (O.R. 2.73,
95% CI 1.06±6.99, P ˆ 0.1). The gender-speci®c analysis
indicated that the risk for low birth weight was greater for
boys than for girls (O.R. 3.7 vs. O.R. 2.9, respectively).
DISCUSSION
In the present retrospective nested case-control study,
exposure to shortwave was found to be associated with low
birth weight and with an increased risk for male infants.
These results are in agreement with those of Larsen et al.
[1991] that showed an O.R. of 5.9 (95% CI ˆ 1.0±28.6) for
low birth weight of boys in the high exposure group versus
an O.R. of 0.7 (95% CI ˆ 0.0±3.6) for girls. In the same
cohort, Larsen [1991] showed that the crude association
between the indices of exposure to high frequency electromagnetic radiation and the occurrence of malformations
were weak and non-signi®cant because the CIs were wide
and included unity.
Our results also showed that congenital malformations
were not found to be associated signi®cantly with exposure
to shortwaves during pregnancy after controlling for the
different confounders. Ultrasound equipment was found to
be associated with congenital malformations. Different
results were reported among physiotherapists in Sweden by
Exposure index
Shortwaves
No exposure
Exposed
1^9 hours/week
> 10 hours/week
Heavy lifting
No exposure
Exposed
5^25 times/week
> 25 times/week
Ultrasound
No exposure
Exposed
1^9 hours/week
> 10 hours/week
Cases
(n)
Controls
(n)
O.R.
95% CI
P
8
23
14
9
323
310
222
88
2.99
2.54
4.12
1.32^6.79
1.05^6.17
1.54^11.01
0.006
0.033
0.002
16
15
8
7
306
327
216
111
0.87
0.70
1.20
0.42^1.80
0.29^1.68
0.48^3.01
0.72
0.43
0.68
8
23
11
12
230
403
225
178
1.64
1.40
1.90
0.72^3.70
0.55^3.55
0.77^4.84
0.230
0.47
0.15
Kallen et al. [1982] who demonstrated that the use of
shortwave equipment ``often'' or ``daily'' was more common
among the cases with malformed or perinatally dead infants
than the controls with normal infants (33 vs. 14%, P ˆ 0.03).
The ®ndings of this study can be random insofar as
statistical signi®cance is borderline and multiple comparisons were performed for all subjects. However, it should be
borne in mind that the primary objective of this study was to
test the hypothesis of one of them, i.e., of a possible harmful
effect on pregnant physiotherapists working with shortwave
equipment.
We cannot exclude the possibility that exposure to
shortwave and ultrasound were closely associated or that
they were associated with a third unidenti®ed exposure that
could be harmful. We controlled for several confounders
during the index pregnancy, the ones that were found to
be risk factors for adverse pregnancy outcomes, such as the
mother's age during each pregnancy, heavy lifting, febrile
disease, use of drugs and tobacco, consumption of alcohol
during pregnancy, and number of pregnancies. None of them
could explain the tendencies of risk associated with
shortwave exposure.
We could not obtain information from medical records,
thus the possibility of misclassi®cation bias cannot be
excluded. The information on pregnancy outcome and on
the different exposures was based on the physiotherapists'
reports, so information bias might have been responsible
Pregnancy Outcome of Shortwave-Exposed Physiotherapists
for the apparent associations, especially since the hypothesis
of the hazard of work with shortwave equipment may
well have been known to some of the physiotherapists
beforehand. However, two aspects of the results tend to
negate this bias. One is the dose±response relationship,
i.e., the O.R. for congenital malformations and low birth
weight increased with increasing doses of exposure. The
second is that we would expect the same trend with
spontaneous abortions and prematurity, and this was not the
case.
Recall bias of the exposures cannot be excluded since
varying periods of time had passed since exposure. Exposure
evaluation is traditionally based on indirect data, such as
those from the kind of interviews conducted by us and by
others. For example, Larsen and Skotte [1991] compared the
assessment of shortwave exposure based upon observations which included objective measurements of exposure
and compared them to the information given by exposed
female diathermy operators by means of interviews. According to this study, it was possible to discriminate between
recent high and low peak exposure. Moreover, an interview
index re¯ecting the duration of the exposure correlated to
some extent with the corresponding measurements of that
study.
We had no access to the registration of all certi®ed
physiotherapists in Israel, and we cannot arrive at any
conclusions about the compliance of this studied group,
thus there is a potential selective participation bias. We took
three measures to minimize such a selection bias and to
increase participation among physiotherapists: (1) we
offered the participants the possibility of answering the
questionnaire anonymously; (2) questionnaires were also
mailed to male physiotherapists (not included in any
analysis); and (3) non-relevant questions (e.g., on orthopedic problems at work) were included in the questionnaire.
One obvious issue related to the de®nition of the source
population is that some pregnant female physiotherapists
may stop working altogether and, therefore, they may drop
off the roster of Israeli physiotherapists. This has the
potential to bias the ®ndings toward the null if, under
the assumption of a deleterious effect of shortwaves, these
women were exposed early in their pregnancy, experienced
bleeding or some other symptom due to the exposure and
then subsequently stopped working and dropped out of the
target population. In spite of the potential for a healthy
worker effect and its bias toward the null, we provided
evidence in this study for the deleterious effects of shortwave on some of the pregnancy outcomes.
We had explored exposure to shortwaves in the Israeli
physiotherapy departments in an earlier investigation. The
electromagnetic ®eld strengths from shortwave diathermy
equipment in six physiotherapy departments produced
values above the reference levels (61.4 v/m electric ®eld
and 0.18 A/m magnetic ®eld) which were detected up to one
503
meter from electrodes and cables when shortwave equipment was used in the continuous mode [Lerman et al.,
1996].
Little is known about the mechanisms by which shortwaves might harm the pregnancy or cause fetal damage. In
mammals, shortwaves are teratogenic and embryolethal at
intensities which produce signi®cant hyperthermia in the
dams [Lary et al., 1986]. Others have provided evidence
which indicates that increases in mortality and resorption
are probably related to peak body temperature and its
duration regardless of the method by which the temperature
elevation is elicited [Michaelson and Elson, 1996]. Thus,
thermal stress appears to be the primary mechanism by
which radiofrequency energy absorption exerts a teratogenic
action. It is not known whether occupational exposure of
physiotherapists could produce hyperthermia in the fetus of
pregnant women. In experiments on rats, hyperthermia
during speci®c developmental stages produces not only fetal
resorption or malformation but also retardation of growth.
This might explain an association between exposure to
shortwaves early in pregnancy and low birth weight, as was
shown in this study, after controlling for possible maternal
lifestyle factors also associated with low birth weight, such
as smoking. Finally, a non-thermal effect of shortwave
exposure may also explain the phenomenon [Tofani et al.,
1986].
In conclusion, this study provides further support to
earlier reports that occupational exposure of female
physiotherapists to shortwave radiation during pregnancy
could have detrimental effects on pregnancy outcome, and
that shortwave use during pregnancy may be considered as
being a potential reproductive hazard.
REFERENCES
Brown Woodman PDC, Hadley JA, Richardson L, Bright D, Porter D.
1989. Evaluation of reproductive function of female rats exposed to
radiofrequency ®elds (27.12 MHz) near a short wave diathermy device.
Health Phys 56:521±525.
Guberan E, Campana A, Faval P, Guberan M, Sweetriam PM, Tuyn J,
Usel M. 1994. Gender ratio of offspring and exposure to shortwave
radiation among female physiotherapists. Scand J Work Environ
Health 20:345±348.
Kallen B, Malmquist G, Moritz U. 1982. Delivery outcome among
physiotherapists in Sweden: Is non-ionizing a fetal hazard? Arch
Environ Health 37:81±84.
Larsen A. 1991. Congenital malformations and exposure to highfrequency electromagnetic radiation among Danish physiotherapists.
Scand J Work Environ Health 17:318±323.
Larsen A, Olsen J, Svane O. 1991. Gender-speci®c reproductive
outcome and exposure to high frequency electromagnetic radiation
among physiotherapists. Scand J Work Environ Health 17:324±
329.
Larsen A, Skotte J. 1991. Can exposure to electromagnetic radiation in
diathermy operators be estimated from interview data?: A pilot study.
Am J Ind Med 15:51±57.
504
Lerman et al.
Lary JM, Conorer DL, Johnson PH, Hornung RW. 1986. Dose response
relationship between body temperature and birth defects in radiofrequency-irradiated rats. Bioelectromagnetics 7:141±149.
Lerman Y, Jacubovich R, Caner A, Ribak J. 1996. Electromagnetic
®elds from shortwave diathermy equipment in physiotherapy departments. Physiotherapy 82:456±458.
Michaelson SM, Elson EC. 1996. Interaction of nonmodulated and
pulse modulated radiofrequency ®elds with living matter: experimental
results. In: Polk C, Postow E, editors. Handbook of biological effects of
electromagnetic ®elds, 2nd Ed. New York: CRC Press.
Quellet-Hellstrom R, Stewart WF. 1993. Miscarriages among female
physical therapists who report using radio-and microwave frequency
electromagnetic radiation. Am J Epidemiol 10:775±785.
Taskinen H, Kyyronen P, Hemminki K. 1990. Effects of ultrasound
shortwaves and physical exertion on pregnancy outcome in physiotherapists. J Epidemiol Community Health 44:196±201.
Tofani S, Agnesod G, Ossola P, Ferrini S, Bussi R. 1986. Effects of
continuous low-level exposure to RF radiation on intrauterine development in rats. Health Phys 51:489±499.