Professional Documents
Culture Documents
Environmental Research
journal homepage: www.elsevier.com/locate/envres
A R T I C LE I N FO A B S T R A C T
Keywords: Exposure to extremely low-frequency electromagnetic fields (ELF-EMF) was in 2002 classified as a possible
Electromagnetic fields human carcinogen, Group 2B, by the International Agency for Research on Cancer at WHO based on an increased
ELF-EMF risk for childhood leukemia. In case-control studies on brain and head tumours during 1997–2003 and
Acoustic neuroma 2007–2009 we assessed life-time occupations in addition to exposure to different agents. The INTEROCC ELF-
Occupational exposure
EMF Job-Exposure Matrix was used for associating occupations with ELF-EMF exposure (μT) with acoustic
Job-exposure matrix
neuroma. Cumulative exposure (μT-years), average exposure (μT) and maximum exposed job (μT) were calcu-
lated. No increased risk for acoustic neuroma was found in any category. For cumulative exposure in the highest
exposure category 8.52+ μT years odds ratio (OR) = 1.2, 95% confidence interval (CI) = 0.8–2.0, p linear
trend = 0.37 was calculated. No statistically significant risks were found in the time windows 1–14 years, and
15+ years, respectively. In conclusion occupational ELF-EMF was not associated with an increased risk for
acoustic neuroma.
∗
Corresponding author.
E-mail addresses: michael.carlberg@environmentandcancer.com (M. Carlberg), tarmo.koppel@ttu.ee (T. Koppel), mikko.ahonen@tutanota.com (M. Ahonen),
lennart.hardell@environmentandcancer.com (L. Hardell).
1
Department of Oncology, Faculty of Medicine and Health, Fakultetsgatan 1, Örebro University, SE-701 82, Örebro, Sweden
https://doi.org/10.1016/j.envres.2020.109621
Received 14 January 2020; Received in revised form 8 April 2020; Accepted 29 April 2020
Available online 07 May 2020
0013-9351/ © 2020 Elsevier Inc. All rights reserved.
M. Carlberg, et al. Environmental Research 187 (2020) 109621
neuroma (Forséen et al., 2006). No increased risk for acoustic neuroma and the working areas with respect to the workers’ position.
was found for occupational and residential exposure to EMFs, based on Our case-control studies had detailed occupational history including
only 32 cases, however (Baldi et al., 2011). Extremely low-frequency job titles, branch of different occupations and years for the specific jobs.
(ELF)-EMF was in 2002 classified by the International Agency for Re- Thus, it was possible to calculate ELFEMF job exposure for cases and
search on Cancer (IARC) as “possibly carcinogenic to humans”, Group controls using a job-exposure matrix (JEM).
2B based on an increased risk for childhood leukemia (IARC, 2002).
There was no consistent pattern of an association between brain tu- 2. Materials and methods
mours and residential exposure to ELF-EMF. Data for occupational
cancer was limited. In summary the IARC panel concluded that the 2.1. Data material
evidence for an increased risk associated with ELF-EMF exposure for
other diseases than childhood leukemia was inadequate. Similar methods were used in all our studies; detailed information
During the last couple of decades an increasing number of studies has been published previously (Hardell et al., 2006, 2013a). In brief, 6
have associated brain tumours with use of wireless phones (Hardell administrative regions with oncology centres covering Sweden regis-
et al., 2006, 2013a; Interphone Study Group, 2010; Coureau et al., tered new cancer cases. Regarding 1997–2003, cases and controls
2014; Hardell and Carlberg, 2015). During use they emit radio- covered central Sweden (Hardell et al., 2006), whereas the 2007–2009
frequency electromagnetic fields (RF-EMF). In May 2011 IARC eval- study included the whole country (Hardell et al., 2013a). The oncology
uated the carcinogenic potential from RF-EMF. The expert group clas- centres reported new cases with histopathologically verified brain tu-
sified RF-EMF in the frequency range 30 kHz–300 GHz as “possibly mour, either benign or malignant, to us during these periods, although
carcinogenic to humans”, Group 2B (Baan et al., 2011; IARC, 2013). the actual reporting interval varied from centre to centre. In total, 4038
Regarding use of mobile phones and the risk for acoustic neuroma a brain tumour cases (2437 cases 1997–2003 and 1601 cases 2007–2009)
meta-analysis of Interphone (Interphone Study Group, 2011) and were included. Of these, 3563 (2158 cases 1997–2003 and 1405 cases
Hardell et al., (2013b) yielded statistically significant increased risk for 2007–2009) answered the questionnaire. Among the responding cases,
cumulative ipsilateral use ≥1640 h; odds ratio (OR) = 2.71 95% 1498 were diagnosed with a malignant brain tumour and 2068 with a
confidence interval (CI) = 1.72–4.28 (Belpomme et al., 2018). Gender- benign brain tumour (3 were diagnosed with both malignant and be-
specific analyses in our study yielded similar results (Hardell et al., nign tumour). Regarding acoustic neuroma, 338 were included from the
2013b). Cumulative use of wireless phones gave OR = 2.9, 95% brain tumour cases (257 from 1997 to 2003 and 81 from 2007 to 2009)
CI = 1.5–5.6 for men in the fourth quartile and OR = 1.9, 95% and of these, 316 answered the questionnaire (243 from 1997 to 2003
CI = 1.1–3.4 for women; thus the results for both genders were sta- and 73 from 2007 to 2009).
tistically significant with 95% CI overlapping ORs (data not shown in Men and women, aged 20–80 years (1997–2003) and 18–75 years
article). (2007–2009) at the time of diagnosis, were included. Only living cases
Similarly, as in Interphone, our results on use of wireless phones and were included after asking the responsible physician for permission
brain tumour risk were based on case-control studies. We used a before inclusion in the study. Tumour localization in the brain was
structured questionnaire but with certain differences compared to the based on reports to the cancer registries and medical records, which
Interphone study, such as that we used postal questionnaires sent to were obtained after informed consent from the patients.
cases and controls, supplemented over the phone instead of personal Controls were ascertained from the Swedish Population Registry.
interviews. In contrast to our studies even bedside interviews of cases The registry is continuously updated, so that each person could be
were performed in Interphone. Furthermore, we assessed in addition to traced by a unique ID number. It also records the address of each
mobile phones also use of cordless phones (DECT); the latter use not person. For each case, one control subject of the same gender and in the
assessed by Interphone. Detailed comparison of the studies may be same 5-year group was drawn at random from the Population Registry.
found elsewhere (Hardell et al., 2008). They were assigned the same year for cut-off of all exposure as the year
Wherever alternating electric current is used, this generates ELF- of diagnosis of the respective case. In total, 4038 controls (2437 con-
EMFs. Most often exposure to ELF-EMFs is from machines operating on trols 1997–2003 and 1601 controls 2007–2009) were included; of
the mains power, which operates at 50 Hz frequency. 50 Hz mains these, 3530 controls (2162 controls 1997–2003 and 1368 controls
power is used in most of the world, except for Americas and some parts 2007–2009) answered the questionnaire. All these controls were used
of Asia where 60 Hz used. Next to the frequency, the main exposure in the analysis of risk of acoustic neuroma.
feature is the amplitude of the wave. Magnetic fields depend on the Exposure was assessed using a mailed questionnaire sent to each
current – the more electrical power consumed the higher the magnetic person. The questionnaire contained a number of questions relating to
field. Therefore, powerful electrical machinery provides high ELF-EMF the overall working history, exposure to different chemicals and other
exposure to workers working nearby these machines. High exposure to agents, smoking habits, X-ray investigations of the head and neck, and
the ELF magnetic field may also take place next to generation and heredity traits for cancer. Regarding use of a mobile phone and cordless
distribution of electrical power, such as power lines, transformers etc. phone, time period, average daily use (min per day), use of hands-free
Electrical motors, which are used in many industrial processes, are device, and external antenna in a car were asked for. The ear mostly
another source of high ELF magnetic field exposure. used during phone calls, or equally both, was also noted.
Electrical motors and electromagnets, which incorporate coils pro- When questionnaire answers were unclear, they were resolved by
duce often higher magnetic fields than the cables powering these. The phone by trained interviewers using a written protocol for clarification
number of turns of electrically conductive element i.e. wire by which of each question. The interviewer supplemented the whole ques-
the electrical current flows, also determines the resultant magnetic field tionnaire during the phone call. Each questionnaire had received a
– the more there are turns, the higher the magnetic field. Therefore, unique ID-number that did not disclose whether it was a case or a
where workers use or are next to powerful electrical motors they are control; i.e. the interviewer was unaware of the status during further
exposed to strong ELF-EMFs. Where strong magnetic fields occur, like in data processing. All information was coded and entered into a database;
electric transport, industrial settings etc. the workers' exposure level case or control status was not disclosed until statistical analyses were
depends on the distance to the source. People working with hand-held undertaken.
electrical power tools have strong magnetic fields exposure, whereas In this study we included acoustic neuroma cases. All controls were
others further away from even more strong sources, may be exposed at used as comparison group. This was possible since we adjusted for
much less levels. Therefore, a crucial role in determining workers' ex- potential confounding factors such as year of diagnosis, age at diag-
posure to the magnetic field is from the design of the instrumentation nosis, gender and socioeconomic index (SEI). SEI was divided into three
2
M. Carlberg, et al. Environmental Research 187 (2020) 109621
tions 1988 (ISCO88) four-digit codes for most jobs included; ISCO68 Mean age (median; min-max) 52 (53; 24–80) 54 (56; 20–80)
five-digit codes were used for more specific electrical jobs. The online Gender
version of the JEM is available at: Men 141 45.5 1472 42.2
http://radiation.isglobal.org/index.php/ca/radiation-programme- Women 169 54.5 2013 57.8
Time period
projects/2-general/55-database-elfjem.
1997–2003 237 76.5 2132 61.2
Job codes were coded using the Nordisk Yrkesklassificering (NYK 2007–2009 73 23.5 1353 38.8
85; five-digit codes) system and their validity was checked before they SEI-code
were translated to the International Standard Classification of Blue-collar worker 142 45.8 1568 45.0
White-collar worker 148 47.7 1695 48.6
Occupations 1988 (ISCO88; four-digit codes) using a coding key pro-
Self-employed 20 6.5 222 6.4
vided by Dr Erik Bihagen at Stockholm University (Bihagen, 2007). For
translation to the 1968 ISCO version for specific jobs (ISCO68; five-digit
codes) we compared with the NYK 85 system manually and selected the Table 2
most proper codes to be translated. Job exposure the year before di- Odds ratio (OR) and 95% confidence interval (CI) for acoustic neuroma cases
agnosis was excluded. No job was reported for 6 cases and 45 controls, (n = 310) for occupational exposure to ELF-EMF. Population based controls
so they were excluded from the analyses. (n = 3485) were used. Subjects with no coded occupation were excluded, 6
acoustic neuroma cases, and 45 controls. Unconditional logistic regression
2.2. Statistical methods analysis was adjusted for age at diagnosis, gender, socio-economic index (SEI)
and year of diagnosis. Exposure the year before diagnosis was excluded (“1-year
StataSE 12.1 (Stata/SE 12.1 for Windows; StataCorp., College lag”).
Station TX) was used for the analyses. Odds ratios (OR) and 95% con- Exposure metric Acoustic neuroma (n = 310)
fidence intervals (CI) were calculated using unconditional logistic re-
Ca/Co OR 95% CI
gression including acoustic neuroma cases and the whole control
sample (i.e. matched to both malignant and benign cases) to increase Cumulative exposure (μT-years)
the power.
Cumulative exposure (μT-years), average exposure (μT) and max- < 2.33 90/870 1.0 –
2.33- < 3.79 81/872 1.0 0.7–1.4
imum exposed job (μT) were calculated for the cases and controls for
3.79- < 5.55 59/869 0.8 0.5–1.2
lifetime work history and in time windows. Two sets of time windows 5.55- < 8.52 42/525 0.9 0.6–1.4
were analyzed, 1–14 and 15+ years before diagnosis. Cut points at the 8.52+ 38/349 1.2 0.8–2.0
25th, 50th, 75th and 90th percentile for controls were used to cate- p, linear trend 0.37
gorize the exposure variables with the lowest category (< 25th per- Average exposure (μT)
< 0.11 80/830 1.0 –
centile) as reference group (OR = 1.0). Tests for linear trends were
0.11- < 0.13 73/912 0.8 0.6–1.2
performed using the Wald χ2 test with the median of each category 0.13- < 0.18 64/871 0.7 0.5–1.04
included as an ordinal variable in the analyses. 0.18- < 0.27 57/523 1.0 0.7–1.5
In all analyses adjustment was made for the matching variables 0.27+ 36/349 1.0 0.6–1.5
p, linear trend 0.64
gender, age (as a continuous variable) and year of diagnosis and also for
Maximum exposed job (μT)
SEI divided into three categories (blue-collar worker, white-collar < 0.13 82/823 1.0 –
worker, self-employed). 0.13- < 0.16 70/812 0.8 0.6–1.2
Restricted cubic splines were used to illustrate the relationship be- 0.16- < 0.24 73/968 0.7 0.5–0.96
tween cumulative exposure to ELF-EMF (μT-years) in time windows and 0.24- < 0.60 45/532 0.7 0.5–1.1
0.60+ 40/350 1.0 0.6–1.5
acoustic neuroma. As suggested by Harrell Jr (Harrell, 2001), four knots
p, linear trend 0.54
were used at the 5th, 35th, 65th, and 95th percentiles.
Cut points at 25th, 50th, 75th and 90th percentile for controls.
3. Results
we found no statistically significant linear trend for increasing ex-
The results were based on analyses of 310 acoustic neuroma cases posure. Cumulative exposure in the highest exposure category, 8.52+
(92% of all included; 6.5% did not answer the questionnaire and 1.7% μT-years, gave OR = 1.2, 95% CI = 0.8–2.0, p, linear trend = 0.37.
were excluded since they had no job codes registered) and 3485 con- Cumulative exposure in different time windows before diagnosis is
trols (86% of included; 12.6% did not answer the questionnaire and shown in Table 3 and Figs. 1 and 2. No statistically significant risks or
1.1% were excluded since they had no job codes registered). linear trends were found. Cumulative exposure in the highest exposure
The mean age of the acoustic neuroma cases was 52 years (median group 2.75+ μT-years yielded OR = 1.3, 95% CI = 0.8–2.0 (p, linear
53, range 24–80) and of the controls 54 years (median 56, range trend = 0.26) in the latency group 1–14 years, see Table 2. For longer
20–80). Of the cases there were 141 men and 169 women, versus 1472 latency time, 15+ years, OR = 1.1, 95% CI = 0.6–1.8 was calculated
men and 2013 women among the controls. The mean number of jobs for in the highest exposure group 6.59+ μT-years (p, linear trend = 0.74),
cases were 2.9 (median = 3, min = 1, max = 10) and for controls 2.7 see Table 3.
(median = 2, min = 1, max = 12). Main characteristics of the study In a separate analysis we grouped latency in 1–4, 5–9 and 10+
population are displayed in Table 1. years. In the highest exposure category 0.69+ μT-years we calculated
Table 2 displays cumulative exposure in μT-years, average exposure OR = 1.2, 95% CI = 0.7–1.8 for tumour induction period 1–4 years,
in μT, and maximum exposure job (μT). No statistically significant in- exposure 0.92+ μT-years in the latency group 5–9 years yielded
creased or decreased risk was found for any of the studied variables and
3
M. Carlberg, et al. Environmental Research 187 (2020) 109621
Table 3
Odds ratio (OR) and 95% confidence interval (CI) for acoustic neuroma
(n = 310) for occupational exposure to ELF-EMF in time windows; 1–14 and
15+ years before diagnosis. Unconditional logistic regression, adjusted for age
at diagnosis, gender, socioeconomic index (SEI) and year of diagnosis was used.
Exposure the year before diagnosis was excluded (“1-year lag”).
Cumulative exposure (μT-years) Acoustic neuroma (n = 310)
Ca/Co OR 95% CI
1–14 years
4
M. Carlberg, et al. Environmental Research 187 (2020) 109621
EMF exposure and acoustic neuroma. Forséen, U.M., Lönn, S., Ahlbom, A., et al., 2006. Occupational magnetic field exposure
and the risk of acoustic neuroma. Am. J. Ind. Med. 49, 112–118.
Hardell, L., Carlberg, M., 2015. Mobile phone and cordless phone use and the risk for
Funding glioma - analysis of pooled case-control studies in Sweden, 1997-2003 and 2007-
2009. Pathophysiology 22, 1–13.
This study was supported by Kone Foundation, Helsinki, Finland, Hardell, L., Carlberg, M., 2019. Comments on the US National Toxicology Program
technical reports on toxicology and carcinogenesis study in rats exposed to whole-
Cancerhjälpen, Sweden, and Pandora-Foundation for Independent body radiofrequency radiation at 900 MHz and in mice exposed to whole-body
Research, Berlin, Germany. Support by a grant from Mr Brian Stein is radiofrequency radiation at 1,900 MHz. Int. J. Oncol. 54, 111–127.
acknowledged. Funding sources had no role in study design, data col- Hardell, L., Carlberg, M., Söderqvist, F., Hansson Mild, K., 2013a. Case-control study of
the association between malignant brain tumours diagnosed between 2007 and 2009
lection, analysis and interpretation of the data, in the writing of the and mobile and cordless phone use. Int. J. Oncol. 43, 1833–1845.
manuscript, or in the decision to publish the results. Hardell, L., Carlberg, M., Hansson Mild, K., 2006. Pooled analysis of two case-control
studies on use of cellular and cordless telephones and the risk for malignant brain
tumours diagnosed in 1997–2003. Int. Arch. Occup. Environ. Health 79, 630–639.
Ethics approval and consent to participate
Hardell, L., Carlberg, M., Hansson Mild, K., 2008. Methodological aspects of epidemio-
logical studies on the use of mobile phones and their association with brain tumors.
All study subjects participated after informed consent; the studies Open Environ. Sci. 2, 54–61.
were approved by the Ethical Committee (Örebro County Hospital DNR Hardell, L., Carlberg, M., Söderqvist, F., Hansson Mild, K., 2013b. Pooled analysis of case-
control studies on acoustic neuroma diagnosed 1997-2003 and 2007-2009 and use of
351/96, Uppsala University DNR, 2005:367). mobile and cordless phones. Int. J. Oncol. 43, 1036–1044.
Harrell Jr., F.E., 2001. Regression Modeling Strategies. With Application to Linear
CRediT authorship contribution statement Models, Logistic Regression and Survival Analysis. Springer, New York.
IARC, 2002. Monograph - Non-ionizing Radiation, Part 1: Static and Extremely Low-
Frequency (ELF) Electric and Magnetic Fields, vol. 80 World Health Organization,
Michael Carlberg: Conceptualization, Formal analysis, Data cura- International Agency for Research on Cancer, Lyon. http://monographs.iarc.fr/ENG/
tion. Tarmo Koppel: Validation, Writing - review & editing. Mikko Monographs/vol80/mono80.pdf.
IARC, 2013. Monograph - Non-ionizing Radiation, Part 2: Radiofrequency
Ahonen: Validation, Writing - review & editing. Lennart Hardell: Electromagnetic Fields, vol. 102 World Health Organization, International Agency for
Conceptualization, Supervision, Writing - original draft. Research on Cancer, Lyon. http://monographs.iarc.fr/ENG/Monographs/vol102/
mono102.pdf.
Interphone Study Group, 2010. Brain tumour risk in relation to mobile telephone use:
Declaration of competing interest results of the INTERPHONE international case-control study. Int. J. Epidemiol. 39,
675–694.
Interphone Study Group, 2011. Acoustic neuroma risk in relation to mobile telephone use:
The authors declare that they have no known competing financial
results of the INTERPHONE international case-control study. Canc. Epidemiol. 35,
interests or personal relationships that could have appeared to influ- 453–464.
ence the work reported in this paper. Larjavaara, S., Feychting, M., Sankila, R., et al., 2011. Incidence trends of vestibular
schwannomas in Denmark, Finland, Norway and Sweden in 1987-2007. Br. J. Canc.
105, 1069–1075.
References Mairs, R.J., Hughes, K., Fitzsimmons, S., et al., 2007. Microsatellite analysis for de-
termination of the mutagenicity of extremely low-frequency electromagnetic fields
Asthagiri, A.R., Parry, D.M., Butman, J.A., et al., 2009. Neurofibromatosis type 2. Lancet and ionising radiation in vitro. Mutat. Res. 626, 34–41.
373, 1974–1986. National Toxicology Program, 2018. NTP Technical Report on the Toxicology and
Baan, R., Grosse, Y., Lauby-Secretan, B., et al., 2011. Carcinogenicity of radiofrequency Carcinogenesis Studies in Hsd:Sprague Dawley Sd Rats Exposed to Whole-Body Radio
electromagnetic fields. Lancet Oncol. 12, 624–626. Frequency Radiation at a Frequency (900 MHz) and Modulations (GSM and CDMA)
Baldi, I., Coureau, G., Jaffe, A., et al., 2011. Occupational and residential exposure to Used by Cell Phones, vol. 595 NTP TR March 26-28, 2018. https://ntp.niehs.nih.
electromagnetic fields and risk of brain tumors in adults: a case-control study in gov/ntp/about_ntp/trpanel/2018/march/tr595peerdraft.pdf.
Gironde, France. Int. J. Canc. 129, 1477–1484. National Toxicology Program, 2018. NTP Technical Report on the Toxicology and
Belpomme, D., Hardell, L., Belyaev, I., et al., 2018. Thermal and non-thermal health ef- Carcinogenesis Studies in B6C3F1/N Mice Exposed to Whole-Body Radio Frequency
fects of low intensity non-ionizing radiation: an international perspective. Environ. Radiation at a Frequency (1,900 MHz) and Modulations (GSM and CDMA) Used by
Pollut. 242, 643–658. Cell Phones, vol. 596 NTP TR March 26-28, 2018. https://ntp.niehs.nih.gov/ntp/
Bihagen, E., 2007. Nya möjligheter för stratifieringsforskning i Sverige: internationella about_ntp/trpanel/2018/march/tr596peerdraft.pdf.
yrkesklassificeringar och stratifieringsmått över tid [In Swedish]. Sociol. Forsk. 44, Ohgaki, H., 2009. Epidemiology of brain tumors. Methods Mol. Biol. 472, 323–342.
52–67. Qi, G., Zuo, X., Zhou, L., et al., 2015. Effects of extremely low-frequency electromagnetic
Carlberg, M., Koppel, T., Ahonen, M., Hardell, L., 2017. Case-control study on occupa- fields (ELF-EMF) exposure on B6C3F1 mice. Environ. Health Prev. Med. 20, 287–293.
tional exposure to extremely low-frequency electromagnetic fields and glioma risk. Schneider, A.B., Ron, E., Lubin, J., et al., 2008. Acoustic neuromas following childhood
Am. J. Ind. Med. 60, 494–503. radiation treatment for benign conditions of the head and neck. Neuro Oncol. 10,
Carlberg, M., Koppel, T., Ahonen, M., Hardell, L., 2018. Case-control study on occupa- 73–78.
tional exposure to extremely low-frequency electromagnetic fields and the associa- Soffritti, M., Tibaldi, E., Padovani, M., et al., 2016a. Life-span exposure to sinusoidal-50
tion with meningioma. BioMed Res. Int. 5912394 2018. Hz magnetic field and acute low-dose γ radiation induce carcinogenic effects in
Cea-Soriano, L., Wallander, M.A., García Rodríguez, L.A., 2012. Epidemiology of me- Sprague–Dawley rats. Int. J. Radiat. Biol. 92, 202–214.
ningioma in the United Kingdom. Neuroepidemiology 39, 27–34. Soffritti, M., Tibaldi, E., Padovani, M., et al., 2016b. Synergism between sinusoidal-50 Hz
Coureau, G., Bouvier, G., Lebailly, P., et al., 2014. Mobile phone use and brain tumours in magnetic field and formaldehyde in triggering carcinogenic effects in male Sprague-
the CERENAT case-control study. Occup. Environ. Med. 71, 514–522. Dawley rats. Am. J. Ind. Med. 59, 509–521.
Falcioni, L., Bua, L., Tibaldi, E., et al., 2018. Report of final results regarding brain and Turner, M.C., Benke, G., Bowman, J.D., et al., 2014. Occupational exposure to extremely
heart tumors in Sprague-Dawley rats exposed from prenatal life until natural death to low-frequency magnetic fields and brain tumor risks in the INTEROCC study. Canc.
mobile phone radiofrequency field representative of a 1.8 GHz GSM base station Epidemiol. Biomarkers Prev. 23, 1863–1872.
environmental emission. Environ. Res. 165, 496–503.