PU40CH25_Roosli ARjats.

cls December 28, 2018 11:41

Annual Review of Public Health

Brain and Salivary Gland
Tumors and Mobile Phone Use:
Evaluating the Evidence from
Various Epidemiological Study
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

Designs
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

Martin Röösli,1,2 Susanna Lagorio,3

Minouk J. Schoemaker,4 Joachim Schüz,5
and Maria Feychting6
1
Department of Epidemiology and Public Health, Swiss Tropical and Public
Health Institute, 4051 Basel, Switzerland; email: martin.roosli@unibas.ch
2
University of Basel, 4001 Basel, Switzerland
3
Department of Oncology and Molecular Medicine, National Institute of Health,
00162 Rome, Italy
4
Division of Genetics and Epidemiology, The Institute of Cancer Research, London
SM2 5NG, United Kingdom
5
Section of Environment and Radiation, International Agency for Research on Cancer
(IARC), 69372 Lyon, France
6
Institute of Environmental Medicine, Karolinska Institutet, 171 77 Stockholm, Sweden

Annu. Rev. Public Health 2019. 40:25.1–25.18 Keywords

The Annual Review of Public Health is online at
intracranial tumor, central nervous system tumor, glioma, meningioma,
publhealth.annualreviews.org
acoustic neuroma, salivary gland tumor, pituitary gland tumor,
https://doi.org/10.1146/annurev-publhealth-
radiofrequency electromagnetic fields, mobile phones
040218-044037

Copyright © 2019 by Annual Reviews. Abstract

All Rights reserved
Mobile phones (MPs) are the most relevant source of radiofrequency elec-
tromagnetic field (RF-EMF) exposure to the brain and the salivary gland.
Whether this exposure implies a cancer risk has been addressed in sev-
eral case-control and few cohort studies. A meta-analysis of these stud-
ies does not show increased risks for meningioma, pituitary, and salivary
gland tumors. For glioma and acoustic neuroma, the results are heteroge-
neous, with few case-control studies reporting substantially increased risks.

25.1
Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

However, these elevated risks are not coherent with observed incidence time trends, which are
considered informative for this specific topic owing to the steep increase in MP use, the avail-
ability of virtually complete cancer registry data from many countries, and the limited number
of known competing environmental risk factors. In conclusion, epidemiological studies do not
suggest increased brain or salivary gland tumor risk with MP use, although some uncertainty re-
mains regarding long latency periods (>15 years), rare brain tumor subtypes, and MP usage during
childhood.

INTRODUCTION
Exposures from Mobile Phones
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

The first cellular network (1G) introduced in 1979 in Japan and in 1981 in the Nordic countries
used mobile phones (MPs) with antennae mounted on a car or a bag. Handheld MPs with antennae
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

on the handset were introduced in 1983 in the United States and in 1987 in the Nordic countries.
With the deployment of the Global System of Mobile Communication (GSM, 2G) in the early
1990s, the number of MP subscribers started to increase steeply, reaching a penetration rate of
50% in Europe in 2000, in the United States in 2005, in developing countries in 2008, and even
in the least developed countries in 2010 (54). The overall number of MP subscriptions exceeded
the worldwide population in 2016 (54).
Second-, third-, and fourth-generation MPs and cordless phones emit radiofrequency elec-
tromagnetic fields (RF-EMF) in the frequency range of 700–2,700 MHz, and 5G is expected
to use the frequency spectrum up to 80 GHz. Wireless phones, i.e. MPs and cordless phones,
used close to the body produce a near-field exposure situation, and the specific absorption rate
(SAR, in W/kg tissue weight) is the most relevant exposure metric (48). In general, the SAR
decreases with the square of the distance to the source. MPs are relatively strong transmitters
because they have to reach longer distances than do other common RF-EMF sources [e.g., wire-
less local area networks (WLAN), cordless phones]. Thus, for MP users, the most significant ex-
posure contribution to the brain arises from these devices when held to the head during voice
calls (94). Exposure contributions from RF-EMF far-field sources such as WLANs, MP base
stations, or broadcast transmitters to the brain and to the whole body are typically below 10%
(32, 94).
Epidemiologic research on the carcinogenicity of RF-EMF has focused mainly on tumors de-
veloping in the head, in which organs and tissues are more exposed than other parts of the body.
For earlier studies of MP use, cumulative call time was shown to be a good predictor of exposure
(12, 117), but the validity of this exposure proxy lessened with more recent applications (61) and
technologies. The main reason is the adaptive power control in response to the network quality,
which has become very efficient for Universal Mobile Telecommunications System (UMTS, 3G)
technology (34). Minimum emission levels of a UMTS phone used with optimal network quality
can be more than 100,000 times lower than under the worst-case situation of bad network qual-
ity and maximum power emission. In real-world situations, the average output power differences
between GSM calls and UMTS calls are between a factor of 100 and 500 (34, 83). Consequently,
although the amount of MP use has increased over time as MP-related costs have decreased, cu-
mulative RF-EMF exposure among people mainly using their phone in the UMTS era is expected
to be considerably lower than in those with long durations of MP use in the GSM era, as well.
Additional exposure uncertainty comes from the changing usage patterns: Users tend to hold the

25.2 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

phone to the head less frequently now, as compared with the early decades of MP use, but they
attend to the screen more when using various phone applications.

Biological Plausibility for Carcinogenicity

Although MP batteries also produce extremely low-frequency electromagnetic fields (15), RF-
EMF emissions are of more significant concern for potential health risk. Because RF-EMF be-
longs to the nonionizing part of the EMF spectrum, the photon energy is too weak to directly
ionize molecules (18) and thereby cause direct DNA damage. Absorption of RF-EMF is known
to heat biological tissue owing to its electrical conductivity. Apart from the thermal interaction,
despite considerable research efforts, no other mechanism relevant for carcinogenesis has been
consistently established (91). Oxidative stress was seen to be increased after RF-EMF in some in
vitro and in vivo studies (23). However, no convincing evidence of changes in protein and gene
expression induced by low-level RF-EMF was obtained from cell culture or rodent brain tissue
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

experiments (68). In a pooled analysis of 15 in vitro and microarray studies of RF-EMF exposure,
investigators observed no strong link to any known pathway of human diseases, including cancer
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

(81).
Among the numerous animal studies conducted, some results indicated an increased cancer
risk (31, 79) or tumor promotion in mice coexposed to carcinogenic chemicals (64, 111). The
large-scale US National Toxicology Program (NTP) experiment investigated the carcinogenic-
ity of lifetime exposure to RF-EMF in rats (79) and mice (78). Whole-body SAR values up to
6 W/kg were applied in rats and up to 10 W/kg in mice, which is far higher than the whole-body
standard for the public (0.08 W/kg) but within the range of the regulatory limits for localized
sources such as MP handsets for the public (2 W/kg) and for workers (10 W/kg) (48). For male
rats, the NTP concluded that there is “clear evidence of tumors in the hearts of male rats” for the
incidence of heart schwannoma, with 5 cases in the highest GSM exposure group and 6 cases in
the highest CDMA (code-division multiple access) exposure versus 0 cases in the sham-exposed
group. Further, they concluded “some evidence of tumors in the brains and in the adrenal glands
of male rats.” Only “equivocal evidence” of carcinogenicity was seen for all outcomes in female rats
as well as in male and female mice. Thus, the observed carcinogenicity may be causal or a chance
finding due to multiple testing. Alternatively, it may be the consequence of temperature-induced
metabolic changes in male rats, where a measurable increase in core temperature was registered.
The latter might also explain the unexpected significantly longer lifetime of exposed male rats
compared with their sham controls.

Epidemiology of Intracranial and Salivary Gland Tumors

Intracranial tumors are rare diseases, with incidence rates below 10 per 100,000. They are a het-
erogeneous family of neoplasms, including tumors occurring in the brain, cerebral meninges, cra-
nial nerves, and pituitary gland, with more than 100 histologically distinct types (87). Overall, the
most frequently reported histology is meningioma (37%), followed by gliomas (25%), pituitary
gland tumors (16%), and nerve sheath tumors (8%). The vast majority of nerve sheath tumors are
vestibular schwannomas (also known as acoustic neuromas) (80).
Gliomas account for 70% of adult malignant primary brain tumors (92). The role of envi-
ronmental factors in the etiology of glioma is not well understood (80, 90). High-dose ionizing
radiation, as applied for cancer treatment, is an established risk factor (80, 90). Radiation risk de-
creases with age, and the observed median interval between exposure and diagnosis of a radiation-
induced glioma is between 9 and 18 years (69). The evidence concerning other exposures—such as

www.annualreviews.org • Mobile Phone Use and Brain Tumors 25.3

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

smoking and other lifestyle factors, organic chemicals, N-nitroso compounds including passive
smoking, pesticides, extremely-low-frequency magnetic fields, and estrogen-only menopausal
hormone therapy—is too inconsistent to ascribe causation (9, 21, 80, 90).
Also poorly understood is the etiology of meningioma, a slow-growing, mostly benign tumor
originating from the meninges. High-dose ionizing radiation is a causal factor for meningioma,
with a median latency period ranging between 17 and 23 years (69). Meningioma is much more
common in women than men (ratio about 2:1), whereas glioma is somewhat more common in
men.
Acoustic neuroma is a slow-growing benign tumor of the myelin-forming cells of the VIII cra-
nial nerve. Apart from the inherited disorder neurofibromatosis type 2 and high-dose radiation, no
other risk factors are clearly established, although chronic noise exposure is suspected to be a risk
factor (20). Tumors of the pituitary gland are usually benign neoplasms, and the most common
histological type is pituitary adenoma (77). Salivary gland tumors include neoplasms of the parotid
and of other salivary glands. Most salivary neoplasms are benign (70). Epidemiological research
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

on the etiology of childhood CNS tumors came to similar conclusions as did studies on CNS tu-
mors in adults, with no established environmental factors other than high-dose ionizing radiation
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

(86).

STRENGTHS AND LIMITATIONS OF EPIDEMIOLOGICAL STUDY

DESIGNS IN THIS RESEARCH CONTEXT
Case-Control Studies
In a case-control study, a group of individuals with the disease under investigation (cases) and a
group of subjects without the disease (controls) are compared with respect to the exposure of in-
terest. A crucial aspect of this study design is the selection of controls, which provides information
about the exposure distribution in the population from which the cases arose without introducing
selection bias. Selection bias may occur in case-control studies if not everybody who was initially
sampled is willing to participate and if willingness to participate is related to both exposure and
case-control status (95). Nonparticipation validation studies in the Interphone study, for example,
found that non-MP users were less likely to participate than were MP users and that participa-
tion in controls was lower than in cases (51, 73); as a consequence, prevalence of MP use was
more overestimated in controls than in cases, resulting in a downward bias of approximately 10%
(118). Bias (in both directions) can be introduced if cases and controls differ in the completeness
of answers to questions in interviews or questionnaires.
By design, in a case-control study, exposure information must be collected retrospectively, af-
ter identification of cases and controls. In case-control studies on MP use, this data collection has
been done almost exclusively by self-completed questionnaires or interviews, which allows inves-
tigators to obtain detailed information about MP use histories and confounding factors. However,
for both cases and controls, participants have found it very difficult to accurately recall MP use
habits from many years prior, and misclassification inevitably occurs, as demonstrated in several
validation studies (3, 35, 53, 56, 72, 82, 104, 112, 115). Misclassification entirely independent of
the case status, i.e., nondifferential, usually leads to a dilution of the risk estimates toward unity.
There is also evidence of systematic misclassification; light users tended to underestimate, and
heavy users tended to overestimate, their amount of phone use (35, 116); this tendency, although
nondifferential between cases and controls, would lead to an inflation of any association.
Additionally, some researchers are concerned that the disease affects the reporting of the ex-
posure, resulting in recall bias. Cases may overestimate their previous MP use as a potential cause

25.4 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

of their disease, whereas controls may not have thought much about past MP use and will thus
underestimate it. If cases overestimate, and/or controls underestimate, exposure, an overestima-
tion of the risk or a spurious association will occur. The determination that this type of differential
exposure misclassification is indeed relevant comes from a validation study, in which cases tended
to overestimate their MP use further back in time; this behavior was not observed among con-
trols (114). Indication for recall bias may also be derived from a Swedish study that included cases
diagnosed between 2007 and 2009 (34). In this study, the proportion of MP users reporting to
have started to use analog or digital MPs before the corresponding technology was actually im-
plemented was significantly higher in cases than in controls.
One particular challenge is the reporting of the side of the head predominantly used for call-
ing. A recent validation study in young people aged between 10 and 24 years from 12 countries
demonstrated that correlation between self-reported and app-recorded side of use was moder-
ate for right side users and almost nonexistent for left side users (35). For cases, retrospective
assessment of the preferred head side for MP use may also be biased toward the side where a
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

tumor occurred (101). In contrast, controls do not have any motivation to differently report the
side of the head for making MP calls, which will ultimately produce a bias in laterality analyses
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

(101).
Operator-recorded traffic data is assumed to be more reliable than self-reported MP use; some
uncertainties remain, however, as people may use other phones or talk on voice-over-IP (e.g.,
Skype or WhatsApp), which is registered as data transfer and not counted as voice call duration.
To date, only one case-control study on childhood brain tumors considered retrospectively col-
lected operator-recorded MP use (5). When retrospectively collected, operator-recorded data have
additional limitations: Not all operators retain traceable traffic data for long periods; the MP user
may not be the subscriber, e.g., phones may be redistributed within families; and study participants
may not remember their previous phone numbers, which are needed for record identification. The
motivation to remember and share data from old subscriptions may depend on the case-control
status, which could introduce bias.

Cohort Studies
Cohort studies start with a disease-free population and follow the cohort members over time
for occurrence of the studied disease (or diseases), while the exposures of interest are collected
at baseline and ideally on a continual basis during the follow-up period. Comparability between
exposed and unexposed must be ensured, e.g., through control of confounding. Selection bias is
usually not a problem in cohort studies, but it is crucial to have mechanisms in place for follow-up
of the cohort members. Otherwise, if loss to follow-up is related to exposure status, bias might be
introduced. For cancer outcomes that are usually rare, the cohort needs to be very large, which
often leads to the collection of less detailed exposure information than in case-control studies. Too
crude exposure information may hamper the ability to detect effects restricted to small subgroups
with specific exposure characteristics. Only a few cohort studies on tumors of the head and wireless
phone use have been conducted so far (10, 11, 30, 33, 103, 105), and a large international study
(COSMOS) is still ongoing (102, 112). Some of the cohort studies have used only register-based
exposure information, with few details about MP use. In contrast, other studies have used self-
reported information about MP use at baseline, which may change over time and is thus subject to
nondifferential exposure misclassification and may yield a bias toward the null in the case of a true
association. An important difference compared with case-control studies is that when exposure
information is collected before the occurrence of the disease, the likelihood of differential exposure
misclassification, i.e., recall bias, is minimized.

www.annualreviews.org • Mobile Phone Use and Brain Tumors 25.5

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

Ecological Studies
In an ecological study, disease incidence or prevalence is compared in space and/or over time.
Data are usually retrieved from routine statistics, such as cancer registers. Such comparisons of
aggregated data are often affected by confounding or ecological fallacy (95), and thus ecological
studies are usually considered weak and useful only for hypothesis generation. For the specific
question of carcinogenicity of MP use, analyses of cancer incidence time trends may be valuable
for several reasons.
First, prior to the MP era, RF-EMF exposure of the head was negligible except for a few spe-
cific working environments; this level of exposure has changed dramatically since the mid-1990s.
If indeed MP use will increase the risk of developing a tumor, the corresponding cancer inci-
dence rates worldwide should have increased substantially, unless compensated by just as sudden
changes in exposures to other, currently unknown, strong protective or risk factors for head tu-
mors. Second, such analyses do not need individual and complex exposure assessment because
they capitalize from the marked change in exposure on the population level. MP use spread very
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

distinctly in different age and sex groups; i.e., measurable incidence increases should be seen first
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

and be stronger in men compared with women and among those who were in their 30s to 50s
when MP technology began to be used more frequently. Third, time trends of incidence are not
prone to participation bias as seen in case-control studies.
A prerequisite for an evaluation of time trends of tumors of the head is the availability of high-
quality registry data with virtually complete tumor registration over long time periods. Moreover,
one must take into consideration that incidence rates of many brain tumors have been increasing
in many countries since the introduction and more frequent use of magnetic resonance imaging
and computer tomography (71). However, this rise began in the 1980s before MP use had become
widely prevalent. Similarly, changes over time in the registration of benign intracranial tumors
(67), in histology coding practice due to improved diagnostics, or in the population acceptance rate
for autopsy will affect incidence rates of brain tumor subtypes. These caveats should be considered
when interpreting incidence time trends, but such studies are nevertheless informative for hazard
identification in this specific context.

Case-Case Studies
Radiofrequency exposure during MP use is highly localized and declines rapidly with distance
from the exposure source. The energy absorption reaches only a few centimeters into the brain.
Thus, tumors in MP users, if directly caused by the RF-EMF, would be expected to be located
more often close to the exposure source. A few case-case studies have been conducted to test this
hypothesis using different methods.
A case-case analysis may reduce various types of biases, in particular control-selection bias.
However, depending on the method applied, bias may be introduced owing to underlying as-
sumptions about the spatial distribution of tumors in the head. In addition, recall bias may be of
concern if some of the exposure information used in such analyses is collected retrospectively by
interview or questionnaire.

STUDY RESULTS
In the following section, we present an overview of study results on MP use and risk of intracra-
nial and salivary gland tumors, including a meta-analysis of case-control and cohort studies pub-
lished up to December 31, 2017. Meta risk (mRR) estimates are shown for “ever versus never use”

25.6 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

and “long-term use” (i.e., time since first use of at least 10 years). For multicountry studies [i.e.,
Interphone (51, 52)] or studies of the same protocol but in different phases [i.e., Swedish Örebro
(40, 46)], inclusion in the meta-analyses was restricted to the most comprehensive analyses (for
details, see Supplemental Data).

Glioma
Glioma is the most frequently studied type of tumor in relation to MP use (25 case-control and 3
cohort studies on adults; Supplemental Data). Some investigations involved shared populations;
therefore, the meta-analysis was restricted to 12 unique, nonoverlapping studies (2, 7, 11, 22, 33,
40, 46, 50, 51, 74, 109, 120). The Interphone international analysis (51) and the Swedish Örebro
study (40, 46) contribute most of the cases.
On the basis of 4,197 exposed glioma cases, an mRR of 1.00 [95% confidence interval (CI)
0.89–1.13] was obtained for ever users compared with nonusers with substantial heterogeneity
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

across studies (I2 = 60%, p = 0.003) (Supplemental Figure 1). Two studies reported a statistically
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

significantly decreased glioma risk (11, 51), and two studies demonstrated increased risks (2, 40).
The mRR for long-term use (>10 years) was nonsignificantly elevated (mRR = 1.11, 0.85–
1.46) on the basis of 1,018 exposed cases (Figure 1) with considerable heterogeneity across studies
(I2 = 82%, p < 0.001). One cohort study reported a significantly decreased odds ratio (OR) of
0.77 (0.62–0.96) (11) and one case-control study reported an increased OR of 1.69 (1.40–2.03),
which is the pooled estimate of all latency categories >10 years presented in Hardell & Carlberg
(40). Supplemental Figure 2 shows the glioma risk in relation to cumulative duration of MP use.
In 4 out of 7 studies, investigators observed a significant increased risk in the highest category of
MP use: >896 h (OR = 2.89, 1.41–5.93) in the French study (22), >1,640 h (OR = 1.40, 1.03–
1.89) in Interphone (51), and >2,001 and >2,377 h (OR = 3.7, 1.7–7.7 and OR = 2.8, 1.6–4.8) in
the two Swedish studies, respectively (41, 43). No upward trend in OR with increasing cutoffs of
the highest exposure category is derivable from these estimates.
To check the implications of a potential glioma risk from MP use, several studies have assessed
how much the incidence of glioma or brain cancer would increase over time under various risk
and latency time scenarios. The incidence time trends in the United States (66), the Nordic coun-
tries (26, 27), and Australia (19) are not consistent with substantially increased risk from MP use
as observed in some of the studies in Figure 1. A comprehensive analysis of global trends in brain
and CNS tumors, including data from 1993 to 2007 from 96 registries in 39 countries, did not find
an overall pattern supporting the hypothesis of increasing incidence rates following, with some
latency, the time period of MP uptake in different populations, as outlined in the first paragraph
of this article (71). Increases in brain cancer incidences over time were observed in some stud-
ies. However, in most instances the increase did not follow the dissemination of MPs but rather
started earlier (28, 57, 76, 98) or was limited to the elderly (29, 55, 93), among whom MP use was
uncommon until recently. In other publications, the increase was limited to specific topographic
or morphologic subtypes of brain cancer and compensated by a decrease in complementary di-
agnoses, as seen in Israel (8), England (24, 85), Sweden (39), or the United States (65). Thus,
the findings may be explained by changes in the availability of information and coding practices,
particularly related to brain and CNS tumors of unknown type (D43) and brain cancers with
unknown intracerebral location (ICD-O-3 code C71.9) or morphology (e.g., glioma malignant
NOS, ICD-O-3 code 9380) (1, 58).
A case-case analysis of a subset of Interphone data from 7 European countries was performed
to assess whether gliomas of 888 cases occurred more often in brain areas closest (<5 cm) to a MP
held to the ear and areas most exposed to RF-EMF emitted by the device (63). However, no such

www.annualreviews.org • Mobile Phone Use and Brain Tumors 25.7

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

Glioma (long-term use) Meningioma (long-term use)

STUDY SEX CASES RR (95% CI) STUDY SEX CASES RR (95% CI)

Cohort Cohort
Frei et al. 2011 M 117 1.04 (0.85, 1.26) Frei et al. 2011 M 21 0.90 (0.57, 1.42)
Frei et al. 2011 F 10 1.04 (0.56, 1.95) Frei et al. 2011 F 8 0.93 (0.46, 1.87)
Benson et al. 2014 F 135 0.77 (0.62, 0.96) Benson et al. 2014 F 63 1.08 (0.78, 1.49)
2
Subtotal (I = 52.2%, p = 0.124) 0.92 (0.72, 1.16) 2
Subtotal (I = 0.0%, p = 0.794) 1.00 (0.78, 1.29)

Case control Case control

Interphone 2010 M+F 252 0.98 (0.76, 1.26) Interphone 2010 M+F 110 0.83 (0.61, 1.14)
Coureau et al. 2014 M+F 22 1.61 (0.85, 3.09) Coureau et al. 2014 M+F 10 1.57 (0.64, 3.86)
Yoon et al. 2015 M+F 100 1.04 (0.52, 2.09) Carlberg & Hardell 2015 M + F 346 1.10 (0.92, 1.32)
Hardell & Carlberg 2015b M + F 382 1.69 (1.40, 2.03) Subtotal (I 2 = 37.0%, p = 0.204) 1.02 (0.80, 1.30)
Subtotal (I 2 = 76.0%, p = 0.006) 1.30 (0.90, 1.87)
Overall (I 2 = 0.0%, p = 0.596) 1.03 (0.90, 1.17)
Overall (I 2 = 82.0%, p < 0.001) 1.11 (0.85, 1.46)

0.2 1.0 5.0 0.2 1.0 5.0

Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

Relative risk Relative risk

Access provided by 95.85.69.66 on 01/12/19. For personal use only.

Neuroma (long-term use) Pituitary tumors (long-term use)

STUDY SEX CASES RR (95% CI) STUDY SEX CASES RR (95% CI)

Cohort Cohort
Schüz et al. 2011b M 15 0.88 (0.52, 1.48) Benson et al. 2014 F 11 1.61 (0.78, 3.35)
Benson et al. 2014 F 14 1.17 (0.60, 2.27) Subtotal (I 2 = .%, p = .) 1.61(0.78, 3.34)
Subtotal (I 2 = 0.0%, p = 0.509) 0.98 (0.65, 1.48)
Case control
Case control Schoemaker & M+F 24 1.00 (0.50, 1.90)
Interphone 2011 M+F 68 0.76 (0.52, 1.11) Swerdlow 2009
Han et al. 2012 M+F 92 1.29 (0.69, 2.43) Shrestha et al. 2015 M+F 7 0.59 (0.21, 1.65)
Hardell et al. 2013b M+F 58 2.49 (1.74, 3.56) Subtotal (I 2 = 0.0%, p = 0.400) 0.86 (0.49, 1.50)
Pettersson et al. 2014 M+F 103 1.11 (0.76, 1.61)
2
Subtotal (I = 85.8%, p < 0.001) 1.29 (0.74, 2.23) Overall (I 2 = 20.8%, p = 0.283) 1.07 (0.64, 1.77)

Overall (I 2 = 78.3%, p < 0.001) 1.19 (0.80, 1.79)

0.2 1.0 5.0 0.2 1.0 5.0

Relative risk Relative risk

Salivary tumors (long-term use)

STUDY SEX BEHAVIOR CASES RR (95% CI)

Case control
Hardell et al. 2004 M+F Any 6 0.65 (0.27, 1.59)
Söderqvist et al. 2012 M+F Malignant 2 0.30 (0.10, 1.40)
Lönn et al. 2006 M+F Malignant 2 0.40 (0.10, 2.60)
Lönn et al. 2006 M+F Benign 7 1.40 (0.50, 3.90)
Sadetzki et al. 2008 M+F Malignant 1 0.47 (0.05, 4.51)
Sadetzki et al. 2008 M+F Benign 12 0.93 (0.44, 1.98)

Overall (I 2 = 0.0%, p = 0.491) 0.74 (0.48, 1.15)

0.2 1.0 5.0

Relative risk

Figure 1
Meta-analyses of tumors of the head and long-term (>10 years) mobile phone use. Note, odds ratios for Hardell & Carlberg 2015 (40)
(glioma) and Hardell et al. 2013 (44) (neuroma) have been derived by pooling their odds ratios of all latency categories >10 years for
mobile phone use. The orange dashed line represents the mean meta risk (mRR) of all studies of the corresponding graph. Where only
one study is shown, no measure of heterogeneity can be provided. Abbreviations: CI, confidence interval; F, female; I², percentage of
variation across studies that is due to heterogeneity rather than chance; M, male; p, p-value of the heterogeneity test; RR, relative risk.

25.8 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

pattern was found, and the mean distance between tumor midpoint and the phone axis at the ear
was similar among regular MP users and never or nonregular users. This study did not use any
of the self-reported exposure data except usage status and is therefore unlikely to be affected by
recall bias, although nondifferential exposure misclassification cannot be avoided. In an additional
case-case study of a subset of Interphone data from five other countries, the duration and amount
of MP use among people with tumors in highly exposed areas of the brain were compared with
the corresponding characteristics in patients with tumors located in other parts of the brain (16).
On the basis of 11 exposed cases, this study found some indications that people with gliomas in
the most exposed brain areas are more likely to be long-term MP users. Self-reported duration of
use in this paper may be subject to recall bias for more distant use, as indicated in an Interphone
validation study (114). In another analysis of 792 gliomas from the Interphone study, investigators
observed a statistically significant association between the intracranial distribution of gliomas and
the self-reported location of the phone (36). However, as acknowledged by the authors, this type
of analysis is potentially influenced by recall bias with respect to the laterality of MP use.
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

Meningioma
Overall, 18 case-control studies and 2 cohort studies on meningioma have been conducted (see
Supplemental Data). In the meta-analysis of 9 unique, nonoverlapping studies with a total of
2,686 exposed cases (2, 7, 11, 17, 22, 33, 46, 50, 51) ever use of MP was inversely associated with
meningioma risk (mRR = 0.91; 95% CI 0.84–0.98) (Supplemental Figure 1) and not associated
with long-term use of >10 years (mRR = 1.03, 0.90–1.17) on the basis of 558 exposed cases, with
no heterogeneity between studies in both meta-analyses (Figure 1). In a French study (22), cumu-
lative duration of >896 h resulted in an OR of 2.57 (1.02 to 6.44), and in a Swedish study >1,486 h
of cumulative MP use yielded an OR of 1.3 (1.1–1.6). In the Interphone study, meningioma
risk for >1,640 h of cumulative MP use was not significantly elevated [OR = 1.15 (0.81–1.62)].
The few published time trend analyses do not indicate an increased incidence among men since
the introduction of MP, whereas an increase in meningioma incidence in women started before
the introduction of MP (27). The only case-case study to date did not observe that MP use was
more common in people with a meningioma in the most exposed brain regions (16).

Acoustic Neuroma
Nineteen case-control studies and 2 cohort studies on acoustic neuroma have been conducted, and
11 studies with 1,546 exposed cases were included in the meta-analysis (see Supplemental Data).
The pattern of results is similar to that observed for glioma. Neither ever MP use (mRR = 1.02,
95% CI 0.84–1.24, number of exposed cases: 1,546) (Supplemental Figure 1) nor long-term use
(mRR = 1.19, 0.80–1.79, n = 350) (Figure 1) was associated with acoustic neuroma risk. The
heterogeneity across studies is substantial in both meta-analyses. In relation to the cumulative
duration of MP use, four studies found increased risk estimates for the highest usage category,
although the estimates were not always statistically significant (Supplemental Figure 3): >680 h
(OR = 1.46, 0.98–2.17), >1,001 h (OR = 3.1, 1.5–6.4), >1,487 h (OR = 2.6, 1.5–4.4) in three
Swedish studies (42, 44, 84), and >1,640 h (OR = 1.32, 0.88–1. 97) in Interphone (52). Like for
glioma, no upward trend in OR with increasing cutoffs was seen.
Sparse data available on trends in acoustic neuroma incidence do not indicate an increase since
MP use became widespread (62, 75). A case-case analysis of 787 acoustic neuroma cases from
Japan found some indication of increased risk for ipsilateral use (97), in particular among heavy
users (>20 min/day). Cases with ipsilateral frequent use were found to have tumors with smaller

www.annualreviews.org • Mobile Phone Use and Brain Tumors 25.9

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

diameters, which may suggest a detection bias because hearing capacity decreased with progressing
disease. Thus, people using the ear with the tumor may realize sooner that they might have a
unilateral hearing loss. Such detection bias would explain the seeming association between the
side of MP use and occurrence of the tumor. A Swedish case-control study also indicated that
laterality analysis for this specific type of tumor could be biased (84).

Pituitary Tumors
Only 4 case-control studies (45, 100, 106, 110) and 1 cohort study (10) addressed the risk of pitu-
itary tumor from MP use, contributing 375 exposed cases to the meta-analysis. Overall, ever use
of MPs was not associated with pituitary tumor risk (mRR = 0.86; 95% CI 0.56–1.31), although
between-study heterogeneity was high (Supplemental Figure 1). Risk for long-term use was
1.07 (0.65–1.77) on the basis of 3 studies with 42 exposed cases (Figure 1).
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

Salivary Gland Tumors

Seven case-control studies and one cohort study on the risk of salivary gland tumors were included
in the meta-analysis. On the basis of 657 exposed benign or malignant salivary gland tumor cases,
no risk increase (mRR = 0.92, 0.80–1.06) among ever MP users was observed (Supplemental
Figure 1), with little heterogeneity between studies. Similarly, there was no indication of a long-
term risk (mRR = 0.74, 0.48–1.15) (Figure 1). No increase in incidence of malignant parotid
gland tumors was observed in Swedish and Nordic adults between 1970 and 2009 (107).

Childhood Brain Tumors

So far, results on the association between wireless phone use and childhood brain tumors in
children and adolescents are available from only one case-control study. Participants were aged
7–19 years and diagnosed with a brain tumor between 2004 and 2008 in Denmark, Norway,
Sweden, and Switzerland (CEFALO study) (5). Overall, the results of the CEFALO study do not
suggest a causal association because the brain tumor risk was not elevated in brain regions that are
most exposed when using a MP, and no consistent exposure–response association was observed in
relation to several self-reported and operator-recorded exposure measures, despite some sporadic
statistically significant associations. Another study in 10- to 24-year-old individuals (MOBI-Kids)
is ongoing (96).
Notable increases in astrocytoma risk among persons who started to use wireless phones before
age 20 years (fivefold for MP and fourfold for cordless phones) were observed in a Swedish case-
control study (38). Now, most young adults have started to use MPs before the age of 20, and
therefore risks of such magnitude are not compatible with incident time trends in Sweden or
other countries for this age group (26, 66). Other case-control studies have not found increased
risks in the youngest age group (60).
A time trend study of brain tumor incidence in children and adolescents (5 to 19 years) found
decreasing rather than increasing rates between 2000 and 2008 in Sweden (4). Stable trends were
observed in the Nordic countries for children aged 0–10 years between 1985 and 2006 (99) and
for individuals aged 5–19 years between 1990 and 2008 (108), in England among individuals aged
10–20 years between 1998 and 2007 (25), in Australia for individuals aged 0–19 years between
2000 and 2008 (29), and in the United States for individuals aged 0–19 years between 1977 and
2006 (49) and for those aged 5–19 years between 1990 and 2007 (13).

25.10 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

DISCUSSION AND CONCLUSION

In 2011, the IARC (International Agency for Research on Cancer) classified RF-EMF as “possibly
carcinogenic to humans” (Group 2B) based on limited evidence in humans and in experimental
animals (6, 47). The epidemiology IARC Working Group based their evaluation mainly on
findings from the Interphone and Swedish Örebro case-control studies of glioma and acoustic
neuroma and stated, “While both of these are susceptible to bias, the Working Group concluded
that these findings could not be dismissed as reflecting bias alone, and that a causal interpretation
was possible” (47, p. 410). Since then, several new studies or study updates have been conducted
(see cumulative meta-analysis in Supplemental Figure 4). In our meta-analysis of long-term
MP users, no indication emerged for an increased risk of meningioma, pituitary, and salivary
gland tumors, whereas meta-estimates of glioma and acoustic neuroma risk were slightly above
1, with confidence intervals including the null effect (Figure 1) and considerable across-study
heterogeneity. A sensitivity analysis using different data sets of unique and nonoverlapping
studies shows that our findings are robust to the choice of data set as long as they do not overlap
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

(Supplemental Figure 5). The observed absence of risk is in line with earlier meta-analyses (59,
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

91), whereas some recent meta-analyses reported significantly increased brain tumor risk from
long-term MP use (14, 88, 119). However, they paid less attention to avoid multiple counting of
the same individual data or combined different disease entities.
In light of the inconsistent epidemiological study results for glioma and acoustic neuroma,
the most relevant question is whether some of the studies showing no association missed a true
risk or whether some of the studies showing an association are, in fact, falsely positive. The risk
might have been underestimated owing to nondifferential exposure misclassification, in particular
by the cohort studies lacking detailed exposure information, relying on subscriber status before a
given time point (33), or relying on answers to a few basic questions on how often (daily versus
less often) and how long a MP had been used (10). This misclassification is expected to dilute
any exposure–response relation if there is a true association. The ongoing COSMOS study is
collecting operator data prospectively and will not suffer from this type of bias (102). In case-
control studies, bias to null from substantial nondifferential exposure misclassification may over-
compensate some recall bias (73, 116). Thus, it is theoretically conceivable that a real risk went
undetected.
However, the mRR estimates of glioma and acoustic neuroma in long-term users were driven
mainly by the pooled Örebro studies with average ORs for all MP latency categories >10 years
of 1.69 (1.40–2.03) for glioma (40) and of 2.49 (1.74–3.56) for acoustic neuroma (44). Simple
calculations demonstrate that such excess risks would not have been unnoticed in clinical practice
by now. The populations from the Nordic countries were among the first to use MPs regularly,
and a 50% penetration rate was achieved in Europe in 2000. Now, in Sweden substantially more
than 50% of the population is a long-term MP user, and an excess glioma risk on the order of
60–70% would yield an increase of at least 30% in glioma incidence rates, which has not been
observed in Swedish people aged <70 years (Supplemental Figure 6). An observed excess risk
of about 150% for acoustic neuroma would produce an even stronger increase in incidence rates.
Published time trend analyses do not indicate any noticeable increase in brain tumor incidence
since the introduction of MPs. Nevertheless, these studies cannot prove the absence of risk, as
they are not sensitive to small increases in incidence of rare histologic subtypes. Current time
trend analyses would not yet pick up a risk increase occurring at latency periods of more than 15–
20 years. However, assuming a similar latency for nonionizing radiation as observed for ionizing
radiation, one would expect that any relevant risk should already have started to emerge by now
(26, 66, 71, 98).

www.annualreviews.org • Mobile Phone Use and Brain Tumors 25.11

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

These inconsistencies should encourage investigators to revisit those case-control studies with
significant excess risk to investigate what design or conduct feature led to the overestimation of
risk. False-positive findings could be produced by recall bias, as discussed above. Thus, a compar-
ison of the exposure distribution in the controls with public statistics is another cross-check to
evaluate the plausibility of self-reported MP use. In the four Örebro case-control studies includ-
ing cases diagnosed between 1994–2003 and 2007–2009, the proportion of MP users in controls
has not increased at the same pace as in the Swedish population according to the Swedish Post
and Telecom Agency (89). Because of the rapid uptake of MP use over time, it is important that
MP exposure is evaluated up to the same calendar period for cases and controls in order to avoid
bias (113).
Recall bias is also a likely explanation for the increased risk for ipsilateral use observed in some
studies because in these studies contralateral tended to be protective, which is biologically implau-
sible (101). For acoustic neuroma, this type of analysis is particularly vulnerable to bias owing to
potential diagnostic detection bias.
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

In summary, current evidence from all available studies including in vitro, in vivo, and epidemi-
ological studies does not indicate an association between MP use and tumors developing from the
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

most exposed organs and tissues. Given the large amount of research on this topic, any potentially
undetected risk is expected to be small from an individual perspective and might concern long la-
tency periods (>15 years), rare brain tumor subtypes, and MP usage during childhood. To address
such small risks, high-quality research with accurate exposure assessment is needed, taking into
account that MP call duration alone is not expected to adequately reflect RF-EMF exposure to
the brain.

DISCLOSURE STATEMENT
M.F. is vice chairman of the International Commission on Non-Ionizing Radiation Protection,
an independent body setting guidelines for nonionizing radiation protection. She has served
as advisor to a number of national and international public advisory and research steering
groups concerning the potential health effects of exposure to nonionizing radiation, including the
World Health Organization. M.R. is member of the International Commission on Non-Ionizing
Radiation Protection. From 2011 to 2018, M.R. was an unpaid member of the foundation board of
the Swiss Research Foundation for Electricity and Mobile Communication, a non-profit research
foundation at ETH Zurich. Neither industry nor nongovernmental organizations are represented
on the scientific board of the foundation.

ACKNOWLEDGMENTS
M.S. is funded by the UK-based charity Breast Cancer Now. The Institute of Cancer Research ac-
knowledges National Health Service funding to the Royal Marsden/Institute of Cancer Research
National Institute for Health Research (NIHR) Biomedical Research Centre (BRC).

LITERATURE CITED
1. Ahlbom A, Feychting M, Holmberg L, Johansson LA, Mathiesen T, et al. 2015. Comments on Hardell
and Carlberg Increasing Rates of Brain Tumors in the Swedish National Inpatient Register and the
Causes of Death Register. Int. J. Environ. Res. Public Health 2015, 12, 3793–3813. Int. J. Environ. Res.
Public Health 12:11662–64

25.12 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

2. Auvinen A, Hietanen M, Luukkonen R, Koskela RS. 2002. Brain tumors and salivary gland cancers
among cellular telephone users. Epidemiology 13:356–59
3. Aydin D, Feychting M, Schüz J, Andersen TV, Poulsen AH, et al. 2011. Impact of random and system-
atic recall errors and selection bias in case–control studies on mobile phone use and brain tumors in
adolescents (CEFALO study). Bioelectromagnetics 32:396–407
4. Aydin D, Feychting M, Schüz J, Röösli M, CEFALO Study Team. 2012. Childhood brain tumours
and use of mobile phones: comparison of a case-control study with incidence data. Environ. Health
11:35
5. Aydin D, Feychting M, Schüz J, Tynes T, Andersen TV, et al. 2011. Mobile phone use and brain tumors
in children and adolescents: a multicenter case-control study. J. Natl. Cancer Inst. 103:1264–76
6. Baan R, Grosse Y, Lauby-Secretan B, El Ghissassi F, Bouvard V, et al. 2011. Carcinogenicity of radiofre-
quency electromagnetic fields. Lancet Oncol. 12:624–26
7. Baldi I, Coureau G, Jaffré A, Gruber A, Ducamp S, et al. 2011. Occupational and residential exposure to
electromagnetic fields and risk of brain tumors in adults: a case-control study in Gironde, France. Int. J.
Cancer 129:1477–84
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

8. Barchana M, Margaliot M, Liphshitz I. 2012. Changes in brain glioma incidence and laterality correlates
with use of mobile phones—a nationwide population based study in Israel. Asian Pac. J. Cancer Prev.
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

13:5857–63
9. Benson VS, Kirichek O, Beral V, Green J. 2015. Menopausal hormone therapy and central nervous
system tumor risk: large UK prospective study and meta-analysis. Int. J. Cancer 136:2369–77
10. Benson VS, Pirie K, Schüz J, Reeves GK, Beral V, et al. 2013. Mobile phone use and risk of brain
neoplasms and other cancers: prospective study. Int. J. Epidemiol. 42:792–802
11. Benson VS, Pirie K, Schüz J, Reeves GK, Beral V, Green J. 2014. Authors’ response to: the case of
acoustic neuroma: comment on mobile phone use and risk of brain neoplasms and other cancers. Int. J.
Epidemiol. 43:275
12. Berg G, Schüz J, Samkange-Zeeb F, Blettner M. 2005. Assessment of radiofrequency exposure from
cellular telephone daily use in an epidemiological study: German Validation study of the international
case-control study of cancers of the brain–INTERPHONE-Study. J. Expo. Anal. Environ. Epidemiol.
15:217–24
13. Boice JD Jr., Tarone RE. 2011. Cell phones, cancer, and children. J. Natl. Cancer Inst. 103:1211–13
14. Bortkiewicz A, Gadzicka E, Szymczak W. 2017. Mobile phone use and risk for intracranial tumors and
salivary gland tumors—a meta-analysis. Int. J. Occup. Med. Environ. Health 30:27–43
15. Calderon C, Ichikawa H, Taki M, Wake K, Addison D, et al. 2017. ELF exposure from mobile and
cordless phones for the epidemiological MOBI-Kids study. Environ. Int. 101:59–69
16. Cardis E, Armstrong BK, Bowman JD, Giles GG, Hours M, et al. 2011. Risk of brain tumours in relation
to estimated RF dose from mobile phones: results from five Interphone countries. Occup. Environ. Med.
68:631–40
17. Carlberg M, Hardell L. 2015. Pooled analysis of Swedish case-control studies during 1997–2003 and
2007–2009 on meningioma risk associated with the use of mobile and cordless phones. Oncol. Rep.
33:3093–98
18. Challis LJ. 2005. Mechanisms for interaction between RF fields and biological tissue. Bioelectromagnetics
Suppl. 7:S98–106
19. Chapman S, Azizi L, Luo Q, Sitas F. 2016. Has the incidence of brain cancer risen in Australia since the
introduction of mobile phones 29 years ago? Cancer Epidemiol. 42:199–205
20. Chen M, Fan Z, Zheng X, Cao F, Wang L. 2016. Risk factors of acoustic neuroma: systematic review
and meta-analysis. Yonsei. Med. J. 57:776–83
21. Connelly JM, Malkin MG. 2007. Environmental risk factors for brain tumors. Curr. Neurol. Neurosci.
Rep. 7:208–14
22. Coureau G, Bouvier G, Lebailly P, Fabbro-Peray P, Gruber A, et al. 2014. Mobile phone use and brain
tumours in the CERENAT case-control study. Occup. Environ. Med. 71:514–22
23. Dasdag S, Akdag MZ. 2016. The link between radiofrequencies emitted from wireless technologies and
oxidative stress. J. Chem. Neuroanat. 75:85–93

www.annualreviews.org • Mobile Phone Use and Brain Tumors 25.13

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

24. de Vocht F. 2016. Inferring the 1985–2014 impact of mobile phone use on selected brain cancer subtypes
using Bayesian structural time series and synthetic controls. Environ. Int. 97:100–7
25. de Vocht F, Burstyn I, Cherrie JW. 2011. Time trends (1998–2007) in brain cancer incidence rates in
relation to mobile phone use in England. Bioelectromagnetics 32:334–39
26. Deltour I, Auvinen A, Feychting M, Johansen C, Klaeboe L, et al. 2012. Mobile phone use and incidence
of glioma in the Nordic countries 1979–2008: consistency check. Epidemiology 23:301–7
27. Deltour I, Johansen C, Auvinen A, Feychting M, Klaeboe L, Schüz J. 2009. Time trends in brain tumor
incidence rates in Denmark, Finland, Norway, and Sweden, 1974–2003. J. Natl. Cancer Inst. 101:1721–24
28. Ding LX, Wang YX. 2011. Increasing incidence of brain and nervous tumours in urban Shanghai, China,
1983–2007. Asian Pac. J. Cancer Prev. 12:3319–22
29. Dobes M, Khurana VG, Shadbolt B, Jain S, Smith SF, et al. 2011. Increasing incidence of glioblastoma
multiforme and meningioma and decreasing incidence of Schwannoma (2000–2008): findings of a mul-
ticenter Australian study. Surg. Neurol. Int. 2:176
30. Dreyer NA, Loughlin JE, Rothman KJ. 1999. Cause-specific mortality in cellular telephone users. JAMA
282:1814–16
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

31. Falcioni L, Bua L, Tibaldi E, Lauriola M, De Angelis L, et al. 2018. Report of final results regarding brain
and heart tumors in Sprague-Dawley rats exposed from prenatal life until natural death to mobile phone
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

radiofrequency field representative of a 1.8GHz GSM base station environmental emission. Environ. Res.
165:496–503
32. Foerster M, Thielens A, Joseph W, Eeftens M, Röösli M. 2018. A prospective cohort study of adoles-
cents’ memory performance and individual brain dose of microwave radiation from wireless communi-
cation. Environ. Health Perspect. 126:077007
33. Frei P, Poulsen AH, Johansen C, Olsen JH, Steding-Jessen M, Schüz J. 2011. Use of mobile phones and
risk of brain tumours: update of Danish cohort study. BMJ 343:d6387
34. Gati A, Hadjem A, Wong M-F, Wiart J. 2009. Exposure induced by WCDMA mobiles phones in oper-
ating networks. IEEE Trans. Wireless Commun. 8:5723–27
35. Goedhart G, van Wel L, Langer CE, de Llobet Viladoms P, Wiart J, et al. 2018. Recall of mobile phone
usage and laterality in young people: the multinational Mobi-Expo study. Environ. Res. 165:150–57
36. Grell K, Frederiksen K, Schüz J, Cardis E, Armstrong B, et al. 2016. The intracranial distribution of
gliomas in relation to exposure from mobile phones: analyses from the INTERPHONE study. Am. J.
Epidemiol. 184:818–28
37. Han YY, Berkowitz O, Talbott E, Kondziolka D, Donovan M, Lunsford LD. 2012. Are frequent dental x-
ray examinations associated with increased risk of vestibular schwannoma? J. Neurosurg. 117(Suppl.):78–
83
38. Hardell L, Carlberg M. 2009. Mobile phones, cordless phones and the risk for brain tumours. Int. J.
Oncol. 35:5–17
39. Hardell L, Carlberg M. 2015a. Increasing rates of brain tumours in the Swedish national inpatient reg-
ister and the causes of death register. Int. J. Environ. Res. Public Health 12:3793–813
40. Hardell L, Carlberg M. 2015b. Mobile phone and cordless phone use and the risk for glioma—analysis
of pooled case-control studies in Sweden, 1997–2003 and 2007–2009. Pathophysiology 22:1–13
41. 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 79:630–39
42. Hardell L, Carlberg M, Hansson Mild K. 2006. Pooled analysis of two case-control studies on the use
of cellular and cordless telephones and the risk of benign brain tumours diagnosed during 1997–2003.
Int. J. Oncol. 28:509–18
43. Hardell L, Carlberg M, Söderqvist F, Mild KH. 2013a. Case-control study of the association between
malignant brain tumours diagnosed between 2007 and 2009 and mobile and cordless phone use. Int. J.
Oncol. 43:1833–45
44. Hardell L, Carlberg M, Söderqvist F, Mild KH. 2013b. Pooled analysis of case-control studies on acous-
tic neuroma diagnosed 1997–2003 and 2007–2009 and use of mobile and cordless phones. Int. J. Oncol.
43:1036–44

25.14 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

45. Hardell L, Hallquist A, Mild KH, Carlberg M, Påhlson A, Lilja A. 2002. Cellular and cordless telephones
and the risk for brain tumours. Eur. J. Cancer Prev. 11:377–86
46. Hardell L, Näsman A, Påhlson A, Hallquist A, Hansson Mild K. 1999. Use of cellular telephones and
the risk for brain tumours: a case-control study. Int. J. Oncol. 15:113–16
47. IARC Work. Group Eval. Carcinog. Risks Hum. 2013. Non-Ionizing Radiation, Part 2: Radiofrequency
Electromagnetic Fields. Lyon: IARC
48. ICNIRP (Int. Comm. Non-Ioniz. Radiat. Prot.). 1998. Guidelines for limiting exposure to time-varying
electric, magnetic, and electromagnetic fields (up to 300 GHz). Health Phys. 74:494–522
49. Inskip PD, Hoover RN, Devesa SS. 2010. Brain cancer incidence trends in relation to cellular telephone
use in the United States. Neuro Oncol. 12:1147–51
50. Inskip PD, Tarone RE, Hatch EE, Wilcosky TC, Shapiro WR, et al. 2001. Cellular-telephone use and
brain tumors. N. Engl. J. Med. 344:79–86
51. Interphone Study Group. 2010. Brain tumour risk in relation to mobile telephone use: results of the
INTERPHONE international case-control study. Int. J. Epidemiol. 39:675–94
52. Interphone Study Group. 2011. Acoustic neuroma risk in relation to mobile telephone use: results of
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

the INTERPHONE international case-control study. Cancer Epidemiol. 35:453–64

53. Inyang I, Benke G, Morrissey J, McKenzie R, Abramson M. 2009. How well do adolescents recall use
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

of mobile telephones? Results of a validation study. BMC Med. Res. Methodol. 9:36
54. ITU (Int. Telecommun. Union). 2017. Key ICT indicators for developed and developing countries and
the world (totals and penetration rates). ITU Key ICT Indicators. https://idp.nz/Global-Rankings/ITU-
Key-ICT-Indicators/6mef-ytg6
55. Kim SJ, Ioannides SJ, Elwood JM. 2015. Trends in incidence of primary brain cancer in New Zealand,
1995 to 2010. Aust. N. Z. J. Public Health 39:148–52
56. Kiyohara K, Wake K, Watanabe S, Arima T, Sato Y, et al. 2018. Long-term recall accuracy for mobile
phone calls in young Japanese people: a follow-up validation study using software-modified phones.
J. Expo. Sci. Environ. Epidemiol. 28:166–72
57. Kohler BA, Ward E, McCarthy BJ, Schymura MJ, Ries LA, et al. 2011. Annual report to the nation on
the status of cancer, 1975–2007, featuring tumors of the brain and other nervous system. J. Natl. Cancer
Inst. 103:714–36
58. Kopel E. 2012. Incidence shifts within central nervous system malignancies. Epidemiology 23:767–68
59. Lagorio S, Röösli M. 2014. Mobile phone use and risk of intracranial tumors: a consistency analysis.
Bioelectromagnetics 35:79–90
60. Lahkola A, Auvinen A, Raitanen J, Schoemaker MJ, Christensen HC, et al. 2007. Mobile phone use and
risk of glioma in 5 North European countries. Int. J. Cancer 120:1769–75
61. Langer CE, de Llobet P, Dalmau A, Wiart J, Goedhart G, et al. 2017. Patterns of cellular phone use
among young people in 12 countries: implications for RF exposure. Environ. Int. 107:65–74
62. Larjavaara S, Feychting M, Sankila R, Johansen C, Klaeboe L, et al. 2011. Incidence trends of vestibular
schwannomas in Denmark, Finland, Norway and Sweden in 1987–2007. Br. J. Cancer 105:1069–75
63. Larjavaara S, Schüz J, Swerdlow A, Feychting M, Johansen C, et al. 2011. Location of gliomas in relation
to mobile telephone use: a case-case and case-specular analysis. Am. J. Epidemiol. 174:2–11
64. Lerchl A, Klose M, Grote K, Wilhelm AF, Spathmann O, et al. 2015. Tumor promotion by exposure to
radiofrequency electromagnetic fields below exposure limits for humans. Biochem. Biophys. Res. Commun.
459:585–90
65. Li K, Lu D, Guo Y, Wang C, Liu X, et al. 2018. Trends and patterns of incidence of diffuse glioma in
adults in the United States, 1973–2014. Cancer Med. 7:5281–90
66. Little MP, Rajaraman P, Curtis RE, Devesa SS, Inskip PD, et al. 2012. Mobile phone use and glioma risk:
comparison of epidemiological study results with incidence trends in the United States. BMJ 344:e1147
67. McCarthy BJ, Kruchko C, Dolecek TA. 2013. The impact of the Benign Brain Tumor Cancer Registries
Amendment Act (Public Law 107–260) on non-malignant brain and central nervous system tumor in-
cidence trends. J. Registry Manag. 40:32–35
68. McNamee JP, Chauhan V. 2009. Radiofrequency radiation and gene/protein expression: a review. Radiat.
Res. 172:265–87

www.annualreviews.org • Mobile Phone Use and Brain Tumors 25.15

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

69. McNeill KA. 2016. Epidemiology of brain tumors. Neurol. Clin. 34:981–98
70. Mehanna H, McQueen A, Robinson M, Paleri V. 2013. Salivary gland swellings. Clin. Otolaryngol. 38:58–
65
71. Miranda-Filho A, Piñeros M, Soerjomataram I, Deltour I, Bray F. 2017. Cancers of the brain and CNS:
global patterns and trends in incidence. Neuro Oncol. 19:270–80
72. Mireku MO, Mueller W, Fleming C, Chang I, Dumontheil I, et al. 2018. Total recall in the SCAMP
cohort: validation of self-reported mobile phone use in the smartphone era. Environ. Res. 161:1–
8
73. Momoli F, Siemiatycki J, McBride ML, MÉ Parent, Richardson L, et al. 2017. Probabilistic multiple-
bias modeling applied to the Canadian data from the Interphone Study of mobile phone use and risk of
glioma, meningioma, acoustic neuroma, and parotid gland tumors. Am. J. Epidemiol. 186:885–93
74. Muscat JE, Malkin MG, Thompson S, Shore RE, Stellman SD, et al. 2000. Handheld cellular telephone
use and risk of brain cancer. JAMA 284:3001–7
75. Nelson PD, Toledano MB, McConville J, Quinn MJ, Cooper N, Elliott P. 2006. Trends in acoustic
neuroma and cellular phones: Is there a link? Neurology 66:284–85
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

76. Nomura E, Ioka A, Tsukuma H. 2011. Trends in the incidence of primary intracranial tumors in Osaka,
Japan. Jpn. J. Clin. Oncol. 41:291–94
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

77. Ntali G, Wass JA. 2018. Epidemiology, clinical presentation and diagnosis of non-functioning pituitary
adenomas. Pituitary 21:111–18
78. NTP (Natl. Toxicol. Program). 2018. Toxicology and carcinogenesis studies in B6C3F1/N mice exposed to
whole-body radio frequency radiation at a frequency (1900 MHz) and modulations (GSM and CDMA) used by
cell phones. NTP Tech. Rep. 596, Natl. Inst. Health, Bethesda, MD. https://www.niehs.nih.gov/ntp-
temp/tr596_508.pdf
79. NTP (Natl. Toxicol. Program). 2018. Toxicology and carcinogenesis studies in Hsd:Sprague Dawley SD rats
exposed to whole-body radio frequency radiation at a frequency (900 MHz) and modulations (GSM and CDMA)
used by cell phones. NTP Tech. Rep. 595, Natl. Inst. Health, Bethesda, MD. https://www.niehs.nih.gov/
ntp-temp/tr595_508.pdf
80. Ostrom QT, Gittleman H, Stetson L, Virk SM, Barnholtz-Sloan JS. 2015. Epidemiology of gliomas.
Cancer Treat Res. 163:1–14
81. Parham F, Portier CJ, Chang X, Mevissen M. 2016. The use of signal-transduction and metabolic path-
ways to predict human disease targets from electric and magnetic fields using in vitro data in human cell
lines. Front. Public Health 4:193
82. Parslow RC, Hepworth SJ, McKinney PA. 2003. Recall of past use of mobile phone handsets. Radiat.
Prot. Dosim. 106:233–40
83. Persson T, Tornevik C, Larsson LE, Loven J. 2012. Output power distributions of terminals in a 3G
mobile communication network. Bioelectromagnetics 33:320–25
84. Pettersson D, Mathiesen T, Prochazka M, Bergenheim T, Florentzson R, et al. 2014. Long-term mobile
phone use and acoustic neuroma risk. Epidemiology 25:233–41
85. Philips A, Henshaw DL, Lamburn G, O’Carroll MJ. 2018. Brain tumours: rise in glioblastoma multi-
forme incidence in England 1995–2015 suggests an adverse environmental or lifestyle factor. J. Environ.
Public Health 2018:7910754
86. Pollack IF, Jakacki RI. 2011. Childhood brain tumors: epidemiology, current management and future
directions. Nat. Rev. Neurol. 7:495–506
87. Pouchieu C, Baldi I, Gruber A, Berteaud E, Carles C, Loiseau H. 2016. Descriptive epidemiology and
risk factors of primary central nervous system tumors: current knowledge. Rev. Neurol. 172:46–55
88. Prasad M, Kathuria P, Nair P, Kumar A, Prasad K. 2017. Mobile phone use and risk of brain tumours: a
systematic review of association between study quality, source of funding, and research outcomes. Neurol.
Sci. 38:797–810
89. PTS (Post Telecom Auth.). 2011. The Swedish Post and Telecom Agency: statistics portal. Online portal, PTS,
Stockholm, accessed May 24. http://sokstat.pts.se/en
90. Reni M, Mazza E, Zanon S, Gatta G, Vecht CJ. 2017. Central nervous system gliomas. Crit. Rev. Oncol.
Hematol. 113:213–34

25.16 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

91. Repacholi MH, Lerchl A, Röösli M, Sienkiewicz Z, Auvinen A, et al. 2012. Systematic review of wireless
phone use and brain cancer and other head tumors. Bioelectromagnetics 33:187–206
92. Ricard D, Idbaih A, Ducray F, Lahutte M, Hoang-Xuan K, Delattre JY. 2012. Primary brain tumours
in adults. Lancet 379:1984–96
93. Röösli M, Michel G, Kuehni CE, Spoerri A. 2007. Cellular telephone use and time trends in brain
tumour mortality in Switzerland from 1969 to 2002. Eur. J. Cancer Prev. 16:77–82
94. Roser K, Schoeni A, Struchen B, Zahner M, Eeftens M, et al. 2017. Personal radiofrequency electro-
magnetic field exposure measurements in Swiss adolescents. Environ. Int. 99:303–14
95. Rothman KJ, Greenland S. 1998. Modern Epidemiology. Philadelphia: Lippincott Williams & Wilkins.
2nd ed.
96. Sadetzki S, Langer CE, Bruchim R, Kundi M, Merletti F, et al. 2014. The MOBI-Kids study proto-
col: challenges in assessing childhood and adolescent exposure to electromagnetic fields from wireless
telecommunication technologies and possible association with brain tumor risk. Front. Public Health
2:124
97. Sato Y, Akiba S, Kubo O, Yamaguchi N. 2011. A case-case study of mobile phone use and acoustic
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

neuroma risk in Japan. Bioelectromagnetics 32:85–93

98. Sato Y, Kiyohara K, Kojimahara N, Yamaguchi N. 2016. Time trend in incidence of malignant neo-
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

plasms of the central nervous system in relation to mobile phone use among young people in Japan.
Bioelectromagnetics 37:282–89
99. Schmidt LS, Schmiegelow K, Lahteenmaki P, Träger C, Stokland T, et al. 2011. Incidence of childhood
central nervous system tumors in the Nordic countries. Pediatr. Blood Cancer 56:65–69
100. Schoemaker MJ, Swerdlow AJ. 2009. Risk of pituitary tumors in cellular phone users: a case-control
study. Epidemiology 20:348–54
101. Schüz J. 2009. Lost in laterality: interpreting “preferred side of the head during mobile phone use and
risk of brain tumour” associations. Scand. J. Public Health 37:664–67
102. Schüz J, Elliott P, Auvinen A, Kromhout H, Poulsen AH, et al. 2011a. An international prospective
cohort study of mobile phone users and health (Cosmos): design considerations and enrolment. Cancer
Epidemiol. 35:37–43
103. Schüz J, Jacobsen R, Olsen JH, Boice JD Jr., McLaughlin JK, Johansen C. 2006. Cellular telephone use
and cancer risk: update of a nationwide Danish cohort. J. Natl. Cancer Inst. 98:1707–13
104. Schüz J, Johansen C. 2007. A comparison of self-reported cellular telephone use with subscriber data:
agreement between the two methods and implications for risk estimation. Bioelectromagnetics 28:130–
36
105. Schüz J, Steding-Jessen M, Hansen S, Stangerup SE, Cayé-Thomasen P, et al. 2011b. Long-term mobile
phone use and the risk of vestibular schwannoma: a Danish nationwide cohort study. Am. J. Epidemiol.
174:416–22
106. Shrestha M, Raitanen J, Salminen T, Lahkola A, Auvinen A. 2015. Pituitary tumor risk in relation to
mobile phone use: a case-control study. Acta Oncol. 54:1159–65
107. Shu X, Ahlbom A, Feychting M. 2012. Incidence trends of malignant parotid gland tumors in Swedish
and Nordic adults 1970 to 2009. Epidemiology 23:766–67
108. Söderqvist F, Carlberg M, Hansson Mild K, Hardell L. 2011. Childhood brain tumour risk and its
association with wireless phones: a commentary. Environ. Health 10:106
109. Spinelli V, Chinot O, Cabaniols C, Giorgi R, Alla P, Lehucher-Michel MP. 2010. Occupational and
environmental risk factors for brain cancer: a pilot case-control study in France. Presse Med. 39:e35–44
110. Takebayashi T, Varsier N, Kikuchi Y, Wake K, Taki M, et al. 2008. Mobile phone use, exposure to
radiofrequency electromagnetic field, and brain tumour: a case-control study. Br. J. Cancer 98:652–59
111. Tillmann T, Ernst H, Streckert J, Zhou Y, Taugner F, et al. 2010. Indication of cocarcinogenic poten-
tial of chronic UMTS-modulated radiofrequency exposure in an ethylnitrosourea mouse model. Int. J.
Radiat. Biol. 86:529–41
112. Toledano MB, Auvinen A, Tettamanti G, Cao Y, Feychting M, et al. 2018. An international prospective
cohort study of mobile phone users and health (COSMOS): factors affecting validity of self-reported
mobile phone use. Int. J. Hyg. Environ. Health 221:1–8

www.annualreviews.org • Mobile Phone Use and Brain Tumors 25.17

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)
 PU40CH25_Roosli ARjats.cls December 28, 2018 11:41

113. Turner MC, Sadetzki S, Langer CE, Villegas R, Figuerola J, et al. 2016. Investigation of bias related to
differences between case and control interview dates in five INTERPHONE countries. Ann. Epidemiol.
26:827–32.e2
114. Vrijheid M, Armstrong BK, Bédard D, Brown J, Deltour I, et al. 2009. Recall bias in the assessment of
exposure to mobile phones. J. Expo. Sci. Environ. Epidemiol. 19:369–81
115. Vrijheid M, Cardis E, Armstrong BK, Auvinen A, Berg G, et al. 2006. Validation of short term recall of
mobile phone use for the Interphone study. Occup. Environ. Med. 63:237–43
116. Vrijheid M, Deltour I, Krewski D, Sanchez M, Cardis E. 2006. The effects of recall errors and of selection
bias in epidemiologic studies of mobile phone use and cancer risk. J. Expo. Sci. Environ. Epidemiol. 16:371–
84
117. Vrijheid M, Mann S, Vecchia P, Wiart J, Taki M, et al. 2009. Determinants of mobile phone output
power in a multinational study: implications for exposure assessment. Occup. Environ. Med. 66:664–71
118. Vrijheid M, Richardson L, Armstrong BK, Auvinen A, Berg G, et al. 2009. Quantifying the impact of
selection bias caused by nonparticipation in a case-control study of mobile phone use. Ann. Epidemiol.
19:33–41
Annu. Rev. Public Health 2019.40. Downloaded from www.annualreviews.org

119. Wang Y, Guo X. 2016. Meta-analysis of association between mobile phone use and glioma risk. J. Cancer
Res. Ther. 12:C298–300
Access provided by 95.85.69.66 on 01/12/19. For personal use only.

120. Yoon S, Choi J-W, Lee E, An H, Choi HD, Kim N. 2015. Mobile phone use and risk of glioma: a
case-control study in Korea for 2002–2007. Environ. Health Toxicol. 30:e2015015

25.18 Röösli et al.

Review in Advance first posted on
January 11, 2019. (Changes may
still occur before final publication.)