ArticlePDF AvailableLiterature Review

Abstract and Figures

Since the International Agency for Research on Cancer recently classified radiofrequency electromagnetic fields, such as those emanating from mobile and cordless phones, as possibly carcinogenic to humans (group 2B), two additional reports relevant to the topic have been published. Both articles were new updates of a Danish cohort on mobile phone subscribers and concern the possible association between assumed use of mobile phones and risk of brain tumors. The aim of the present review is to reexamine all four publications on this cohort. In brief, publications were scrutinized, and in particular, if the authors made explicit claims to have either proved or disproved their hypothesis, such claims were reviewed in light of applied methods and study design, and in principle, the stronger the claims, the more careful our review. The nationwide Danish cohort study on mobile phone subscribers and risk of brain tumors, including at best 420,095 persons (58% of the initial cohort), is the only one of its kind. In comparison with previous investigations, i.e., case-control studies, its strength lies in the possibility to eliminate non-response, selection, and recall bias. Although at least non-response and recall bias can be excluded, the study has serious limitations related to exposure assessment. In fact, these limitations cloud the findings of the four reports to such an extent that render them uninformative at best. At worst, they may be used in a seemingly solid argument against an increased risk--as reassuring results from a large nationwide cohort study, which rules out not only non-response and recall bias but also an increased risk as indicated by tight confidence intervals. Although two of the most comprehensive case-control studies on the matter both have methodological limitations that need to be carefully considered, type I errors are not the only threats to the validity of studies on this topic--the Danish cohort study is a textbook example of that.
Content may be subject to copyright.
Rev Environ Health 2012;27(1):51–58 © 2012 by Walter de Gru yter • Berlin • Boston. DOI 10.1515/reveh-2 012-00 04
*Corresponding author: Fredrik S ö derqvist, DmedSc, Department
of Oncology, University Hospital, SE-701 82 Ö rebro, Sweden
E-mail: fredrik. soderqvist@orebroll.se
Received December 13, 2011; accepted January 12, 2012
Review of four publications on the Danish cohort study on
mobile phone subscribers and risk of brain tumors
Keywords: acoustic neuroma; bias; digital enhanced cordless
telecommunication (DECT) phone; glioma; high-grade astro-
cytoma; meningioma; mobile phone.
Background
Today, about 5 billion persons worldwide are estimated to use a
mobile phone. In addition, cordless desktop phones of the digi-
tal enhanced cordless telecommunication (DECT) type are also
commonly used. Because these phones are often held close to
the head when used for talking, the near-fi eld exposure to radio-
frequency electromagnetic fi eld (RF-EMF) radiation on the
brain is relatively high. This exposure has generated concerns
about adverse health effects, mainly an increased risk of brain
tumors. Indeed, the two most comprehensive studies so far point
to an increased risk of glioma (1 3) and acoustic neuroma (4,
5) associated with long-term use of mobile phones and/or cord-
less phones. Although the evidence is not yet conclusive, the
results cannot be dismissed as refl ecting bias alone (6) . Thus, a
causal interpretation of an association between wireless phone
RF-EMF exposure and brain tumors is possible.
A recent evaluation of the carcinogenicity of RF-EMF
radiation, including the frequency range 30 kHz–300 GHz,
was made by the International Agency for Research on
Cancer (IARC) in May 2011. Both occupational sources and
personal devices, such as mobile phones and cordless phones,
were included in the evaluation. RF-EMF exposure was clas-
sifi ed as possibly carcinogenic to humans, group 2B (6) .
Regarding wireless phones and the risk for brain tumors, the
evidence came mainly from the Hardell group in Sweden and
the INTERPHONE study group (6) . Both contributed with
case-control studies, but the INTERPHONE study included
several research centers from different countries. The studies
were performed during a comparable period with no overlap
of cases between the Swedish part of INTERPHONE and the
Hardell studies. A study from Japan gave further evidence of
an increased risk for acoustic neuroma associated with ipsi-
lateral mobile phone use (7) . There were additional studies
on the matter, three early case-control studies (8 10) and two
reports on a Danish cohort study (11, 12) ; however, these
studies were regarded as less informative or inconclusive.
After the outcome of the IARC evaluation was made public
in June 2011, two additional reports on the topic were quickly
published. Both were new updates on the Danish cohort of
mobile phone subscribers, and both showed no increased risk.
One focused on acoustic neuroma (13) , whereas the other
presented results for all cancers and separately for glioma and
meningioma (14) . The aim of this review is to reexamine all
four reports published on the Danish cohort study.
Fredrik S ö derqvist *, Michael Carlberg
and Lennart Hardell
Department of Oncology , University Hospital, SE-701 82
Ö rebro , Sweden
Abstract
Background: Since the International Agency for Research
on Cancer recently classifi ed radiofrequency electromag-
netic fi elds, such as those emanating from mobile and cord-
less phones, as possibly carcinogenic to humans (group 2B),
two additional reports relevant to the topic have been pub-
lished. Both articles were new updates of a Danish cohort on
mobile phone subscribers and concern the possible associa-
tion between assumed use of mobile phones and risk of brain
tumors. The aim of the present review is to reexamine all four
publications on this cohort.
Methods: In brief, publications were scrutinized, and in
particular, if the authors made explicit claims to have either
proved or disproved their hypothesis, such claims were
reviewed in light of applied methods and study design, and
in principle, the stronger the claims, the more careful our
review.
Results: The nationwide Danish cohort study on mobile
phone subscribers and risk of brain tumors, including at best
420,095 persons (58 % of the initial cohort), is the only one
of its kind. In comparison with previous investigations, i.e.,
case-control studies, its strength lies in the possibility to elim-
inate non-response, selection, and recall bias. Although at
least non-response and recall bias can be excluded, the study
has serious limitations related to exposure assessment. In fact,
these limitations cloud the fi ndings of the four reports to such
an extent that render them uninformative at best. At worst,
they may be used in a seemingly solid argument against an
increased risk as reassuring results from a large nationwide
cohort study, which rules out not only non-response and recall
bias but also an increased risk as indicated by tight confi dence
intervals.
Conclusion: Although two of the most comprehensive case-
control studies on the matter both have methodological limi-
tations that need to be carefully considered, type I errors are
not the only threats to the validity of studies on this topic
the Danish cohort study is a textbook example of that.
Authenticated | fredrik.soderqvist@orebroll.se
Download Date | 6/15/12 10:16 PM
52 S ö derqvist et al.: Mobile phones and risk of brain tumors
Methods
First, we address the initial design of the cohort, and then the four
reports are presented and discussed in the order in which they were
published. For each report, we begin by presenting the main fi ndings,
which are taken mainly from the abstract of the respective report. The
publications are scrutinized, and in particular, if the authors make
explicit claims to have either proved or disproved their hypothesis,
the purpose of the discussion will be to critically review those claims
in light of applied methods and study design, and in principle, the
stronger the claims, the more careful the review. For each report, we
also present and discuss further results, given that these are important
for the respective publication and its conclusions. Because all four
reports stem from the same cohort, methodological features are much
alike. The main strengths and weaknesses are therefore not repeat-
edly discussed unless it is necessary for the respective publication,
e.g., due to specifi c results. To further examine and contextualize
methodological issues of importance, the review also discusses other
studies on the matter. However, the latter is done only when relevant
for the results and discussion presented by the authors of the four
publications under review. For example, if the authors relate their
results to those in other studies in a questionable manner, our discus-
sion includes associated studies and fi ndings as well. Lastly, because
the fourth publication was accompanied by a most peculiar editorial,
we will also include that in the review, especially because it played a
large part in the report of the media of the study, at least in Sweden.
The Danish cohort study on mobile phone
subscribers
Cohort design
The cohort was based entirely on record linkage. Mobile
phone subscribers in Denmark from January 1, 1982, until
December 31, 1995, were identifi ed from the computerized
les of the two Danish operating companies, TeleDenmark
Mobil and Sonafon, which partly also funded the study. The
les included name of the subscriber (a person or a company),
address, and type of mobile phone used: analog phones of
either the Nordic Mobile Telephony (NMT) 450 or NMT
900 type or digital phone of the Global System for Mobile
Communications (GSM) type. Using the date of last invoice,
calculating the duration of use in months was possible,
although only for digital phones. For analog phones, no such
information was available; time from date of subscription was
used as a surrogate for duration of use. A total of 723,421 sub-
scribers were initially identifi ed. Of these, 200,507 corporate
users were excluded because information on the individual was
not available. The remaining subjects were linked to the fi les
of the Central Population Register. Table 1
details the num-
ber of persons fi rst identifi ed and later excluded in the initial
cohort setup. The fi nal cohort consisted of 58 % of the initially
identifi ed subscribers.
First publication Overall, 3391 cancers were reported in
the fi rst publication by Johansen et al. (11) . That number was
434 fewer than expected, giving a standardized incidence ratio
(SIR) of 0.89. A substantial proportion of the cancers occurred
at smoking-related cancer sites. No increased risk was seen
Table 1 The Danish cohort study with initially identifi ed and
excluded subjects.
Total number identifi ed 723,421
Excluded persons in total 303,326
Corporate users 200,507
E r r o r s i n n a m e 4 9 , 3 5 2
E r r o r s i n a d d r e s s 1 0 , 1 8 3
Address not residential 11,687
D u p l i c a t e s 1 0 , 6 7 9
Subscriptions after 1995 17,921
Under the age of 18 2550
R e s i d e n t s i n G r e e n l a n d /
Faroe Islands
3 9 4
Refused to participate 53
Remaining cohort 420,095 (58 % of total number)
Men 357,550 (85 % of the cohort)
Women 62,545 (15 % of the cohort)
for brain or nervous system cancers, and risk did not vary by
duration of mobile phone use (or rather subscription years),
time since fi rst subscription, age at fi rst subscription, or type
of mobile phone. The authors concluded that, In summary,
this fi rst nationwide cancer incidence study evaluating cellular
telephone use provides no support for an association between
use of these telephones and risk of brain cancer . ”.
Follow-up of the cohort began at the date of fi rst subscrip-
tion and ended at the date of death, emigration, or December
31, 1996. Overall, there was a statistically signifi cant decrease
of SIR for all cancer in men, 0.86; 95 % confi dence interval
(CI), 0.83 0.90, probably due to confounding. For example,
no adjustment was made for socioeconomic factors (income
and smoking). No statistically signifi cant increased risk was
found for any cancer site. There was no statistically signifi -
cant increased or decreased risk for brain and nervous system
tumors (n = 154). Most of the calculations were based on low
numbers, e.g., glioma (n = 66), meningioma (n = 16), cranial
nerve sheet tumors (n = 7). Only 24 cases had a latency time
of 5 years and the average follow-up time was 3.1 years.
The duration of subscription of digital phone was for the
“ highest ” category 3 years (n = 9). Actually, over two-thirds
of the subscriptions began in 1994 and 1995. In other words,
the majority of the cohort members had 2 years or less of sub-
scription time. Thus, in relation to assumed latency time for
brain tumors, SIR was calculated from extremely low number
of lifetime years. Noteworthy is that both the INTERPHONE
study group and the Hardell group used a minimum latency
time of 1 year, i.e., exposure was censored at 1 year before
the reference date.
To our knowledge, the Danish study is the only nationwide
cohort on the topic. In comparison with other investigations,
i.e., case-control studies, its strength lies in the possibility
to eliminate non-response, selection, and recall bias and, to
some degree, the fi rst publication also succeeded in doing
so. However, its many weaknesses and short follow-up cloud
the fi ndings to such an extent that renders it uninformative,
at best. Regarding exposure assessment, for example, no
Authenticated | fredrik.soderqvist@orebroll.se
Download Date | 6/15/12 10:16 PM
S ö derqvist et al.: Mobile phones and risk of brain tumors 53
for the whole group of brain and nervous system tumors
(p for trend=0.51). Separate calculations were not presented
for the largest group of interest, glioma, a tumor type for
which other studies at the time had shown excess risk for
latency period 10 years [for overview, see Hardell et al.
(15) ]. Instead, combined results for brain tumors were given
with a statistically signifi cant decreased SIR of 0.66 (in the
10 year latency group). Because there is no plausible bio-
logical evidence for an inverse association between RF fi elds
and brain tumors, this fi nding most likely refl ects substantial
bias in the study.
Also, the second publication had indications of exposure
misclassifi cation. Self-reported use of mobile phones was
assessed in the Danish INTERPHONE study (17) . Of the
controls, 822 subjects were identifi ed within the subscriber
list of the cohort using personal identifi cation numbers. It
was possible to classify them as being either regular mobile
phone users or not during 1982 1995. Of these subscrib-
ers, only 61 % reported making or receiving at least one call
a week during the last 6 months (12, 18) . Furthermore, the
fact that users who began their subscription after 1995 were
included in the unexposed reference population must have
given considerable exposure misclassifi cation. The authors
expressed little concern about this but rather made the fol-
lowing claim, Despite uncertainties in estimating actual tele-
phone use, the consistency of the fi ndings with case-control
studies conducted in Denmark and in other parts of the world
is reassuring. First, the number of subscription years is hardly
an estimate of actual use. The study could not even discrimi-
nate between the lightest and the heaviest users because no
cumulative quantifi cation of exposure was ever made, other
than subscription years as a less valid surrogate. Second, the
results were consistent only with the three early case-control
studies that the IARC working group dismissed as less infor-
mative (see Introduction). The results were not consistent
with the Hardell studies, which at that time had both the larg-
est numbers of cases and controls and the longest time since
rst use of wireless phones, as we have summarized (5, 19) .
Therefore, the statement on consistency by Sch ü z et al. (12)
is more likely to refl ect the limitations, which the three early
case-control studies share with the Danish cohort study, rather
than credible evidence against an association between use of
mobile phones and risk of brain tumors.
Third publication Approximately 2.9 million of the
Danish population of 5.5 million in total was included in this
record linkage study on acoustic neuroma published in 2011
by Sch ü z et al. (13) . Of the 2.9 million subjects, 420,095
were mobile phone subscribers that started their subscription
1987 1995 and, in accordance with the aim of the study, had
lasted for 11 years, i.e., 1998 2006, during which the tumor
cases were ascertained. No evidence of an increased risk was
found for 11 years of subscription (adjusted incidence rate
ratio = 0.87; 95 % CI, 0.52 1.46). Tumors did not occur more
on the right side of the head said to be the preferred side
among Danes when making or receiving mobile phone calls,
and the tumors were not of larger size among mobile phone
subscribers than expected (multivariate p = 0.95). The authors
information on cumulative hours of use was assessed, neither
in this nor its succeeding publications. No information on lat-
erality was assessed either, i.e., whether the phone was mostly
used on the same side of the head as the tumor developed.
Also, because this was a register-based study, no use of cord-
less phone was assessed. Those who used a cordless desktop
phone but did not have a mobile phone subscription were thus
classifi ed as unexposed, and there was no possibility to cal-
culate cumulative risk for those who had used both a mobile
and a cordless phone.
The Danish cohort study is not the only one on the mat-
ter that has either disregarded or not adequately analyzed
and presented use of cordless phone. Indeed, all studies but
those from the Hardell group (15) suffer from this drawback.
Cordless phones are often used equally much or even more
than mobile phones, and the emissions of these phones are
in the same order of magnitude (16) , which is why the IARC
evaluation was based on use of both mobile and cordless
phones, hence, the term wireless phones . The frequency of
cordless phone use has increased considerably since the 1990s.
Including persons with such use in the unexposed group
would preclude the possibility of detecting an increased risk.
Furthermore, regarding exposure misclassifi cation, it seems
as if the reference category, i.e., the Danish population not
only included the 303,326 excluded subscribers but also the
cohort members (see below). Note that all except the last of
the mentioned limitations in exposure assessment apply also
to the three subsequent publications on this cohort.
Second publication In the rst update on the Danish
cohort study, Sch ü z et al. (12) reported that mobile phone
use (or rather having a subscription) was not associated
with increased risk for brain tumors (SIR = 0.97) or acoustic
neuroma (SIR = 0.73). Nor was there an increased risk for brain
tumors for long-term subscribers of 10 years (indeed, SIR
was statistically signifi cantly decreased, 0.66; 95 % CI, 0.44
0.95). No trend was seen with time since fi rst subscription.
The authors specifi cally concluded, the narrow CIs provide
evidence that any large association of risk of cancer and
cellular telephone use can be excluded.
The 420,095 mobile phone subscribers were also included
in this updated publication with follow-up through 2002. The
median time since fi rst subscription was 8.0 years. It was
now stated that the cohort members were excluded from the
reference population, which seems not to have been the case
in the fi rst publication. A total of 580 brain and nervous sys-
tem tumors were observed in the cohort. Relatively few were
glioma (n = 257), meningioma (n = 68), or nerve sheet tumors
in cranial nerves (n = 32). A statistically signifi cant decreased
risk for all cancer among men was again obtained (SIR = 0.93;
95 % CI, 0.92 0.95). The SIR for glioma was close to unity,
1.01 (95 % CI, 0.89 1.14). Regarding different anatomical
localizations of glioma, the highest SIR was found for the
temporal lobe where RF-EMF exposure from a mobile phone
would be highest, 1.21 (95 % CI, 0.91 1.58, n = 54). For men-
ingioma, SIR = 0.86 (95 % CI, 0.67 – 1.09, n = 68), and for cra-
nial nerve sheet tumors, SIR = 0.73 (95 % CI, 0.50 – 1.03, n = 32)
were calculated. No effect of years of subscriptions was found
Authenticated | fredrik.soderqvist@orebroll.se
Download Date | 6/15/12 10:16 PM
54 S ö derqvist et al.: Mobile phones and risk of brain tumors
concluded, the results from this comprehensive study do not
support the hypothesis that use of a mobile phone increases
the risk of vestibular schwannoma [acoustic neuroma].
The abstract of this report gives the impression that the
study had great statistical power including 2.9 million sub-
jects, of which 420,950 (additionally 54,350 were excluded,
see below) were mobile phone subscribers, when in fact the
analysis of long-term exposure ( 11 years) was based on only
15 exposed cases, all of whom were men. The latter, however,
was not mentioned in the abstract. Nor was it mentioned that
the analysis of tumor size was based on even fewer cases;
eight had had a subscription for 11 years, all of them men.
As for the risk related to laterality, Sch ü z et al. (20) compared
the location of acoustic neuroma in long-term mobile phone
subscribers with short-term ( < 11 years) and non-subscribers
to see if tumors occurred more frequently on the side that was
assumed to be the mostly exposed. Assumed is the right
word because as mentioned in the Danish cohort, no data on
side of use were ever assessed. Instead, Sch ü z et al. used eco-
logical data from the prospective study COSMOS as proxy
(20) . The authors hypothesized that, because in COSMOS
most (53 % ) mobile phone users preferred using the phone on
the right side, most tumors would appear on that side, if an
increased risk exists. This argument is not very impressive
to begin with, considering the aggregated nature of the data,
and surely, it becomes no stronger when applied on only 15
exposed cases.
To corroborate their overall fi nding of no increased risk,
Sch ü z et al. (13) used another dubious argument. While
acknowledging the limitation of having low numbers of
exposed cases and results with wide CIs, the authors never-
theless claimed that since today the number of never users of
mobile phones is very small, and the proportion of long-term
users is increasing steadily to the extent that any mobile phone
use related vestibular schwannoma [acoustic neuroma] risk
would lead to a marked increase in the incidence rates of ves-
tibular schwannoma. This belief which is frequently used
also as argument in support of null results on risk of glioma
and meningioma (14) is unfounded for two reasons. The fi rst
one is related to latent periods for which the above end points
typically are around 10 to 40 years (21 24) . Only few studies,
however, have exposure duration that exceeds 10 years and
then solely for subgroups with limited numbers. Thus, in most
investigations, the exposure duration was incompatible with
the natural history of the disease for a tumor-initiating effect
to be studied. An initiating effect is what would have the most
direct effect on the incidence. The other reason concerns the
possibility of an early effect on tumor development (promo-
tion) and its consequences on the increase in incidences that
can possibly occur. As discussed in detail by Kundi (25, 26) ,
any such increase is limited by the magnitude of the shift of
the age-incidence function and its slope. In short, in the case
of brain tumors, this makes it unlikely to fi nd an increased
incidence for the time being; an increase may be seen fi rst in
a few years or only in a decade.
In discussing the limitations of their study, Sch ü z et al. (13)
did mention the latency issue, i.e., the still-short time duration
of exposure in relation to the natural history of the disease and
recognize that their fi nding of no increased risk was not unex-
pected. Having said that, the authors then moved on to ques-
tion the Hardell group s fi ndings of increased risks already
in the group of short-term users (5) . Although such fi ndings
are not unproblematic and call for a discussion of the meth-
ods used which has also been given, at length by Hardell
et al. (27) it is a false claim that only our studies found such
results. There are similar tendencies of increased risk also in
the INTERPHONE study on acoustic neuroma (appendix B
in ref. 4). Moreover, by questioning our results and meth-
ods, Sch ü z et al. (13) not only exclude the possibility of an
early effect on tumor development (that they studied them-
selves, see fi rst publication), in principle, they also question
the methods of virtually every other case-control study on the
topic. They claim our fi ndings would most likely be explained
by bias without further specifying what they mean by that
what type of bias ? This is typical of the nature of criticism
directed toward our studies; it is vague, poorly substantiated,
and often based on deviant and/or selective results without a
balanced discussion as to why they deviate. It is assumed that
because it is our results that stand out [which is not entirely
true (28) ], it must be our methods that are fl awed.
Our studies used the national population registry for ran-
dom recruitment of controls and have the highest response
rates among studies on the topic (29) . Thus, it is unlikely
that selection or response bias can explain the results, so we
assume that Sch ü z et al. (13) refer to recall bias. On this point,
we would like to be very clear attempts to estimate the mag-
nitude of recall bias related to long-term use specifi cally has
not been made by any research group. Validating cumula-
tive hours of self-reported use by, e.g., operator-verifi ed data
going back > 10 years in time is practically impossible on a
large data set. Within the INTERPHONE project, validation
studies were performed but only on short-term use (30 32) ,
and those studies showed considerable non-differential mis-
classifi cation of exposure, whereas differential misclassifi ca-
tion was less pronounced. Of course, we cannot exclude the
infl uence of recall bias on the results in our studies; however,
the important question is whether such an infl uence can fully
explain the observed risks ? According to the IARC evalua-
tion, they cannot (6) . Moreover, if recall bias would explain
the increased risk reported for high-grade glioma, why do
we not see the same pattern of increased risk for low-grade
glioma or for meningioma when the same methods were
used ? Interestingly, a prediction recently delineated by Kundi
(26) may provide an explanation of these differences. The
prediction has to do with the shift of the age-incidence func-
tion and its slope. Given that there is an increased risk and all
brain tumors are equally affected, tumor subtypes that have
a steeper slope of the age-log-incidence function should be
associated with higher risks. Hence, a higher risk would be
expected for glioblastoma (astrocytoma grade IV) than for
low-grade astrocytoma. For the same reason, i.e., the slope of
the age-log-incidence function for different tumor subtypes,
Kundi also elaborate on why the narrowing down of age
range in studies (as was done in the INTERPHONE study as
well as in the Danish cohort study) carries the risk of reducing
the ORs for mobile phone use further.
Authenticated | fredrik.soderqvist@orebroll.se
Download Date | 6/15/12 10:16 PM
S ö derqvist et al.: Mobile phones and risk of brain tumors 55
Fourth publication Frei et al. (14) reported no overall
increased risk of tumors of the central nervous system.
This result was true also when restricted to the individuals
with the longest mobile phone use, 13 years of assumed
subscription. The authors concluded that, In this update of a
large nationwide cohort study of mobile phone use, there were
no increased risks of tumors of the central nervous system,
providing little evidence for a causal association.
This time, the number of the cohort was reduced to 358,403
(49.5 % ) of the initially identifi ed subscribers (n = 723,421).
The major additional exclusion (n = 54,350) was due to record
linkage with the Danish so-called CANULI cohort (33) on
socioeconomic factors. That register started in 1990 and
included subjects from the age of 30 years while subscrip-
tion holders aged 18 29 years were excluded from the mobile
phone cohort; the same was the case for the third publication
(see above). Follow-up for the occurrence of cancer started at
January 1, 1990 or at age of 30, whichever occurred later, and
ended on the date of cancer diagnosis, date of death, date of
emigration, or December 31, 2007, whichever came fi rst. In
total, 3117 subscribers with previous cancer were excluded
from the cohort. The person time within the fi rst year of sub-
scription (n = 1556) or censored before the fi rst subscription
(n = 2660) did not contribute to person-years in the study. Nine
were excluded due to “other problems identifi ed later in data”.
Thus, fi nally < 50 % of the initial cohort was now included.
The study period was during 1990 2007, but the cohort
was established during 1982 1995. In the fi rst two publica-
tions (11, 12) , no lag time was used, but this time, some early
cases diagnosed with a nervous system tumor with latency
period up to 8 years were obviously disregarded. Cancer
cases before 1990 were disregarded because, as mentioned,
the CANULI cohort started in 1990 (33) . The authors did not
discuss the impact of the exclusion of these subscribers on
the results. This exclusion would include the early users of
analog phones, which seem to have had higher emissions of
RF-EMF than the later digital system. The authors themselves
also stated the following in their discussion, we found indi-
cations that early subscription holders before 1995 were in
fact heavier users (based on outgoing calls) compared with
all subscription holders in the years 1996 – 2002. ” “ Heavier ”
in this context refers to 23 min/week. Nevertheless, analy-
sis of any early effect in that group who used phones with
the highest emissions was most likely hampered. Moreover,
also the youngest users, aged 18 29 years that had previously
been included, were now excluded from the cohort. The lat-
ter is strange because, according to the authors, adjustment
of disposable income and level of education ( fully adjusted
model ) had no substantial effect on the risk estimate (i.e., the
results were the same as in the analysis adjusted only for age
and calendar period; basic model ). Adjustment for age and
calendar period should be possible also for the youngest users
because this information does not depend on data from the
CANULI cohort. Note also in the case of the third publication
that adjustments for variables obtained from CANULI (this
time also including marital status) did not have any notable
effect on the result (13) .
The exclusion of young subscribers could be of impor-
tance because previous studies have indicated highest risk
in subjects that started the use of a mobile or cordless phone
before the age of 20 years (34) [see discussion on reason
why (26) above]. We should mention that in the fi rst publi-
cation (11), results were given for three age groups at entry,
0 49 [obviously it should be 18 49 and not 0 49], 50 64,
and 65 years (Table 3 in ref. 11). In the three following
publications (12, 13, 14), no results were given for age at
entry but would certainly have been of interest as well as
taking the larger number of cancer cases in the present pub-
lication although the age group 18 29 years was excluded.
Furthermore, in the fi rst publication, the results were pre-
sented for different types of mobile phones ( analog, ana-
log and digital, and digital ) because different generations
of phones use different signals and output power. The dura-
tion of digital subscription in years was also analyzed. In the
following publications, these data were lacking despite that
the results would have been based on larger numbers with
longer follow-up. This would have been important informa-
tion, so why was it not presented ?
Like previous publications, no individual data were
given for tumor risk related to the side of the brain where
the mobile phone had been mostly used. The latter would
have been valuable information because an increased risk
has been consistently associated with ipsilateral use, less
so for contralateral use (1 3) . Although the authors stated
that no increased risk was found for temporal glioma (the
most plausible location if mobile phone use is a risk), this
nding is as mentioned of limited value without individual
information on the side that the phone was used in relation
to tumor site and considering all the limitations in exposure
assessment. Regarding exposure assessment, no operator-
verifi ed data on years of subscription were available. Thus,
one would expect considerable misclassifi cation of mobile
phone use both among subscribers and reference population
because no new subscribers were included in the exposed
cohort after 1995. In effect, this absence means that the
longer the follow-up, the more exposure misclassifi cation
would be expected due to the last decade s increasingly
higher penetration rate of wireless phone users in the Danish
population, the reference population. No doubt, the refer-
ence population may have more person-years at risk than the
exposed cohort.
Editorial by Ahlbom and Feychting The publication by
Frei et al. (14) was accompanied by an editorial by Ahlbom
and Feychting from the Karolinska Institute in Sweden.
It began with the statement, Evidence is reassuring, but
continued monitoring of health registers and prospective
cohorts is still warranted. Ahlbom and Feychting (35) then
opened their discussion by pointing out methodological
advantages that the Danish study is thought to have over
other studies on the topic, e.g., that non-response and
selection bias had been eliminated. First, we wonder
whether Ahlbom and Feychting forgot about the 200,507
early corporate users who were excluded from the cohort ?
Second, it is not likely that all previous studies suffered
Authenticated | fredrik.soderqvist@orebroll.se
Download Date | 6/15/12 10:16 PM
56 S ö derqvist et al.: Mobile phones and risk of brain tumors
Instead of addressing the limitations of the Danish cohort
study in full, Ahlbom and Feychting use their space to selec-
tively report on results in our studies, choosing the period
2000 2003 (36, 37) , although the whole investigation period
was 1997 – 2003 (5, 19) .
In support of the fi ndings that Frei et al. (14) present,
Ahlbom and Feychting (35) refer to data on brain tumor inci-
dence from the Swedish Cancer Registry and not from the
Danish Cancer Registry, which would have been more appro-
priate in this case, if at all relevant. The quality of the Swedish
Cancer Registry in reporting of central nervous system tumors,
from non-response and selection bias to such a degree that it
would be a threat to the validity of the results. As mentioned,
all the Hardell studies had high or very high response rates,
and the control groups were recruited randomly using the
Swedish Population Registry (34) .
Although Ahlbom and Feychting (35) were careful to point
out the assumed strengths of the Danish study over others, the
authors seem to have taken its limitations more lightly. They
mention one, having a mobile subscription is not equiva-
lent to using a mobile phone and conversely some users
will be non-subscribers, and conclude that this limitation
is important in a negative study because it leads to exposure
misclassifi cation and would dilute any association between
mobile phone use and risk of cancer. Indeed, it would, and
therefore, we fi nd it contradictory that Ahlbom and Feychting
(35) chose not to mention any further limitations related to
exposure assessment and misclassifi cation because all would
dilute an association.
Ahlbom et al. (18) wrote quite a contrasting letter to the
editor after publication of the fi rst update of the Danish
cohort study. Therein the authors raised several method-
ological shortcomings and questioned the authors con-
clusion that any large association of risk of cancer and
cellular telephone use can be excluded. Why did Ahlbom
and Feychting (35) change in their opinion about the Danish
study ? Although more long-term data were now available
and adjustment for socioeconomic factors could be made,
the update by Frei et al. (14) suffers from basically the same
methodological limitations mainly related to exposure
assessment as the fi rst one did. The effect of these limita-
tions may also as pointed out accumulate with follow-up.
Table 2 Estimated change in incidence rate/year ( % ) and 95 % CI
for brain and central nervous system tumors in Denmark 1970 2009
according to NORDCAN.
Men Women
Change in
incidence
rate/year, %
95 % CI Change in
incidence
rate/year, %
95 % CI
Denmark
1970 2009 1.3 1.1 to 1.5 1.9 1.7 to 2 .2
1970 – 1979 1.9 0.3 to 3.4 2.4 – 0.2 to 5.2
1980 – 1989 1.6 0.1 to 3.1 2.2 0.4 to 4.1
1990 – 1999 1.5 0.2 to 2.8 2.5 0.4 to 4.5
2000 – 2009 2.7 1.1 to 4.3 2.9 0.7 to 5.2
Calculations were based on incidence rates, age adjusted to the world
standard population, and rounded to two decimal places. Linear
regression analysis on the logarithm of the age-adjusted incidence
rates was used to calculate the trends.
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
Incidence per 100,000 person years
20
18
16
14
12
10
8
6
4
2
0
Year
Men, Denmark
Women, Denmark
Figure 1 Estimated incidence rates for brain and central nervous system tumors in men and women in Denmark 1970 2009 according to
NORDCAN. Age adjusted to the world standard population.
Authenticated | fredrik.soderqvist@orebroll.se
Download Date | 6/15/12 10:16 PM
S ö derqvist et al.: Mobile phones and risk of brain tumors 57
particularly high-grade glioma, has been seriously questioned
(38, 39) . Had the authors instead used data from NORDCAN
(40) on Denmark, they would have found a statistically signif-
icant increase in incidence for brain and central nervous sys-
tem tumors (combined) during 2000 2009 (Table 2
; Figure 1 ).
The latter is corroborated by a recent report from the Danish
National Board of Health (Sundhetsstyrelsen) with data
from the national cancer registry (Cancerregisteret), which
show an unexplained increase in brain and nervous system
tumors. Between 2001 and 2010, an increase in incidence was
seen for both genders, 40 % in men and 29 % in women (41) .
Lastly, noteworthy is that Ahlbom was excluded as expert in
the IARC meeting on RF-EMF in May 2011 due to potential
confl ict of interest (42) . No confl ict of interest was noted in
the editorial (35) .
Originators and initial fi nanciers The design and start of
the Danish cohort was initially made in cooperation between
International Epidemiology Institute (IEI), Rockville, MD,
USA, and the Danish Cancer Society. Two persons from IEI,
John D. Boice Jr. and Joseph K. Laughlin, were coauthors of
the two fi rst publications of the cohort (11, 12) , and the Danish
Cancer Society was represented by Christoffer Johansen and
J ø rgen H. Olsen, with some additional authors in the second
publication (12) . The cohort was established by grants from
two Danish telecom operation companies (TeleDenmark
Mobil and Sonafon), by IEI, and by the Danish Cancer
Society (11) . From where the IEI got its money for this study
has never been declared.
Interestingly, in the second publication (12) , it was reported
that only the Danish Strategic Research Council and the
Danish Cancer Society supported the cohort. Now, in the pub-
lication by Frei et al. (14) , the two contributors from IEI have
disappeared from the list of authors and only funding from
the Danish Strategic Research Council is acknowledged. The
article by Frei et al. (14) has been cleaned from any asso-
ciation with the telecom industry or IEI. In fact, regarding
competing interests, the authors state, no support from any
organization for the submitted work, no fi nancial relation-
ships with any organizations that might have an interest in the
submitted work in the three previous years, no other relation-
ships or activities that could appear to have infl uenced the
submitted work. Certainly, this is a truth with modifi cation.
By restricting to the last 3 years, all previous involvement by
the industry was green washed.
Conclusion
After reviewing the four publications on the Danish cohort
study, one might rightly wonder whether this cohort was
initially set up to show no increased risk. Sch ü z et al. (12)
concluded already in their fi rst update that, the narrow con-
dence intervals provide evidence that any large association
of risk of cancer and cellular telephone use can be excluded.
First, without long-term data, how can any large risk on
the end points in question be excluded ? Second, and more
important, the authors should know that however narrow the
CIs are, such a conclusion cannot be defended if the risk esti-
mates, as discussed above, are biased. In fact, in this context,
a biased study would be more misleading the longer the fol-
low-up: long-term data would accumulate and the CI would
become even narrower. This way, the publication by Frei
et al. (14) can be used in a seemingly solid argument against
an increased risk because this study, a large nationwide
cohort study, rules out not only non-response and recall bias
but also an increased risk as indicated by a tight CI. Although
previous case-control studies all have methodological limi-
tations that need to be carefully considered, type I errors are
not the only threats to the validity of studies in this fi eld of
research this Danish cohort study is a textbook example
of that.
References
1. Hardell L, Carlberg M, Hansson Mild K. Pooled analysis of case-
control studies on malignant brain tumors and the use of mobile
and cordless phones including living and deceased subjects. Int J
Oncol 2011;38:1465 – 74.
2. INTERPHONE Study Group. Brain tumour risk in relation to
mobile telephone use: results of the INTERPHONE international
case-control study. Int J Epidemiol 2010;39:675 94.
3. Cardis E, Armstrong BK, Bowman JD, Giles GG, Hours M,
et al. Risk of brain tumors in relation to estimated RF dose from
mobile phones: results from fi ve Interphone countries. Occup
Environ Med 2011;68:631 – 40.
4. INTERPHONE Study Group. Acoustic neuroma risk in rela-
tion to mobile telephone use: results of the INTERPHONE
international case-control study. Cancer Epidemiol 2011;35:
453 – 64.
5. Hardell L, Carlberg M, Hansson Mild K. Pooled analysis of two
case-control studies on the use of cellular and cordless telephones
and the risk of benign brain tumors diagnosed during 1997 2003.
Int J Oncol 2006;28:509 18.
6. Baan R, Grosse Y, Lauby-Secretan B, El Ghissassi F, Bouvard
V, et al.; WHO International Agency for Research on Cancer
Monograph Working Group. Carcinogenicity of radiofrequency
electromagnetic fi elds. Lancet Oncol 2011;12:624 6.
7. Sato Y, Akiba S, Kubo O, Yamaguchi N. A case-case study
of mobile phone use and acoustic neuroma risk in Japan.
Bioelectromagnetics 2011;32:85 – 93.
8. Muscat JE, Malkin MG, Thompson S, Shore RE, Stellman SD,
et al. Handheld cellular telephone use and risk of brain cancer. J
Am Med Assoc 2000;284:3001 7. Erratum in: J Am Med Assoc
2001;286:1293.
9. Inskip PD, Tarone RE, Hatch EE, Wilcosky TC, Shapiro WR,
et al. Cellular-telephone use and brain tumors. N Engl J Med
2001;344:79 – 86.
10. Auvinen A, Hietanen M, Luukkonen R, Koskela RS. Brain
tumors and salivary gland cancers among cellular telephone
users. Epidemiology 2002;13:356 – 9.
11. Johansen C, Boice J Jr, McLaughlin J, Olsen J. Cellular tele-
phones and cancer a nationwide cohort study in Denmark.
J Natl Cancer Inst 2001;93:203 7.
12. Sch ü z J, Jacobsen R, Olsen JH, Boice JD Jr, McLaughlin JK,
et al. Cellular telephone use and cancer risk: update of a nation-
wide Danish cohort. J Natl Cancer Inst 2006;98:1707 13.
Authenticated | fredrik.soderqvist@orebroll.se
Download Date | 6/15/12 10:16 PM
58 S ö derqvist et al.: Mobile phones and risk of brain tumors
13. Sch ü z J, Steding-Jessen M, Hansen S, Stangerup SE, Cay é -
Thomasen P, et al. Long-term mobile phone use and the risk of
vestibular schwannoma: a Danish nationwide cohort study. Am J
Epidemiol 2011;174:416 – 22.
14. Frei P, Poulsen AH, Johansen C, Olsen JH, Steding-Jessen M,
et al. Use of mobile phones and risk of brain tumors: update of
Danish cohort study. Brit Med J 2011;343:d6387.
15. Hardell L, Mild KH, Carlberg M, S ö derqvist F. Tumour risk
associated with use of cellular telephones or cordless desktop
telephones. World J Surg Oncol 2006;4:74.
16. Redmayne M, Inyang I, Dimitriadis C, Benke G, Abramson MJ.
Cordless telephone use: implications for mobile phone research.
J Environ Monit 2010;12:809 12.
17. Christensen HC, Sch ü z J, Kosteljanetz M, Poulsen HS, Boice
JD Jr, et al. Cellular telephones and risk for brain tumors:
a population-based, incident case-control study. Neurology
2005;64:1189 95. Erratum in: Neurology 2005;65:1324.
18. Ahlbom A, Feychting M, Cardis E, Elliott P. Re: Cellular tele-
phone use and cancer risk: update of a nationwide Danish
cohort study. J Natl Cancer Inst 2007;99:655; author reply
655 – 6.
19. Hardell L, Carlberg M, Hansson Mild K. Pooled analysis of two
case-control studies on use of cellular and cordless telephones
and the risk for malignant brain tumors diagnosed in 1997 2003.
Int Arch Occup Environ Health 2006;79:630 9.
20. Sch ü z J, Elliott P, Auvinen A, Kromhout H, Poulsen AH, et al. An
international prospective cohort study of mobile phone users and
health (Cosmos): design considerations and enrolment. Cancer
Epidemiol 2011;35:37 – 43.
21. Kranzinger M, Jones N, Rittinger O, Pilz P, Piotrowski WP,
et al. Malignant glioma as a secondary malignant neoplasm after
radiation therapy for craniopharyngioma: report of a case and
review of reported cases. Onkologie 2001;24:66 72.
22. Sadetzki S, Chetrit A, Freedman L, Stovall M, Modan B, et
al. Long-term follow-up for brain tumor development after
childhood exposure to ionizing radiation for tinea capitis.
Radiat Res 2005;163:424 32. Erratum in: Radiat Res 2005;
164:234.
23. Umansky F, Shoshan Y, Rosenthal G, Fraifeld S, Spektor S.
Radiation-induced meningioma. Neurosurg Focus 2008;24:E7.
Review.
24. Mohyuddin A, Vokurka EA, Evans DG, Ramsden RT, Jackson
A. Is clinical growth index a reliable predictor of tumour
growth in vestibular schwannomas ? Clin Otolaryngol Allied Sci
2003;28:85 – 90.
25. Kundi M. Mobile phone use and cancer. Occup Environ Med
2004;61:560 – 70, 487.
26. Kundi M. Essential problems in the interpretation of epidemio-
logic evidence for an association between mobile phone use and
brain tumors. CR Physique 2010;11:556 63.
27. Hardell L, Hallquist A, Mild KH, Carlberg M, P å hlson A, et al.
Cellular and cordless telephones and the risk for brain tumors.
Eur J Cancer Prev 2002;11:377 86.
28. Hardell L, Carlberg M, Hansson Mild K. Re-analysis of risk for
glioma in relation to mobile telephone use: comparison with the
results of the Interphone international case-control study. Int J
Epidemiol 2011;40:1126 – 8.
29. Myung SK, Ju W, McDonnell DD, Lee YJ, Kazinets G, et al.
Mobile phone use and risk of tumors: a meta-analysis. J Clin
Oncol 2009;27:5565 – 72.
30. Vrijheid M, Armstrong BK, B é dard D, Brown J, Deltour I, et al.
Recall bias in the assessment of exposure to mobile phones. J
Expo Sci Environ Epidemiol 2009;19:369 81.
31. Vrijheid M, Deltour I, Krewski D, Sanchez M, Cardis E. The
effects of recall errors and of selection bias in epidemiologic
studies of mobile phone use and cancer risk. J Expo Sci Environ
Epidemiol 2006;16:371 – 84.
32. Vrijheid M, Cardis E, Armstrong BK, Auvinen A, Berg G,
et al.; Interphone Study Group. Validation of short term recall of
mobile phone use for the Interphone study. Occup Environ Med
2006;63:237 – 43.
33. Dalton SO, Steding-Jessen M, Gislum M, Frederiksen K,
Engholm G, et al. Social inequality and incidence of and survival
from cancer in a population-based study in Denmark, 1994
2003: background, aims, material and methods. Eur J Cancer
2008;44:1938 – 49.
34. Hardell L, Carlberg M. Mobile phones, cordless phones and the
risk for brain tumors. Int J Oncol 2009;35:5 17.
35. Ahlbom A, Feychting M. Mobile telephones and brain tumors.
Brit Med J 2011;343:d6605.
36. Hardell L, Carlberg M, Mild KH. Case-control study of the asso-
ciation between the use of cellular and cordless telephones and
malignant brain tumors diagnosed during 2000 2003. Environ
Res 2006;100:232 – 41.
37. Hardell L, Carlberg M, Hansson Mild K. Case-control study
on cellular and cordless telephones and the risk for acous-
tic neuroma or meningioma in patients diagnosed 2000 2003.
Neuroepidemiology 2005;25:120 – 8.
38. Barlow L, Westergren K, Holmberg L, Talback M. The complete-
ness of the Swedish Cancer Register: a sample survey for year
1998. Acta Oncol 2009;48:27 – 33.
39. Bergenheim T, Malmstrom A, Bolander H, Michanek A,
Stragliotto G, et al. [Registration on regional basis of patients
with primary brain tumors. Regional differences disclosed].
Lakartidningen 2007;104:332 – 41.
40. NORDCAN. Available at: http://www-dep.iarc.fr/NORDCAN/
english/frame.asp. Accessed 8 December 2011.
41. Sundhedsstyrelsen (National Board of Health Denmark). Tal og
Analyse: Cancerregisteret 2010. Available at: http://www.sst.
dk/publ/Publ2011/DAF/Cancer/Cancerregisteret2010.pdf. Last
accessed 16 March 2012.
42. International Agency for Research on Cancer (IARC). IARC
Monographs on the Evaluation of Carcinogenic Risks to Humans
list of participants. Available at: http://monographs.iarc.fr/
ENG/Meetings/vol102-participants.pdf. Last accessed 16 March
2012.
Authenticated | fredrik.soderqvist@orebroll.se
Download Date | 6/15/12 10:16 PM
... Several statements by ICNIRP are made without any scientific references. On the other hand, the Danish cohort study on mobile phone use (59) is included, in spite of the fact that it was judged by IARC (1,2), as well as in our review (60), to be uninformative. A biased article written by authors including ICNIRP members, used to 'prove' the no-risk paradigm for RF radiation carcinogenesis (23), is cited by ICNIRP. ...
... Notably, this cohort study is now included in the risk calculations, although Martin Röösli was a member of the IARC evaluation group and should have been aware of the IARC decision. The numerous shortcomings in the Danish cohort study, discussed in detail in a peer-reviewed article (60), are omitted in the article by Röösli et al (23). ...
Article
Full-text available
The fifth generation, 5G, of radiofrequency (RF) radiation is about to be implemented globally without investigating the risks to human health and the environment. This has created debate among concerned individuals in numerous countries. In an appeal to the European Union (EU) in September 2017, currently endorsed by >390 scientists and medical doctors, a moratorium on 5G deployment was requested until proper scientific evaluation of potential negative consequences has been conducted. This request has not been acknowledged by the EU. The evaluation of RF radiation health risks from 5G technology is ignored in a report by a government expert group in Switzerland and a recent publication from The International Commission on Non-Ionizing Radiation Protection. Conflicts of interest and ties to the industry seem to have contributed to the biased reports. The lack of proper unbiased risk evaluation of the 5G technology places populations at risk. Furthermore, there seems to be a cartel of individuals monopolizing evaluation committees, thus reinforcing the no-risk paradigm. We believe that this activity should qualify as scientific misconduct. © This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) License.
... Since the IARC evaluation in 2011, more human epidemiological studies have been published that support a causal association between RF radiation and brain and head tumors. A Danish cohort study on 'mobile phone users' (35,36) is not included herein due to serious methodological shortcomings in the study design [see Söderqvist et al (37)]. The study by Benson et al (38) is of limited value since the use of cordless phones was not included, mobile phone use was assessed only at baseline and no information on tumor laterality, including ipsilateral versus contralateral use was given. ...
... The same cohort has also been used for studying brain tumor risk. Due to serious methodological problems, including the misclassification of exposure the study has been evaluated to be uninformative (8,37). ...
Article
Full-text available
During the use of handheld mobile and cordless phones, the brain is the main target of radiofrequency (RF) radiation. An increased risk of developing glioma and acoustic neuroma has been found in human epidemiological studies. Primarily based on these findings, the International Agency for Research on Cancer (IARC) at the World Health Organization (WHO) classified in May, 2011 RF radiation at the frequency range of 30 kHz‑300 GHz as a 'possible' human carcinogen, Group 2B. A carcinogenic potential for RF radiation in animal studies was already published in 1982. This has been confirmed over the years, more recently in the Ramazzini Institute rat study. An increased incidence of glioma in the brain and malignant schwannoma in the heart was found in the US National Toxicology Program (NTP) study on rats and mice. The NTP final report is to be published; however, the extended reports are published on the internet for evaluation and are reviewed herein in more detail in relation to human epidemiological studies. Thus, the main aim of this study was to compare earlier human epidemiological studies with NTP findings, including a short review of animal studies. We conclude that there is clear evidence that RF radiation is a human carcinogen, causing glioma and vestibular schwannoma (acoustic neuroma). There is some evidence of an increased risk of developing thyroid cancer, and clear evidence that RF radiation is a multi‑site carcinogen. Based on the Preamble to the IARC Monographs, RF radiation should be classified as carcinogenic to humans, Group 1.
... shortcomings in the study were discussed in a peer-reviewed article [100] concluding that: "After reviewing the four publications on the Danish cohort study, one might rightly wonder whether this cohort was initially set up to show no increased risk." ...
... Unfortunately, neither study was originally designed to assess the effects of mobile phone use; hence, both had poor measures of cellphone use. Plus, the Danish study classified as non-cellphone users many heavy cellphone users who had cellphones purchased by their businesses which reduced risk estimates below 1 [23]. Although the Million Women Study did not find an association between cellphone use and glioma risk, the study found that those who used cellphones ≥10 years (RR = 2.46, 95% CI = 1.07-5.64) ...
Article
Full-text available
We appreciate Frank de Vocht and Martin Röösli’s interest [1] in our study and their recognition of the importance of conducting regular, systematic syntheses of the epidemiological evidence regarding the tumor risk from mobile phone (i.e., cellphone) use [...]
... They concluded that overall, epidemiological evidence does not suggest increased brain or salivary gland tumor risk with mobile phone (MP) use, although the authors admitted that some uncertainty remains regarding long latency periods (>15 years), rare brain tumor subtypes, and MP usage during childhood. Of concern is that these analyses included cohort studies with poor exposure classification (20). In epidemiological studies, recall bias can play a substantial role in the attenuation of odds ratios toward the null hypothesis. ...
Article
Full-text available
Radiation exposure has long been a concern for the public, policy makers, and health researchers. Beginning with radar during World War II, human exposure to radio-frequency radiation 1 (RFR) technologies has grown substantially over time. In 2011, the International Agency for Research on Cancer (IARC) reviewed the published literature and categorized RFR as a "possible" (Group 2B) human carcinogen. A broad range of adverse human health effects associated with RFR have been reported since the IARC review. In addition, three large-scale carcinogenicity studies in rodents exposed to levels of RFR that mimic lifetime human exposures have shown significantly increased rates of Schwannomas and malignant gliomas, as well as chromosomal DNA damage. Of particular concern are the effects of RFR exposure on the developing brain in children. Compared with an adult male, a cell phone held against the head of a child exposes deeper brain structures to greater radiation doses per unit volume, and the young, thin skull's bone marrow absorbs a roughly 10-fold higher local dose. Experimental and observational studies also suggest that men who keep cell phones in their trouser pockets have significantly lower sperm counts and significantly impaired sperm motility and morphology, including mitochondrial DNA damage. Based on the accumulated evidence, we recommend that IARC re-evaluate its 2011 classification of the human carcinogenicity of RFR, and that WHO complete a systematic review of multiple other health effects such as sperm damage. In the interim, current knowledge provides justification for governments, public health authorities, and physicians/allied health professionals to warn the population that having a cell phone next to the body is harmful, and to support measures to reduce all exposures to RFR.
... This has in part been based on a much criticized Danish cohort study on persons with mobile phone subscriptions and assumed mobile phone use with funding from the telecom industry [67]. The study was not based on sound epidemiological principles and had several methodological limitations mainly due to poor exposure assessment that render it to be uninformative at best [68]. Some of the many shortcomings include the following. ...
Article
Full-text available
Objective. Bradford Hill’s viewpoints from 1965 on association or causation were used on glioma risk and use of mobile or cordless phones. Methods. All nine viewpoints were evaluated based on epidemiology and laboratory studies. Results. Strength: meta-analysis of case-control studies gave odds ratio (OR) = 1.90, 95% confidence interval (CI) = 1.31–2.76 with highest cumulative exposure. Consistency: the risk increased with latency, meta-analysis gave in the 10+ years’ latency group OR = 1.62, 95% CI = 1.20–2.19. Specificity: increased risk for glioma was in the temporal lobe. Using meningioma cases as comparison group still increased the risk. Temporality: highest risk was in the 20+ years’ latency group, OR = 2.01, 95% CI =1.41–2.88, for wireless phones. Biological gradient: cumulative use of wireless phones increased the risk. Plausibility: animal studies showed an increased incidence of glioma and malignant schwannoma in rats exposed to radiofrequency (RF) radiation. There is increased production of reactive oxygen species (ROS) from RF radiation. Coherence: there is a change in the natural history of glioma and increasing incidence. Experiment: antioxidants reduced ROS production from RF radiation. Analogy: there is an increased risk in subjects exposed to extremely low-frequency electromagnetic fields. Conclusion. RF radiation should be regarded as a human carcinogen causing glioma.
... In a record linkage study from Denmark, mobile phone subscribers from January 1, 1982, until December 31, 1995, were identified from the computerised files of the 2 Danish operating companies, TeleDenmark Mobil and Sonofon, which also partly funded the study [9]. It has produced four articles that we thoroughly reviewed [11]. We concluded that its many limitations -embedded in the study design from the very beginning and mainly related to poor exposure assessment -obscure the findings of the four reports to such an extent that, at best, render them uninformative. ...
Article
Full-text available
We made a pooled analysis of two case-control studies on malignant brain tumours with patients diagnosed during 1997-2003 and 2007-2009. They were aged 20-80 years and 18-75 years, respectively, at the time of diagnosis. Only cases with histopathological verification of the tumour were included. Population-based controls, matched on age and gender, were used. Exposures were assessed by questionnaire. The whole reference group was used in the unconditional regression analysis adjusted for gender, age, year of diagnosis, and socio-economic index. In total, 1498 (89%) cases and 3530 (87%) controls participated. Mobile phone use increased the risk of glioma, OR=1.3, 95% CI=1.1-1.6 overall, increasing to OR=3.0, 95% CI=1.7-5.2 in the >25 year latency group. Use of cordless phones increased the risk to OR=1.4, 95% CI=1.1-1.7, with highest risk in the >15-20 years latency group yielding OR=1.7, 95% CI=1.1-2.5. The OR increased statistically significant both per 100h of cumulative use, and per year of latency for mobile and cordless phone use. Highest ORs overall were found for ipsilateral mobile or cordless phone use, OR=1.8, 95% CI=1.4-2.2 and OR=1.7, 95% CI=1.3-2.1, respectively. The highest risk was found for glioma in the temporal lobe. First use of mobile or cordless phone before the age of 20 gave higher OR for glioma than in later age groups. Copyright © 2014 Elsevier Ireland Ltd. All rights reserved.
Article
The MOBI-Kids case-control study on wireless phone use and brain tumor risk in childhood and adolescence included the age group 10–24 years diagnosed between 2010 and 2015. Overall no increased risk was found although for brain tumors in the temporal region an increased risk was found in the age groups 10–14 and 20–24 years. Most odds ratios (ORs) in MOBI-Kids were <1.0, some statistically significant, suggestive of a preventive effect from RF radiation; however, this is in contrast to current knowledge about radiofrequency (RF) carcinogenesis. The MOBI-Kids results are not biologically plausible and indicate that the study was flawed due to methodological problems. For example, not all brain tumor cases were included since central localization was excluded. Instead, all brain tumor cases should have been included regardless of histopathology and anatomical localization. Only surgical controls with appendicitis were used instead of population-based controls from the same geographical area as for the cases. In fact, increased incidence of appendicitis has been postulated to be associated with RF radiation which makes selection of control group in MOBI-Kids questionable. Start of wireless phone use up to 10 years before diagnosis was in some analyses included in the unexposed group. Thus, any important results demonstrating late carcinogenesis, a promoter effect, have been omitted from analysis and may underestimate true risks. Linear trend was in some analyses statistically significant in the calculation of RF-specific energy and extremely low frequency (ELF)-induced current in the center of gravity of the tumor. Additional case-case analysis should have been performed. The data from this study should be reanalyzed using unconditional regression analysis adjusted for potential confounding factors to increase statistical power. Then all responding cases and controls could be included in the analyses. In sum, we believe the results as reported in this paper seem uninterpretable and should be dismissed.
Article
Currently the fifth generation, 5G, for wireless communication is about to be rolled out worldwide. Many persons are concerned about potential health risks from radiofrequency radiation. In September 2017, a letter was sent to the European Union asking for a moratorium on the deployment until scientific evaluation has been made on potential health risks (http://www.5Gappeal.eu). This appeal has had little success. The Health Council of the Netherlands released on September 2, 2020 their evaluation on 5G and health. It was largely based on a World Health Organization draft and report by the Swedish Radiation Safety Authority, both criticized for not being impartial. The guidelines by the International Commission on Non-Ionizing Radiation Protection were recommended to be used, although they have been considered to be insufficient to protect against health hazards (http://www.emfscientist.org). The Health Council Committee recommended not to use the 26 GHz frequency band until health risks have been studied. For lower frequencies, the International Commission on Non-Ionizing Radiation Protection guidelines were recommended. The conclusion that there is no reason to stop the use of lower frequencies for 5G is not justified by current evidence on cancer risks as commented in this article. A moratorium is urgently needed on the implementation of 5G for wireless communication.
Article
Emerging studies have provided evidence on the carcinogenicity of radiofrequency radiation (RFR) from cell phones. This study aims to test the genetic susceptibility on the association between cell phone use and thyroid cancer. Population-based case-control study was conducted in Connecticut between 2010 and 2011 including 440 thyroid cancer cases and 465 population-based controls with genotyping information for 823 single nucleotide polymorphisms (SNPs) in 176 DNA genes. We used multivariate unconditional logistic regression models to estimate the genotype-environment interaction between each SNP and cell phone use and to estimate the association with cell phone use in populations according to SNP variants. Ten SNPs had P < 0.01 for interaction in all thyroid cancers. In the common homozygote groups, no association with cell phone use was observed. In the variant group (heterozygotes and rare homozygotes), cell phone use was associated with an increased risk for rs11070256 (odds ratio (OR): 2.36, 95% confidence interval (CI): 1.30-4.30), rs1695147 (OR: 2.52, 95% CI: 1.30-4.90), rs6732673 (OR: 1.59, 95% CI: 1.01-2.49), rs396746 (OR: 2.53, 95% CI: 1.13-5.65), rs12204529 (OR: 2.62, 95% CI: 1.33-5.17), and rs3800537 (OR: 2.64, 95% CI: 1.30-5.36) with thyroid cancers. In small tumors, increased risk was observed for 5 SNPs (rs1063639, rs1695147, rs11070256, rs12204529 and rs3800537), In large tumors, increased risk was observed for 3 SNPs (rs11070256, rs1695147, and rs396746). Our result suggests that genetic susceptibilities modify the associations between cell phone use and risk of thyroid cancer. The findings provide more evidence for RFR carcinogenic group classification.
Article
Full-text available
To investigate the risk of tumours in the central nervous system among Danish mobile phone subscribers. Nationwide cohort study. Denmark. All Danes aged ≥ 30 and born in Denmark after 1925, subdivided into subscribers and non-subscribers of mobile phones before 1995. Risk of tumours of the central nervous system, identified from the complete Danish Cancer Register. Sex specific incidence rate ratios estimated with log linear Poisson regression models adjusted for age, calendar period, education, and disposable income. 358,403 subscription holders accrued 3.8 million person years. In the follow-up period 1990-2007, there were 10,729 cases of tumours of the central nervous system. The risk of such tumours was close to unity for both men and women. When restricted to individuals with the longest mobile phone use--that is, ≥ 13 years of subscription--the incidence rate ratio was 1.03 (95% confidence interval 0.83 to 1.27) in men and 0.91 (0.41 to 2.04) in women. Among those with subscriptions of ≥ 10 years, ratios were 1.04 (0.85 to 1.26) in men and 1.04 (0.56 to 1.95) in women for glioma and 0.90 (0.57 to 1.42) in men and 0.93 (0.46 to 1.87) in women for meningioma. There was no indication of dose-response relation either by years since first subscription for a mobile phone or by anatomical location of the tumour--that is, in regions of the brain closest to where the handset is usually held to the head. In this update of a large nationwide cohort study of mobile phone use, there were no increased risks of tumours of the central nervous system, providing little evidence for a causal association.
Article
Full-text available
Vestibular schwannomas grow in the region within the brain where most of the energy by radiofrequency electromagnetic fields from using mobile phones is absorbed. The authors used 2 Danish nationwide cohort studies, one a study of all adult Danes subscribing for a mobile phone in 1995 or earlier and one on sociodemographic factors and cancer risk, and followed subjects included in both cohorts for occurrence of vestibular schwannoma up to 2006 inclusively. In this study including 2.9 million subjects, a long-term mobile phone subscription of ≥11 years was not related to an increased vestibular schwannoma risk in men (relative risk estimate = 0.87, 95% confidence interval: 0.52, 1.46), and no vestibular schwannoma cases among long-term subscribers occurred in women versus 1.6 expected. Vestibular schwannomas did not occur more often on the right side of the head, although the majority of Danes reported holding their mobile phone to the right ear. Vestibular schwannomas in long-term male subscribers were not of larger size than expected. Overall, no evidence was found that mobile phone use is related to the risk of vestibular schwannoma. Because of the usually slow growth of vestibular schwannoma and possible diagnostic delay, further surveillance is indicated.
Article
Full-text available
The objective of this study was to examine the associations of brain tumours with radio frequency (RF) fields from mobile phones. Patients with brain tumour from the Australian, Canadian, French, Israeli and New Zealand components of the Interphone Study, whose tumours were localised by neuroradiologists, were analysed. Controls were matched on age, sex and region and allocated the 'tumour location' of their matched case. Analyses included 553 glioma and 676 meningioma cases and 1762 and 1911 controls, respectively. RF dose was estimated as total cumulative specific energy (TCSE; J/kg) absorbed at the tumour's estimated centre taking into account multiple RF exposure determinants. ORs with ever having been a regular mobile phone user were 0.93 (95% CI 0.73 to 1.18) for glioma and 0.80 (95% CI 0.66 to 0.96) for meningioma. ORs for glioma were below 1 in the first four quintiles of TCSE but above 1 in the highest quintile, 1.35 (95% CI 0.96 to 1.90). The OR increased with increasing TCSE 7+ years before diagnosis (p-trend 0.01; OR 1.91, 95% CI 1.05 to 3.47 in the highest quintile). A complementary analysis in which 44 glioma and 135 meningioma cases in the most exposed area of the brain were compared with gliomas and meningiomas located elsewhere in the brain showed increased ORs for tumours in the most exposed part of the brain in those with 10+ years of mobile phone use (OR 2.80, 95% CI 1.13 to 6.94 for glioma). Patterns for meningioma were similar, but ORs were lower, many below 1.0. There were suggestions of an increased risk of glioma in long-term mobile phone users with high RF exposure and of similar, but apparently much smaller, increases in meningioma risk. The uncertainty of these results requires that they be replicated before a causal interpretation can be made.
Article
Full-text available
We studied the association between use of mobile and cordless phones and malignant brain tumours. Pooled analysis was performed of two case-control studies on patients with malignant brain tumours diagnosed during 1997-2003 and matched controls alive at the time of study inclusion and one case-control study on deceased patients and controls diagnosed during the same time period. Cases and controls or relatives to deceased subjects were interviewed using a structured questionnaire. Replies were obtained for 1,251 (85%) cases and 2,438 (84%) controls. The risk increased with latency period and cumulative use in hours for both mobile and cordless phones. Highest risk was found for the most common type of glioma, astrocytoma, yielding in the >10 year latency group for mobile phone use odds ratio (OR) = 2.7, 95% confidence interval (CI) = 1.9-3.7 and cordless phone use OR = 1.8, 95% CI = 1.2-2.9. In a separate analysis, these phone types were independent risk factors for glioma. The risk for astrocytoma was highest in the group with first use of a wireless phone before the age of 20; mobile phone use OR = 4.9, 95% CI = 2.2-11, cordless phone use OR = 3.9, 95% CI = 1.7-8.7. In conclusion, an increased risk was found for glioma and use of mobile or cordless phone. The risk increased with latency time and cumulative use in hours and was highest in subjects with first use before the age of 20.
Article
Due to the close proximity of a mobile phone to the head when placing a call, concerns have been raised that exposure from microwaves during mobile phone use may exert adverse health effects and, in particular, may increase the risk of brain tumours. In response to these concerns epidemiological studies have been conducted, most applying the case-control design. While epidemiology can provide decisive evidence for an association between an exposure and a disease fundamental problems arise if exposure is short compared to the natural history of the disease. For brain tumours latencies of decades have been implicated making special considerations about potential effects of exposures necessary that commence during an already growing tumour. It is shown that measures of disease risk like odds ratios and relative risks can under such circumstances not be interpreted as indicators of a long term effect on incidences in the exposed population. Besides this problem, the issues of a suitable exposure metric and the selection of endpoints are unresolved. It is shown that the solution of these problems affords knowledge about the mechanism of action by which exposure increases the risk of manifest disease.
Article
Methods An interview-based case-control study with 2708 glioma and 2409 meningioma cases and matched controls was conducted in 13 countries using a common protocol. Results A reduced odds ratio (OR) related to ever having been a regular mobile phone user was seen for glioma [OR 0.81; 95% confidence interval (CI) 0.70-0.94] and meningioma (OR 0.79; 95% CI 0.68-0.91), possibly reflecting participation bias or other methodological limitations. No elevated OR was observed >= 10 years after first phone use (glioma: OR 0.98; 95% CI 0.76-1.26; meningioma: OR 0.83; 95% CI 0.61-1.14). ORs were = 1640 h, the OR was 1.40 (95% CI 1.03-1.89) for glioma, and 1.15 (95% CI 0.81-1.62) for meningioma; but there are implausible values of reported use in this group. ORs for glioma tended to be greater in the temporal lobe than in other lobes of the brain, but the CIs around the lobe-specific estimates were wide. ORs for glioma tended to be greater in subjects who reported usual phone use on the same side of the head as their tumour than on the opposite side. Conclusions Overall, no increase in risk of glioma or meningioma was observed with use of mobile phones. There were suggestions of an increased risk of glioma at the highest exposure levels, but biases and error prevent a causal interpretation. The possible effects of long-term heavy use of mobile phones require further investigation.