Printable version (PDF format)

In March this year Khurana, Teo, Kundi, Hardell & Carlberg published a literature review that investigated possible associations of long term mobile phone use and brain tumours (Khurana et al., Surgical Neurology, March 26, 2009). The paper attracted considerable media interest here in Australia, where the issue was broadcast on such programs as Channel 9’s “60 Minutes”, and ABC’s “Lateline”. Given the wide publicity, it is possible that that the use of this paper in the media may have misled audiences about this issue, thus the ACRBR has put together these brief comments on the paper.

Summary of Khurana et al (2009)

The paper reports on a number of issues pertaining to the debate on whether mobile phones cause brain tumours. These include:

• A meta-analysis of the epidemiological literature relating to brain tumours and mobile phone use of > 10 years, where it is reported that there is an overall association between mobile phone use and glioma, but not acoustic neuroma nor meningioma;
• A meta-analysis of the epidemiological literature relating to brain tumours and mobile phone use of > 10 years on the same side of the head as the tumours, where it is reported that such use is associated with increased risk of glioma and acoustic neuroma, but not meningioma;
• A report of the Central Brain Tumor Registry of the United States (CBTRUS) incidence rates from some regions of the USA, where Khurana et al report that, although there are issues to do with comparing recent and older brain tumour incidence rates, they think there might be an increase;
• A brief review of the BioInitiative Report, where it is claimed that the Report has shown many negative health effects of low level EMF exposure; and
• A brief review of the current literature relating to experimental research (i.e. research more able to demonstrate causation) that has addressed the issue of low level EMF and health, arguing that it shows there are negative health consequences of mobile phone exposure

1/ A meta-analysis of the epidemiological literature on associations of brain tumours and mobile phone use of ≥ 10 years

Although the focus of the paper relates to tumours that are reported to be on the same side of the head (ipsilateral) as patients used their mobile phones, Khurana et al. also conduct a meta-analysis regardless of tumour laterality. Here they report that there is an overall association between mobile phone use and glioma, but not acoustic neuroma nor meningioma, for those who used a mobile phone for ≥ 10 years. However the methods section suggests that this analysis is not appropriate, as they have excluded research that did not look at laterality, which makes the data incomplete. Further, there are few details of how the meta-analysis was conducted. This is particularly important as such meta-analyses generally require the assumption of homogeneity to be met in order for the results to be meaningful (and no test of homogeneity is provided).

2/ A meta-analysis of the epidemiological literature on associations of brain tumours and mobile phone use of ≥ 10 years on the same side of the head as the tumours

This is the focus of the paper, where it is reported that there is an increased risk of glioma and acoustic neuroma (but not meningioma) for those who reported using a mobile phone for ≥ 10 years on the same side of the head as the tumour. As the literature has convincingly demonstrated that there is no such association for shorter latencies, this would indeed be an important finding. However, this analysis suffers from the same lack of detail as described above, and so no conclusions can be drawn from this analysis. It may also be noted that as a meta-analysis, this does not present new data, but rather a different analysis of data that has already been reported. This should have been able to increase the number of cases in the analysis in order to provide a more precise measure of association. This could have been very useful as the number of cases in each study who reported using a mobile phone on the same side of the head as their brain tumour is very small, again making individual studies inconclusive.

3/ A report of CBTRUS brain tumour incidence rates from the USA

Given that the use of mobile telephony has increased dramatically since the 1980s, if mobile phones were indeed causing brain tumours then one would expect that brain tumour incidence rates would also be increasing. So far such increases in incidence rates have not been found. Khurana et al consider a data set concerning this from a sample of the USA population (CBTRUS 2007-2008 Report), acknowledge that the data does not suggest there is such an increase in malignant brain tumours, but then go on to say that they still believe there is an increase in benign brain tumour incidence rates.

This latter statement is particularly important because it was taken by the media to support the hypothesis that mobile phones cause cancer. However, it is important to note that these figures are only from a relatively small sample, and are not consistent with the figures released from the substantially larger International Agency for Research on Cancer (IARC) World Cancer Report 2008. That is, IARC estimate that there is no actual increase in cancers, but rather that due to recent improvements in diagnostic techniques that slightly more cancers (in the vicinity of 1% annually) are detected. The technique that has resulted in the detection of more brain tumours is magnetic resonance imaging (MRI).

Another issue that may relate to the increase in the CBTRUS incidence rates is that a new law came into force in 2004 in the USA (US Public Law 107-260). This law resulted in increased surveillance of benign tumours, and may explain why the CBTRUS figures do not represent international trends. The ACRBR thus agrees with IARC that current data are “not showing any excess risk of brain cancers and other neoplasms associated with the use of mobile phones”.

4/ A brief review of the BioInitiative Report

Khurana et al provide a brief review of the BioInitiative Report, which is a web-based report that does not follow normal scientific peer-review procedures² and claims that there is substantial evidence showing mobile phones are harmful. If this was an accurate report of the state of science, then this would indeed suggest that mobile phones are harmful.

However Khurana et al. do not provide a balanced appraisal of the BioInitiative Report, a report which is not consistent with the current state of science as espoused by leading independent expert committees (such as the World Health Organisation). A more detailed critique of that report can be found on the ACRBR FAQ website. The interested reader may also wish to consult other independent critiques of the report, such as by the Health Council of the Netherlands.

5/ A brief review of the current literature pertaining to experimental research (i.e. research more able to demonstrate causation) that has addressed the issue of low level EMF and health

Khurana et al. argue that this literature shows there are negative health consequences of mobile phone exposure. However, as per the BioInitiative Report (see above), this does not provide a balanced appraisal of the literature, and reaches conclusions that contradict those of major international expert committees. The current state of science is that there have not been any consistent demonstrations of health effects resulting from mobile phone-like exposures (see the World Health Organisation for more information on this).

Part of the difficulty may arise from the authors’ failure to appreciate science as a body of research, where instead they have focused on particular studies that reported health effects without regard for relevant conflicting data. This strategy is unusual given that the authors acknowledge at one point that there have not been consistent findings of harm (i.e. “although the literature is inconsistent in terms of experimental reproducibility”, page 7), but then they ignore this variability and conclude that reported effects are real, regardless of whether other studies have found the opposite.

This is particularly important as science operates in a probabilistic fashion, with all results referred to as relatively likely or unlikely. This method is typically set up so that 1 in 20 results that are reported will appear to be real, but may actually be due to chance alone. Thus without considering the other research that fails to find an effect, it is a little like concluding that you are magic because you can toss a coin 4 times and get heads each time (i.e. it would not seem so remarkable if one considered the other 15 experiments where you tried to do this but failed).

Conclusions

In summary, the ACRBR believes it is important to recognize that Khurana et al. do not present any new data that is relevant to the mobile phone health debate. The preexisting data they consider are not synthesised in a meaningful fashion. Many of the conclusions made in the paper contradict those made by international expert committees, without providing adequate reasons for rejecting the standard view. On the contrary, we believe that the standard view of science, which is that there is currently no evidence that mobile phones have any negative health effects (as espoused by such groups as the World Health Organisation; see above), is an accurate reflection of the literature to date. __________________________

1 Note that ‘benign’ and ‘malignant’ are not always applicable to the particular brain tumours reported in this literature. For example, acoustic neuromas are typically referred to as benign, but can also have fatal consequences. For more information on such distinctions, please visit http://brain.mgh.harvard.edu/PatientGuide.htm.

2 It should be noted that some BioInitiative Report authors have also released individual papers that have followed normal peer review processes (for example, Carpenter & Sage, Rev Environ Health 2008, 23(2):91-117). As there are commonalities between aspects of the BioInitiative Report and these papers, the interested reader may wish to consult the latter for a peer-reviewed discussion of some of the issues addressed in the present Position Statement.

Printable version (PDF format)

In 2007 a group of interested individuals collated a series of views on the non-ionising radiation health debate. This was entitled the BioInitiative Report1, a web document dated August 31, 2007. The BioInitiative Report presents a series of views that argue for a change in public exposure standards, but which are largely inconsistent with current scientific consensus. The ACRBR have received numerous queries about this report from the general public, and have provided this document to answer a few questions to clarify its perspective on the report.

Do the BioInitiative Report authors represent an authoritative international body?

Often in assessing public health issues, bodies are formed to evaluate evidence and offer recommendations about particular issues. The model that most scientific expert bodies in this area (e.g. World Health Organisation (WHO)) employ is to engage independent experts to provide a review and recommendations on an issue. Independent experts are engaged because it is meant to provide an objective evaluation of the issue. This contrasts strongly with the BioInitiative Report, which is the result of the opinions of a self-selected group of individuals who each have a strong belief that does not accord with that of current scientific consensus. An indication of this may be seen in the group’s stated purpose, which is “to document the reasons why current public exposure standards for non-ionizing electromagnetic radiation are no longer good enough to protect public health” (Section 2, page 1), rather than to provide a scientific evaluation of the issue. Similarly, the standard model normally seeks a consensus view. In terms of the BioInitiative Report, the preface by Carpenter and Sage state that this is not a consensus document, but is rather a collection of individual views, where “the information and conclusions in each chapter are the responsibilities of the authors of that chapter” (Section i, page 1). Thus the ‘Summary for the Public and Conclusions’, released both independently and as part of the full Report, should be read as Sage’s view on the matter, and there is no indication in the Report that the authors of other chapters share her views. This does not mean that what is written in the Report is invalid, but it means that we need to evaluate the content of the report itself, and cannot rely on there being a consensus from an independent authoritative body to help us judge the merits of these conclusions.

What is the scientific status of the BioInitiative Report?

In science we generally differentiate between peer-reviewed and non-peer-reviewed publications, where the peer-review comes from independent experts in the area. The reason for this is that peer-reviewed work is only published after independent scientific peers have reviewed the work and agreed with its scientific merit, making it easier for the reader to be confident with conclusions drawn in the publication. Conversely, without independent peer review, there is far less opportunity to correct errors and ensure that the conclusions are appropriate, and thus scientists treat peer-reviewed publications as their main scientific literature source. It should be noted that this does not mean that publications lacking independent peer review are flawed (or for that matter that peer-reviewed publications are perfect), it is more that scientists would typically withhold judgment about publications until peer review has occurred.

The BioInitiative Report has not undergone such independent peer review, and so the conclusions that it reaches would normally be viewed more as views of some of the authors, rather than strong contributions to science. In fact the Report does not identify the level of review that it has undergone, merely mentioning that “another dozen outside reviewers have looked at and refined the Report” (Section 1, page 4). This is particularly important since many of the statements and conclusions in the Report are contrary to scientific consensus. Thus rigorous scientific evaluation would need to be performed to determine whether the inconsistencies are due to errors in the report, or errors in the scientific consensus. While such independent peer review would normally be undertaken prior to publication (to avoid misleading conclusions should problems be identified), some informal independent peer review has now occurred in response to publication of the BioInitiative Report. For example, the Health Council of the Netherlands (HCN) recently published a report that noted a number of inadequacies in the BioInitiative Report, inadequacies that would normally be addressed during the peer review process2.

Of particular note is that the BioInitiative Report does not appear to apply principles consistently, which biases its conclusions. For example, in arguing for a link between 50/60 Hz power lines and breast cancer, the Report does not consider some of the evidence that argues against such an association. It also provides an argument for excluding other evidence (poor exposure assessment) that is not employed for studies arguing for an association between 50/60 Hz power lines and childhood leukemia (even though they are subject to the same exposure assessment limitations; see Section 12 of the Report). Another issue is that there are statements that do not accord with the standard view of science, and the Report does not provide a reasonable account of why we should reject the standard view in favour of the views espoused in the Report.

Should we be convinced by the BioInitiative Report?

Overall we think that the BioInitiative Report does not progress science, and would agree with the Health Council of the Netherlands2 that the BioInitiative Report is “not an objective and balanced reflection of the current state of scientific knowledge” (page 4). As it stands it merely provides a set of views that are not consistent with the consensus of science, and it does not provide an analysis that is rigorous-enough to raise doubts about the scientific consensus.

It is worth noting that the state of science in this area is continually being debated and updated by a number of expert bodies comprised of the leading experts in this field. For example, the World Health Organisation (WHO) Electromagnetic Fields (EMF) project3, the International Commission on Non-Ionizing Radiation Protection (ICNIRP)4, the UK Mobile Telecommunications and Health Research (MTHR) programme5, and here in Australia the Australian Radiation Protection and Nuclear Science Agency (ARPANSA)6 have all provided authoritative analyses of the electromagnetic radiation bioeffects research. The WHO Environment Health Criteria 238 also provides a thorough analysis of the literature to date in relation to extremely low frequency (ELF, or powerline electromagnetic fields)7. We have provided some web links to these below, and would strongly urge the interested reader to consult these for a balanced perspective on this fascinating research domain.

1 BioInitiative Report: A Rationale for a Biologically-based Public Exposure Standard for Electromagnetic Fields (ELF and RF), August 31, 2007. http://www.bioinitiative.org/report/index.htm
2 Health Council of the Netherlands. BioInitiative report. The Hague: Health Council of the Netherlands, 2008; publication no. 2008/17E. http://www.gr.nl/pdf.php?ID=1743&p=1
3 http://www.who.int/peh-emf/en/
4 http://www.icnirp.de/
5 http://www.mthr.org.uk/documents/MTHR_report_2007.pdf
6 http://www.arpansa.gov.au/mobilephones/index.cfm
7 http://www.who.int/peh-emf/publications/Complet_DEC_2007.pdf

Printable version (PDF format)

What is this study about, where was it done and in what population?

This study was based on a larger Danish Birth Cohort which recruited a total of 101,032 pregnancies between March 1996 and November 2002. The authors investigated associations between in utero (prenatal) and post natal exposures to mobile (cell) phones and behavioural problems in children, using a questionnaire. The study was based on information about children born between 1997 and 1999.

What are the major strengths of this study?

The study had a large sample size (13,159) and a reasonable response rate (65%) and thus was sufficiently powered to detect subtle differences between exposed and unexposed. Almost half the children were unexposed, providing a good comparison group.

What are the weaknesses of the study?

This study has a number of other methodological limitations including:

• Failure to adjust for the effects of prematurity, gestational age and birth weight – all known causes of behavioural problems.
• Maternal recall may be influenced by previous negative or positive reproductive experiences
• Parity (number of pregnancies) may affect recall of exposure
• Paternal psychiatric history not discussed, but is also a risk factor for behavioural problems in children
• No discussion of breast feeding
• Abuse of alcohol and other psychotropic drugs not investigated
• Exposure to ionising radiation e.g. X-rays not assessed.
• Possibility of temporal confounding by other triggers such as effect of season on physical activity

What were the main findings?

There was a significant association (1.8 fold increased risk) between behavioural problems and both prenatal and post natal exposure to mobile phones after adjusting for the confounding effects of the sex of the child, mother’s age, smoking during pregnancy, mother’s psychiatric problems and socio-occupational levels.

How should the results be interpreted?

We need to distinguish between ‘association’ and ‘causation’. In this case, the authors found an association between prenatal and post-natal exposure to mobile phones based on 7 year retrospective recall of mothers and behavioural problems in children. An association means that the exposure and the outcome go together, but doesn’t mean that one necessarily causes the other. For example, it is possible that both mobile phone exposure and behavioural problems in children were due to a common cause, such as overworked parents. Although this was a cohort study, there was still potential for recall bias and associated exposure misclassification. As the authors note in the paper, results should be interpreted with caution.

What conclusions should be drawn?

Although this study demonstrated statistically significant associations between prenatal and postnatal exposures to mobile phones and behavioural difficulties in children, the authors failed to adjust for a number of possible confounding factors. The exposure assessment was limited and probably resulted in misclassification because of maternal recall bias. Therefore there is need for caution in accepting these results at face value.

Comments to Imo.Inyang@med.monash.edu.au

In a self published online report, Canberra based neurosurgeon, Dr Vini Khurana, claims that there is compelling evidence of a link between mobile phone use and brain tumours. The report purports to undertake a wide ranging review of the literature on the subject of mobile phone use and the incidence of brain tumours, mostly based on epidemiological studies undertaken in the last decade or so. The review was not published in a peer reviewed journal and presents no new research findings.

While making strong claims that “There is a growing and statistically significant body of evidence that brain tumours…are linked with “heavy” and “prolonged” mobile phone use…”, the review is inexpert and incomplete, containing a number of factual errors. In particular, Khurana fails to consider the relative scientific quality of different studies. This produces an unbalanced analysis of the literature, which is also selective in support of the author’s claims.

The ACRBR recommends a ‘weight of evidence approach’ when reviewing the RF health effects literature, giving priority to the results of published, peer-reviewed and replicated scientific research. Hence, the epidemiological evidence regarding a possible association between RF exposure and brain tumours has not indicated a causal relationship. For example, “Interphone”, the largest study performed to date, has not found any consistent relationship between mobile phone use and brain tumours, although compilation of data from this study is incomplete. In a recent update on their website (see http://www.iarc.fr/ENG/Units/INTERPHONEresultsupdate.pdf), the International Agency for Research on Cancer (IARC, which co-ordinates the Interphone study) reports little evidence in the main analyses for an overall association between mobile phone use and an increase in the incidence of head and neck tumours The ACRBR has completed a systematic review of all studies published so far (itself awaiting publication) and is also of this view, although this is contingent on the findings when the complete data from Interphone become available later this year.

It is also the view of the leading international health authority, the World Health Organisation (WHO) (see http://www.who.int/peh-emf/about/en/), that research to date has not identified any association between mobile phone base stations and adverse health effects.

However, both the ACRBR and WHO agree there is a need for more research in certain areas to improve knowledge and better evaluate any possible health risk. In response to these concerns, the ACRBR has been funded by the Australian Government to conduct a wide-ranging program of research on neurological and behavioural aspects of mobile telephony, embracing epidemiological, human, animal and cellular studies. The ACRBR also specialises in the measurement and analysis of the absorption of energy from radio devices by biological systems, including humans.

What is this study?
This investigation is part of the larger Interphone study, which is a case-control study examining the association between mobile phone use of up to 10 years and brain tumours. In this study participants were recruited from Finland, Denmark, Norway, Sweden and the South-Eastern region of the United Kingdom. This amounted to 1,522 subjects suffering from a particular brain cancer known as glioma (cases), who were recruited from national cancer registries and hospitals, and 3,301 healthy subjects (controls), who were recruited at random from national population registers or other publicly available lists of wide cross-sections of the community. To reduce systematic differences between the two groups, control subject were chosen to match case subjects in terms of age, gender, and region of residence. The researchers then compared differences in mobile phone usage patterns between the cases and controls to determine whether there was any association between their disease status (case or control) and their mobile phone use.

What were the findings?
No indication was found of increased glioma risk with mobile phone use for categories of ‘regular phone use’ (defined as use at least once a week for six months), duration of use, years since first use, number of calls or cumulative hours of phone use. Results were similar for men and women, for analogue and digital phones, and across the five countries. The authors report a marginally significant increase in the risk of glioma for greater than ten years of use, where the use is reported on the same side of the head as the tumour was diagnosed (ipsilateral). However, they also report a corresponding decrease in the risk of glioma where phone use occurred on the opposite side of the head as the tumour (contralateral). This is widely regarded as indicating recall bias in the results, meaning a skew in results due to inaccurate reporting of phone use by participants. The authors note this limitation in the particular finding and suggest that more research is needed to clarify this result.

How should the results be interpreted?
The overall results of this study, based on a large number of cases and controls, provide strong evidence of no increased risk of glioma associated with mobile phone use of up to 10 years. Results for the sub-category of ipsilateral/contralateral phone use of longer than ten years are drawn from far fewer subjects, providing far less certainty in those results. Additionally, the ipsilateral and contralateral results are contradictory, and the authors caution that such findings are “…difficult to interpret and lend themselves to…non-causal (artefactual) explanations.” In particular, it is worth noting that when the researchers restrict the data contributing to this result to only that which they consider to be of “good or very good” quality, the association reported is no longer evident.

What conclusions can we draw?
Overall, the only strong conclusion that can be drawn is that this study provides no evidence of an increased risk of glioma associated with mobile phone use up to 10 years. Firm results cannot be drawn regarding phone use of longer than 10 years, but it is hoped that the upcoming results from the full Interphone analysis from 13 countries will provide more information.

RMIT today released the findings of an independent review led by A/Prof LaMontagne (http://mams.rmit.edu.au/g60adi0a81r3.pdf), of recent investigations addressing an apparent brain tumour cluster occurring in the top two floors of their Business Studies building in Bourke St, Melbourne. The Australian Centre for Radiofrequency Bioeffects Research (ACRBR) has been monitoring this issue closely due to initial concerns by RMIT staff that this ‘reported cluster’ may be related to telecommunications base stations on the roof of that building.

The LaMontagne review was summarized as concluding that 1/ there is no evidence of a brain tumour cluster, and 2/ there is no evidence of exposures at or above levels of concern for known or suspected occupational or environmental risk factors for brain cancer. The ACRBR has assessed this review, and believe that these conclusions are entirely appropriate.

Consistent with this conclusion, the ACRBR would emphasize the following:

1/ As pointed out by the LaMontagne Review, given the size and age of the workforce over those two floors, 7 cases of malignant tumour is about what would be expected based on the incidence of cancer in the Victorian population.

2/ Radiofrequency (RF) fields in publicly accessible areas in the vicinity of mobile phone base stations are exceedingly low and below Australian and International standards.

3/ Based on extensive investigation and consultation with research and community groups, the leading international health authority, the World Health Organisation (WHO) (see http://www.who.int/peh-emf/about/en/), has concluded that there is no evidence of an association between mobile phone base stations and adverse health effects.

4/ Investigation of apparent cancer clusters is rarely conclusive. Apart from well documented events such as exposure to ionising radiation from atomic blasts or catastrophic reactor leakages and contaminants such as asbestos, it is often difficult to identify an environmental cause. Moreover, the cause of brain tumours is poorly understood (see http://www.ncri.ie/cancerinfo/clusters.shtml).

However, both the ACRBR and WHO agree there is a need for more research in certain areas to improve knowledge and better evaluate any possible health risk. In response to these concerns, the ACRBR has been funded by the Australian Government to conduct a wide-ranging program of research on neurological and behavioural aspects of mobile telephony, embracing epidemiological, human, animal and cellular studies. The ACRBR also specialises in the measurement and analysis of the absorption of energy from radio devices by biological systems, including humans.