Journal Article Analysis
Directions: Respond to each question in one or two sentences
Points
1. What public health problem is the study described in this article attempting to
address?
2. What is the research hypothesis that this study aims to test?
3. What type of epidemiologic study is this?
4. What is the independent variable in this study?
5. What is the (primary) dependent variable in this study?
6. How were cases of brain cancer ascertained in this study?
7. What broader population do you believe the sample of cases represents? (this
will determine the population to which the results of the study can ultimately be
generalized)
8. What exclusion criteria were used?
9. What kind of comparison group was used in this study? How were they selected?
Of what broader population is this group intended to be representative?
10. What measure of association was calculated in this study? What statistical test
was used to test its significance?
11. What did the study find?
12. What practical application do the results of this study have for clinical care of
patients?
13. What practical application do the results of this study have for public health
policy in Canadian provinces?
© International Epidemiological Association 2002
Printed in Great Britain
International Journal of Epidemiology 2002;31:210–217
Brain cancer and occupational exposure to magnetic
fields among men: results from a Canadian
population-based case-control study
Paul J Villeneuve,a,b David A Agnew,c Kenneth C Johnson,a Yang Maoa and the Canadian Cancer
Registries Epidemiology Research Groupd
Background
The relationship between occupational exposure to magnetic fields and brain cancer in men
was investigated using population-based case-control data collected in eight Canadian
provinces. Emphasis was placed on examining the variations in risk across different
histological types.
Methods
A list of occupations was compiled for 543 cases and 543 controls that were individually
matched by age. Occupations were categorized according to their average magnetic field
exposure through blinded expert review (,0.3, 0.3–,0.6, and >0.6 µT). In total, 133 cases
(14%) and 123 controls (12%) were estimated to have at least one occupation whereby
magnetic field exposures exceeded 0.3 µT. Odds ratios (OR) were generated using
conditional logistic regression, and were adjusted for suspected occupational risk factors
for brain cancer.
Results
A non-significantly increased risk of brain cancer was observed among men who had ever
held a job with an average magnetic field exposure .0.6 µT relative to those with exposures
,0.3 µT (OR = 1.33, 95% CI : 0.75–2.36). A more pronounced risk was observed among
men diagnosed with glioblastoma multiforme (OR = 5.36, 95% CI : 1.16–24.78).
Moreover, a cumulative time weighted index score of magnetic field exposure was
significantly related to glioblastoma multiforme (P = 0.02). In contrast, magnetic field
exposures were not associated with astrocytoma or other brain cancers.
Conclusions Our findings support the hypothesis that occupational magnetic field exposure increases the
risk of glioblastoma multiforme.
Magnetic fields, brain cancer, occupation
a Environmental Risk Assessment and Case Surveillance Division, Laboratory
Correspondence: Dr Paul Villeneuve, Department of Epidemiology and Community
Medicine, University of Ottawa, 451 Smyth Road, Ottawa, Ontario, Canada K1H
8M5. E-mail: pvillene@uottawa.ca
Centre for Disease Control, Health Canada, Ottawa, Ontario, Canada K1A 0L2.
b Department of Epidemiology and Community Medicine, University of Ottawa,
451 Smyth Road, Ottawa, Ontario, Canada K1H 8M5.
c
Department of Public Health Sciences, University of Toronto, 12 Queen’s Park
Crescent, Toronto, Ontario, Canada M5S 1A8.
d The Canadian Cancer Registries Epidemiology Research Group comprises a
Principal Investigator from each of the Provincial Cancer Registries involved in
the National Enhanced Cancer Surveillance System: Bertha Paulse, Newfoundland
Cancer Foundation; Ron Dewar, Nova Scotia Cancer Registry; Dagny Dryer,
Prince Edward Island Cancer Registry; Nancy Kreiger, Cancer Care Ontario; Erich
Kliewer, Cancer Care Manitoba; Diane Robson, Saskatchewan Cancer
Foundation; Shirley Fincham, Division of Epidemiology, Prevention and
Screening, Alberta Cancer Board; and Nhu Le, British Columbia Cancer Agency.
The aetiology of brain tumours is not well understood. Ionizing radiation
and a genetic predisposition have been implicated as risk factors,
however, they are thought to account for a small proportion of all such
tumours.1 Positive associations between brain cancer and other
occupational exposures such as vinyl chlorides, 2–4 pesticides5 and
electromagnetic fields6 have been observed in some studies, but, taken as
a whole, the results are inconclusive. Efforts to clarify the role of these
factors are needed, particularly in light of the extremely poor prognosis
for patients diagnosed with these neoplasms.7
During the past decade, a number of studies have examined the
relationship between occupational exposure to magnetic fields and the
occurrence of brain tumours. Several of these studies were performed
within electric utility industry workers and incorporated detailed
210
BRAIN CANCER AND OCCUPATIONAL EXPOSURE TO MAGNETIC FIELDS 211
exposure assessments obtained using either personal monitoring devices
or other sampled measures taken from relevant work-sites.8–15 Despite
elaborate efforts to characterize exposure to 50/60 Hz power frequency
magnetic fields, the findings of these studies have been equivocal. This
may partly be due to the size of the cohorts that have typically yielded a
small number of cases, and consequently, limited the power of the study
to detect effects. Inconsistent results have also been obtained from a series
of population-based casecontrol studies that investigated the association
between occupational magnetic field exposures and brain cancer.16–21
Many of these studies did present risk estimates across different
histological types of brain cancer. However, several were limited by
either
small
sample
sizes,
8,10,12,17
crude
assessment
of
exposure,16,18,22,23 incomplete occupational history,19 lack of data on
Histological type
Astrocytoma
Glioblastoma multiforme
Other
Unknown
ICD-O Codes 1991a
9384, 9400–9421
9440–9442
No. of cases
214
198
9380, 9382, 9391, 9392,
9424, 9430, 9450, 9451,
9460, 9470, 9473
115
8000, 8010, 8900,
9150, 9505, 9990
16
Total
543
a ICD-O codes given by the World Health Organization. 31
potential occupational confounders,11,17,23 and the use of decedent rather
than incident cases.13,16,18,24,25 The identification of brain cancer cases
using death certificates is particularly problematic as such tumours may
represent a metastatic spread from a cancer that originated at another
anatomical site.26
The results from both in vivo and in vitro studies suggest that if
exposures to 60 Hz magnetic fields increase the risk of cancer, it is
through the promotional stage of the carcinogenic process. In the
traditional multistage model, tumour promotion is regarded as an
extended process that requires prolonged or repeated exposure to the
promoting agent.27 Continued exposure to promoting or co-promoting
agents after tumour development may cause the tumour to evolve with
increased metastatic properties.28 Tumour promoters are characterized by
the existence of a threshold, prolonged exposure and reversibility of
effects.29 The present study was undertaken to explore the relationship
between occupational magnetic field exposures and different histological
types of brain cancer. Elevated risks for more aggressive subtypes of brain
cancer would support the hypothesis that magnetic fields act as tumour
promoter.
Using data collected through the Canadian National Enhanced Cancer
Surveillance System (NECSS), we examined the relationship between
occupational exposure to magnetic fields and brain cancer using several
different exposure indices. Occupational magnetic field exposure
assessment was performed by an expert review that was blinded to the
case-control status of the subjects. An important strength of this study is
the ability to derive magnetic field exposures indices that take into
account the complete occupational history of each subject. Perhaps more
importantly, the size of the study is sufficiently large to perform risk
assessment across different histological types of brain cancer.
Subjects and Methods
The NECSS was designed to investigate environmental causes of cancer
using population-based data. The case-control component of the NECSS
collected data between January 1994 to August 1997 in eight Canadian
provinces (Newfoundland, Prince Edward Island, Nova Scotia, Ontario,
Manitoba, Alberta, Saskatchewan and British Columbia). The collection
of data was conducted through the co-operation of Health Canada and the
provincial cancer registries. All brain cancer cases included in the NECSS
were confirmed histologically and cases were defined according to the
International Classification of Diseases, Ninth Revision (ICD-9) rubric
191.30 Benign brain tumours were not included in these analyses.
Analyses are based on a total of 543 brain cancer cases that were
categorized by histological type using the
Table 1 Brain cancers ascertained among men in the National Enhanced
Cancer Surveillance System (NECSS) case-control study, by histological
type, 1994–1997
International Classification of Diseases for Oncology (ICD-O)31 using
the codes shown in Table 1.
The participating provinces attempted to identify eligible brain
cancer cases as early as possible in the registration process in order to
minimize the loss of subjects due to severe illness or death. Of these
eligible cases, data were not collected among those who had died
(23%) or for whom physician consent was not granted (10.2%).
Among those cases that were sent questionnaires 63% were
completed, while the corresponding response rate from the control
population was approximately 65%.
Frequency matching was employed by the investigators of the
NECSS to select population-based controls so as to achieve a similar
age and sex distribution to all cancer cases. There were subtle
differences in the methods that were used to select controls in each
participating province. In Prince Edward Island, Nova Scotia,
Manitoba, Saskatchewan and British Columbia, provincial health
insurance plans were used to obtain a random age- and sex-stratified
sample of the provincial population. In each of these provinces, more
than 95% of residents are covered by these public plans; those
excluded include current military personnel and their families and
indigenous peoples who are covered by other plans. Newfoundland
and Alberta used random digit-dialling to recruit controls while
Ontario used Ministry of Finance data to create a stratified random
sample.
Mailed questionnaires were used to obtain information on subjects’
residential and occupational histories and on other risk factors for
cancer. When necessary, telephone follow-up was used to clarify
responses. The NECSS questionnaire was designed to collect data on
ethnicity, education, income, smoking, height, weight, exposure to
specific occupational carcinogens, physical activity, diet two years
before interview (60-item foodfrequency questionnaire) and general
changes in diet compared with 20 years ago. Subjects were asked
whether they had ever been occupationally exposed to 17 different
agents. Of these, the following exposures have been identified as
possible risk factors for brain cancer: pesticides, herbicides, radiation
sources, and vinyl chlorides.
Each subject was asked to report on all the jobs they had held for at
least one year and all Canadian residences that they had lived in for at
least one year. For each job, subjects were asked to describe their jobtitle, company name, work location, duties, the starting and ending
calendar year of employment, and information on exposure to
workplace odours and tobacco smoke. Residential data that was
collected included address, the occupancy period and the source of
water, the type of heating used and the number of smokers they lived
with.
212 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Although the NECSS also collected data among women, we
decided to restrict magnetic field exposure assessment and analysis to
men for several reasons. First, because most occupational studies of
electromagnetic fields have been conducted using male workers, there
was limited data to characterize occupational exposures to magnetic
fields for women. Second, restricting analyses among men facilitated
comparisons with previously published studies. Finally, as the median
(or mean) age of the female brain cancer cases was 52 years, it was
anticipated that as a whole, there would be little variation in
occupational exposure to magnetic fields as few women in this
populationbased study would have been employed in occupations
characterized by jobs with high magnetic field exposures in their
distant past.
It was determined a priori that occupational magnetic field
exposure would not be assigned by using occupational coding, but
rather through a manual inspection for each subject of several key
variables through expert review. Controls were individually matched
to cases because it would have been quite onerous to code all
occupations held by the entire control population, and the matching
procedure ensured that the age distributions of the case and control
populations were similar. Specifically, one control was randomly
selected for each case and matched within a single year of age. In
total, 543 controls were chosen in this manner from the pool of 4823
NECSS controls with completed questionnaire data.
A list of all the occupations held was compiled for the cases and
matched controls. Each occupation was assigned an exposure value
based on a time-weighted average magnetic flux density for full-time
workers. This exposure assessment also incorporated questionnaire
data that were collected on the job duties and the employment
location. The categories of average exposure were: ,0.3, 0.3–,0.6, and
>0.6 µT. The lower cutpoint of 0.3 µT was chosen to provide
reasonable assurance that occupational exposures in the upper two
categories were greater than background exposure levels that workers
receive at home. Information about the distribution of residential
exposures was obtained from a Canadian study of residential
magnetic field exposures and childhood leukaemia.32 It has been
estimated that the cutpoint of 0.3 µT corresponded to the 82th
percentile for adult exposures in the same homes.33 The occupational
exposure categorizations were derived through expert review (D
Agnew) of the employment variables described above and were
performed blinded to case-control status. There were a total of 3808
unique character string job title descriptors. The assignment of
exposure relied on results from published reports9,34–37 and
consultations with occupational hygienists specializing in the area of
electromagnetic fields. For some occupations that could not be readily
classified, field measurements were performed using a Drexel
Corporation Magnum 310 magnetic field monitor. The upper 0.6 µT
limit was chosen as it was double the lower cutpoint, and split the job
titles with .3 mG into two groups with number of job titles in the ratio
of 2:1. Examples of highly exposed occupations (>0.6 µT) included:
sheet metal workers, telephone cable splicer, projectionists (motion
pictures), welders, electricians, electronic assemblers, and electric
utility workers. Incomplete questionnaire data prevented us from
classifying 42 (1.3%) of the occupations held by the study subjects.
Odds ratios (OR) were estimated using conditional logistic regression
which took into account the matched design of the study. Five different
magnetic field exposure indices were modelled. These included the
highest average occupational exposure to magnetic fields (,0.3, >0.3, >0.6
µT) and the magnetic field exposure received in the job held the longest
(,0.3, 0.3–,0.6 and >0.6 µT). To evaluate the effect of magnetic field
exposures received early or later on in life, we calculated the risk of brain
cancer based on exposure categorizations for subjects’ first and last held
jobs. The last index we examined was a cumulative time-weighted
occupational magnetic field exposure score that was calculated by taking
into account exposure at each job (E), the duration of employment (D)
and whether the work was full-time (F). Mathematically, the cumulative
index score was calculated as follows:
i=j
MFindex = ∑ Ei × Di × Fi
i=1
where E = 0 for jobs with average occupational exposures of ,0.3 µT
= 1 for jobs with average occupational exposure 0.3– ,0.6 µT
= 2 for jobs with average occupational exposure >0.6 µT j =
the total number of jobs held
D = duration of employment (in years) and F = 1
for full-time employment
= 0.5 for part-time or seasonal employment.
Several variables were evaluated to determine whether they confounded
the results. These included self-reported occupational exposures to vinyl
chloride, herbicides, pesticides and radiation sources. Similar to the
assignment of magnetic field exposures, an index of exposure to ionizing
radiation was also constructed through a manual review of the
occupational variables for each job. Cases and controls were classified as
having an annual exposure ,1 or >1 mSv (milliSievert).
Results
A total of 543 brain cancer cases (ICD-9: 191) formed the basis of this
analysis (Table 1). Of these cases, 214 were astrocytomas, 198 were
glioblastoma multiforme, 115 were classified into the ‘other category’.
Sixteen cases could not be categorized because they were lacking
histological data.
The frequency distribution of several key variables is presented for both
cases and controls in Table 2. The matched design of the study ensured
identical age distributions in the case and control series; 64% of the cases
were >45 years of age. The average number of jobs held by each subject
and the length of employment were similar between cases and controls.
Likewise, the total number of subjects with reported workplace exposures
to pesticides, herbicides, radiation sources and vinyl chloride did not
differ appreciably by case-control status. Eighty-six per cent (86%) of
jobs held by cases were determined to have an average magnetic field
exposure of ,0.3 µT. Among controls, the corresponding percentage was
88%. Based on our three-level exposure categorization, 845 subjects
(78%) did not experience a change in the average level of exposure to
magnetic fields based on their lifetime occupational history. Of those that
did
Table 2 Characteristics of study subjects, by case-control status
Variable
Cases
Controls
88
94
35–44
108
112
45–54
129
131
55–64
122
120
65–74
93
86
3
3
Age at interview (years)
,35
75
BRAIN CANCER AND OCCUPATIONAL EXPOSURE TO MAGNETIC FIELDS 213
Average number of jobs held
3.6 (SD = 2.2)
3.5 (SD = 2.1)
Average length of time spent in each
job (in years)
8.1 (SD = 9.3)
8.3 (SD = 9.5)
A statistically not significant increased risk of brain cancer was
observed among those subjects who had ever held a job having
average magnetic field exposures .0.6 µT relative to those whose
highest level was ,0.3 µT (OR = 1.33, 95% CI : 0.75–2.36) (Table 3).
When analyses were restricted to those cases diagnosed with a
glioblastoma multiforme, the resulting risk estimate was considerably
higher (OR = 5.36, 95% CI : 1.16– 24.78). No significant differences
in risk were observed based on the highest level of occupational
magnetic field exposure ever received were for those diagnosed with
astrocytomas or other brain cancers. Similar results were obtained
when risk assessment was performed using the average occupational
magnetic field exposure received in the longest held job (results not
shown). Specifically, among those subjects diagnosed with
glioblastoma multiforme, the risk estimates were most pronounced
among subjects whose longest held job had an average exposure that
exceeded 0.6 µT when compared to those with exposures ,0.3 µT (OR
= 3.70, 95% CI : 0.96–1.20); the corresponding OR for all brain
cancers combined was 1.27 (95% CI : 0.64–2.53).
The results obtained from modelling the relationship between the
incidence of brain cancer and the constructed index that represents a
cumulative lifetime occupational magnetic field exposure score is
presented in Table 4. Consistent with our previous findings, this
continuous index of magnetic field exposure was not significantly
related to the incidence of all brain cancers, astrocytomas nor other
brain cancers. However, for those diagnosed with glioblastoma
multiforme, this exposure index was significantly related to case-
Subjects who worked with the following
for more than one year
Pesticides
77
80
Herbicides
65
65
Radiation sources
32
36
7
9
1690
1654
164
162
>0.6 µT
71
53
Exposure could not be assigned
32
10
543
543
Vinyl chloride
Total no. of jobs held according to average
exposure to magnetic fields
,0.3 µT
0.3–,0.6 µT
Total subjects
experience such a change, 12% of the subjects experienced one change
while the remaining 10% experienced two or more changes during their
occupational history. A greater number of occupations among cases (n =
32) relative to controls (n = 10) could not be categorized according to the
average level of magnetic field exposure. We were unable to classify
these occupations because subjects did not provide data that described
either their job-title or duties.
control status as indicated by the Wald χ2 statistic (P = 0.02), and upon
categorization, those subjects that had an index score >8 had an OR of
2.58 (95% CI : 1.15–5.82) relative to those with a score of zero (results
not shown).
Table 3 The risk of brain cancer according to the highest average level of occupational magnetic field exposure ever received, by histological type, Canadian National
Enhanced Cancer Surveillance System (NECSS), male participants, 1994–1997
Highest average occupational exposure magnetic fields
ever received
Cases
Controls
Odds ratioa
95% CI
Odds ratiob
95% CI
All brain cancers
,0.3 µTc
410
420
1.0
>0.3 µT
133
123
1.11
0.84–1.48
1.12
0.83–1.51
>0.6 µT
42
29
1.38
0.79–2.42
1.33
0.75–2.36
,0.3 µT
163
160
1.0
>0.3 µT
51
54
0.93
0.60–1.44
0.93
0.59–1.47
>0.6 µT
12
16
0.61
0.26–1.49
0.59
0.24–1.45
,0.3 µT
143
156
1.0
>0.3 µT
55
42
1.50
0.91–2.46
1.48
0.89–2.47
>0.6 µT
18
6
5.50
1.22–24.8
5.36
1.16–24.78
,0.3 µT
92
94
1.0
>0.3 µT
23
21
1.11
1.0
Astrocytomas
1.0
Glioblastoma multiforme
1.0
Other
1.0
0.59–2.10
1.10
0.58–2.09
214 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
>0.6 µT
9
7
1.50
0.53–4.21
1.58
0.56–4.50
a Unadjusted odds ratio obtained from the conditional logistic model. b The odds ratio was adjusted for
occupational exposure to ionizing radiation and vinyl chloride. c Referent group.
Table 4 Parameter estimates obtained by modelling the relationship between brain cancer and a cumulative index of occupational magnetic field exposure using
conditional logistic regression, by histological type, Canadian National Enhanced Cancer Surveillance System (NECSS), male participants, 1994–1997
Parameter estimatea for cumulative
Histological type of cancer
index of magnetic field exposure
All brain cancers
Astrocytomas
Glioblastoma multiforme
Other
Odds ratiob
Standard error
P-valuec
0.0173
0.0107
1.02
0.10
–0.0096
0.0192
0.98
0.62
0.0415
0.0177
1.04
0.02
–0.0169
0.0335
0.98
0.61
a The parameter estimate was adjusted for exposure to ionizing radiation and vinyl chloride. b This odds ratio represents the change in risk
of cancer per unit increase in the cumulative index of magnetic field exposure. c The P-value was calculated using the Wald χ2 test statistic.
The risk of brain cancer based on the average field exposure of the
subject’s first or last held job is presented in Table 5. Among subjects
diagnosed with glioblastoma, the OR among those with averages
exposures .0.6 µT relative to those with exposures ,0.3 µT were 4.81
(95% CI : 0.94–24.71) and 12.59 (95% CI : 1.50–150) for the first
and last held job respectively. However, differences between these
two risk estimates should be interpreted cautiously as only one control
had an average exposure >0.6 µT in the last held job.
Discussion
We found that as a whole, brain cancer was not significantly related
to occupational exposure to magnetic fields. However, when the
analyses were restricted by histological type, four indices of
occupational magnetic fields (highest exposure received, exposure
during first job, exposure during last job, exposure during longest
held job, and cumulative exposure) were positively associated with
glioblastoma multiforme. In contrast, no significant associations were
observed with astrocytomas or other brain cancers. The large
variation in risk between astrocytomas and glioblastoma multiforme
Table 5 The risk of brain cancer according to the occupational magnetic field exposure received in the first and last held job, by histological type, Canadian
National Enhanced Cancer Surveillance System (NECSS), male participants, 1994–1997
Average occupational exposure
to magnetic fields
Earliest held job
Last held job
a
Cases
Controls
Odds ratio
458
474
1.0
>0.3–,0.6 µT
43
48
0.89
>0.6 µT
21
12
1.72
180
187
1,0
19
20
0.81
7
4
1.51
163
174
1.0
>0.3–,0.6 µT
15
16
1.21
>0.6 µT
10
3
4.81
101
100
1.0
>0.3–,0.6 µT
8
7
1.11
>0.6 µT
3
5
0.65
Cases
Controls
Odds ratiob
475
490
1.0
0.57–1.37
46
41
1.13
0.72–1.79
0.80–3.66
16
11
1.50
0.69–3.28
186
187
1.0
0.43–1.53
21
20
0.98
0.50–1.92
0.45–5.38
5
7
0.71
1.22–2.27
171
184
1.0
0.55–2.66
17
12
1.99
0.83–4.81
0.94–24.71
8
1
12.59
1.50–105.6
105
105
1.0
0.37–3.33
6
7
0.83
0.25–2.70
0.15–2.77
2
3
0.62
0.10–3.76
95% CI
95% CI
All brain cancers
,0.3 µTc
Astrocytomas
,0.3 µT
>0.3–,0.6 µT
>0.6 µT
Glioblastoma multiforme
,0.3 µT
Other
,0.3 µT
a
Unadjusted odds ratio obtained from the conditional logistic model.
b
The odds ratio was adjusted for occupational exposure to ionizing radiation and vinyl chloride. c Referent group.
BRAIN CANCER AND OCCUPATIONAL EXPOSURE TO MAGNETIC FIELDS 215
requires comment. These two cancers account for approximately 80%
of all gliomas.26 It is generally accepted that astrocytic gliomas that
are classified as grades 1 and 2 are classified as astrocytomas and the
more aggressive forms (grades 3 and 4) are classified as
glioblastomas.26 Indeed, cases of glioblastoma multiforme often
evolve from less malignant forms of astrocytoma, although some
26
cases rise de novo. The results from in vivo and in vitro work
suggest that if magnetic fields influence carcinogenesis, it is through
a promoting effect.27,34,38 For example, 60 Hz magnetic field
exposures were recently shown to increase the rate of proliferation in
astrocytoma cells and potentiate the effect of the phrobol ester
PMA.39 Continued exposure to promoting or co-promoting agents
after tumour development may cause the tumour to evolve with
increased invasive and metastatic properties.28 Although the
underlying mechanisms of carcinogenesis continue to be widely
debated, the increased risk due to exposure to magnetic fields that was
found for more aggressive malignancies (i.e. glioblastoma
multiforme) is consistent with the hypothesis that magnetic fields act
at the promotional stage.
It is possible that our results may be biased due to nonresponse in the
case and control series. Since questionnaires were not mailed out to cases
known to be deceased, our analyses does not include aggressive forms of
brain cancer that were rapidly fatal. To the extent that physician consent
was not given due to the poor health of the cancer patient, additional cases
of advanced disease will also be excluded. In total, almost onethird of
eligible cases were excluded either because the subject had died, or
consent was not given by the physician to approach patients diagnosed
with brain cancer. Of the remaining cases, 63% participated in the study.
Therefore, if magnetic field exposures act as a promoter of brain cancer,
our risk estimates would be attenuated because the risk profiles of less
aggressive brain cancer cases may be more similar to the profiles in the
controls.
A large Tri-Utility study that employed personal monitoring to
construct a job-exposure matrix of magnetic field exposures found an
elevated risk of brain cancer among those with high cumulative
exposures.9 The Tri-Utility study has a considerable number of strengths
including a relatively large sample (n = 250), and workplace exposures
that were inferred using personal monitoring worn by a sample of current
workers. The investigators found that those workers having a cumulative
exposure to magnetic fields that exceeded the median exposure (3.15 µTyears) had a twofold increase in brain cancer risk (OR = 2.0, 95% CI :
0.98–3.9). Contrary to our findings, the increased brain cancer risk in the
Tri-Utility study was observed among those cases diagnosed with an
astrocytoma. Their findings should be interpreted with caution as there
were only five cases in the exposed population, and there were
differences in the follow-up procedures of workers from the Ontario,
Quebec and French utilities. Furthermore, electric utility workers
represent a select subset of individuals that are likely to exhibit less
variation with respect to magnetic fields exposures, demographic
characteristics and other occupational exposure than encountered in our
populationbased sample of individuals.
Many occupations with greater than background levels of exposure to
magnetic fields are also associated with higher exposure to electric fields.
A re-analysis of the French component of the Tri-Utility study observed
a positive relationship between occupational exposure to electric fields
and the incidence of brain cancer and benign tumours. 40 In particular,
subjects having exposures in the 90th percentile had an OR of 3.1 (95%
CI : 1.1–8.7) relative to the baseline group. On the other hand, a reanalysis of the Ontario data found no association between cumulative
electric field exposure and the incidence of brain tumours. 8 Occupational
data for electric field exposures were not assembled for the subjects that
we analysed, and therefore, our risk estimates were unable to be adjusted
for the potential confounding influence of these exposures.
Our results are consistent with findings from a Swedish casecontrol
study of occupational and residential exposure to magnetic fields11 that
observed a significant relationship between magnetic field exposure and
the incidence of astrocytomas grades III and IV (or glioblastoma
multiforme). A nonsignificantly increased risk of astrocytoma grades III
and IV was observed among those having both residential and
occupational exposure .0.2 µT (OR = 2.2, 95% CI : 0.6–8.5). The
precision of this estimate was limited by the fact that only three cases had
high exposures to both residential and occupational magnetic fields.
Unlike the Swedish study which only took into account one occupation
held by the subject (based on census data), our analyses considered all
occupations held. Although we were unable to model residential
magnetic field exposures, in general, the weak correlation between home
and workplace exposures41 reduces the likelihood that our results will be
confounded.
More recently, it has been suggested that the failure to consider
magnetic field frequencies ,20 Hz that emanate from radial tyres may
compromise risk estimates obtained from epidemiological studies. 42
If exposures ,20 Hz are relevant to the biological mechanisms
associated with the development of brain tumours then our risk
estimates may be understated due to increased exposure
misclassification.
We also evaluated the relationship between the total number of
years spent in occupations with exposures of (1) 0.3–0.6 µT and (2)
.0.6 µT. However, the precision of the parameter estimates that were
derived for these two continuous measures of exposures was limited
by the small number of subjects that had such exposure. For example,
only 4.1% of subjects had average occupational magnetic fields that
were >0.6 µT, while 16.1% had exposures that were 0.3–,0.6 µT. For
this reason, we have presented results based on the cumulative
measure of magnetic field exposure that combines information across
the three possible job exposure categories. Comparative analyses of
the risk of glioblastoma multiforme between the first and last held
jobs revealed a more pronounced risk for those jobs held more
recently. However, caution should be exercised when interpreting this
finding due to the small number of subjects with exposure >0.6 µT
and the width of the accompanying confidence intervals.
The results of this study support the hypothesis that occupational
magnetic field exposures play a role in the aetiology of brain cancers.
Despite a sample size that is considerably larger than most studies of
brain cancer and magnetic field exposure, these findings must still be
interpreted cautiously due to a smaller number of cases within each
histological grouping and the unavailability of direct sampled
measures of field exposure. Nonetheless, the elevated risk of
glioblastoma multiforme is of significance and replication of this
study result should be pursued in another population.
Acknowledgements
The authors are grateful to Long On of the Environmental Risk
Assessment and Case Surveillance Division of Health Canada for
preparing the NECSS analysis files and to helpful comments provided
by Erich Kliewer on an early draft of this manuscript. We also thank
the reviewers for their thoughtful comments on the original
submission of this manuscript.
216 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
KEY MESSAGES
•
This study examined the relationship between occupational exposure to magnetic fields and the incidence of brain cancer in
a population-based sample of Canadians.
•
A positive and statistically significant relationship was found between average levels of occupational magnetic field exposure
and the incidence of glioblastoma multiforme, which is a more aggressive subtype of brain cancer.
•
These results support the hypothesis than magnetic fields play a role in the development of brain tumours, and that they may
exert an influence as tumour promoters.
14
References
Tynes T, Reitan JB, Andersen A. Incidence of cancer among workers in
Norwegian hydroelectric power companies. Scand J Work Environ Health
1994;20:339–44.
1
Preston-Martin S. Epidemiology of primary CNS neoplasms. Neurol Clin
1996;14:273–90.
15
Loomis DP, Savitz DA. Mortality from brain cancer and leukaemia among
electrical workers. Br J Ind Med 1990;47:633–38.
2
IARC monographs on the evaluation of the carcinogenic risk of chemicals
to humans. Overall evaluations of carcinogenicity: an updating of IARC
monographs volumes 1 to 42. Supplement 7. Lyon, France: International
Agency for Research on Cancer, 1987.
16
Lin RS, Dischinger PC, Conde J, Farrell KP. Occupational exposure to
electromagnetic fields and the occurrence of brain tumors. An analysis of
possible associations. J Occup Med 1985;27:413–19.
3
Moulin JJ, Portefaix P, Wild P, Mur JM, Smagghe G, Mantout B. Mortality
study among workers producing ferroalloys and stainless steel in France. Br
J Ind Med 1990;47:537–43.
17
Rodvall Y, Ahlbom A, Stenlund C, Preston-Martin S, Lindh T, Spannare B.
Occupational exposure to magnetic fields and brain tumours in central Sweden.
Eur J Epidemiol 1998;14:563–69.
4
Waxweiler RJ, Stringer W, Wagoner JK, Jones J, Falk H, Carter C.
Neoplastic risk among workers exposed to vinyl chloride. Ann NY Acad Sci
1976;271:40–48.
18
Speers MA, Dobbins JG, Miller VS. Occupational exposures and brain cancer
mortality: a preliminary study of east Texas residents. Am J Ind Med
1988;13:629–38.
5
Bohnen NI, Kurland LT. Brain tumor and exposure to pesticides in
humans: a review of the epidemiologic data. J Neurol Sci 1995;132: 110–
21.
19
Floderus B, Persson T, Stenlund C, Wennberg A, Ost A, Knave B. Occupational
exposure to electromagnetic fields in relation to leukemia and brain tumors: a
case-control study in Sweden. Cancer Causes Control 1993;4:465–76.
6
Miller RD, Neuberger JS, Gerald KB. Brain cancer and leukemia and
exposure to power-frequency (50- to 60-Hz) electric and magnetic fields.
Epidemiol Rev 1997;19:273–93.
20
Guenel P, Raskmark P, Andersen JB, Lynge E. Incidence of cancer in persons
with occupational exposure to electromagnetic fields in Denmark. Br J Ind Med
1993;50:758–64.
7
National Cancer Institute of Canada. Canadian Cancer Statistics 1995.
Toronto, 1995.
21
Preston-Martin S, Mack W, Henderson BE. Risk factors for gliomas and
meningiomas in males in Los Angeles County. Cancer Res 1989;49: 6137–43.
8
Miller AB, To T, Agnew DA, Wall C, Green LM. Leukemia following
occupational exposure to 60-Hz electric and magnetic fields among Ontario
electric utility workers [see comments]. Am J Epidemiol 1996; 144:150–60.
9
Thériault G, Goldberg M, Miller AB et al. Cancer risks associated with
occupational exposure to magnetic fields among electric utility workers in
Ontario and Quebec, Canada, and France: 1970–1989 [published erratum
appears in Am J Epidemiol 1994;139:1053] [see comments]. Am J
Epidemiol 1994;139:550–72.
10
11
12
13
Harrington JM, McBride DI, Sorahan T, Paddle GM, van Tongeren M.
Occupational exposure to magnetic fields in relation to mortality from brain
cancer among electricity generation and transmission workers. Occup
Environ Med 1997;54:7–13.
Feychting M, Forssen U, Floderus B. Occupational and residential magnetic
field exposure and leukemia and central nervous system tumors.
Epidemiology 1997;8:384–89.
Sahl JD, Kelsh MA, Greenland S. Cohort and nested case-control studies of
hematopoietic cancers and brain cancer among electric utility workers.
Epidemiology 1993;4:104–14.
22
Johansen C, Olsen JH. Risk of cancer among Danish utility workers— a
nationwide cohort study. Am J Epidemiol 1998;147:548–55.
23
Tynes T, Andersen A, Langmark F. Incidence of cancer in Norwegian workers
potentially exposed to electromagnetic fields. Am J Epidemiol 1992;136:81–88.
24
Savitz DA, Cai J, van Wijngaarden E et al. Case-cohort analysis of brain cancer
and leukemia in electric utility workers using a refined magnetic field jobexposure matrix [In Process Citation]. Am J Ind Med 2000;38:417–25.
25
Baris D, Armstrong BG, Deadman J, Theriault G. A mortality study of electrical
utility workers in Quebec. Occup Environ Med 1996;53:25–31.
26
Inskip PD, Linet MS, Heineman EF. Etiology of brain tumors in adults.
Epidemiol Rev 1995;17:382–414.
27
Cridland N. Effects of power frequency EMF exposures at the cellular level.
Radiation Protection Dosimetry 1997;72:279–90.
28
Loscher W, Mevissen M. Animal studies on the role of 50/60-Hertz magnetic
fields in carcinogenesis. Life Sci 1994;54:1531–43.
29
Savitz DA, Loomis DP. Magnetic field exposure in relation to leukemia and
brain cancer mortality among electric utility workers [published erratum
appears in Am J Epidemiol 1996;144:205]. Am J Epidemiol 1995;141:123–
34.
Pitot HC, Dragan YP. Facts and theories concerning the mechanisms of
carcinogenesis. FASEB J 1991;5:2280–86.
30
World Health Organization. International Classification of Diseases, 9th
Revision. Geneva: WHO, 1985.
BRAIN CANCER AND OCCUPATIONAL EXPOSURE TO MAGNETIC FIELDS 217
31
World Health Organization. International Classification of Diseases for
Oncology. Geneva: WHO, 1976.
32
Green LM, Miller AB, Agnew DA et al. Childhood leukemia and personal
monitoring of residential exposures to electric and magnetic fields in Ontario,
Canada. Cancer Causes Control 1999;10:233–43.
33
Ontario Hydro. Summary of Electric and Magnetic Field Measurements to June
16th. Toronto: Ontario Hydro, 1989.
34
NIEHS. Assessment of Health Effects from Exposure to Power-line Frequency
Electric and Magnetic Fields. Research Triangle Park, NC, US: National
Institute of Health, 1998.
35
Floderus B, Persson T, Stenlund C. Magnetic-field exposures in the workplace:
reference distribution and exposures in occupational groups. Int J Occup
Environ Health 1996;2:226–38.
36
London SJ, Bowman JD, Sobel E et al. Exposure to magnetic fields among
electrical workers in relation to leukemia risk in Los Angeles County. Am J Ind
Med 1994;26:47–60.
37
Bowman JD, Garabrant DH, Sobel E, Peters JM. Exposures to extremely low
frequency (ELF) electromagnetic fields in occupations with elevated leukemia
rates. Appl Ind Hyg 1988;3:189–94.
38
Hester, GL. Electric and magnetic fields: managing an uncertain environment.
Environment 1992;34:7–31.
39
Wei M, Guizzetti M, Yost M, Costa LG. Exposure to 60-Hz magnetic fields and
proliferation of human astrocytoma cells in vitro. Toxicol Appl Pharmacol
2000;162:166–76.
40
Guenel P, Nicolau J, Imbernon E, Chevalier A, Goldberg M. Exposure to 50Hz electric field and incidence of leukemia, brain tumors, and other cancers
among French electric utility workers [see comments]. Am J Epidemiol
1996:144:1107–21.
41
Knave, B. Electric and magnetic fields and health outcomes—an overview.
Scand J Work Environ Health 1994;20:78–89.
42
Milham S, Hatfield JB, Tell R. Magnetic fields from steel-belted radial tires:
implications for epidemiologic studies. Bioelectromagnetics 1999; 20:440–
45.
Copyright of International Journal of Epidemiology is the property of Oxford University Press / UK and its
content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for individual use.
Purchase answer to see full
attachment