select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to paper looking at brain tumour incidence and lifestyle factors

Researchers publishing in the Journal of Environment and Public Health investigate the association between risk of brain tumour and lifestyle factors e.g. mobile phones.

A Before the Headlines analysis accompanied this roundup.


Dr Lion Shahab, Senior Lecturer Epidemiology & Public Health, UCL, said:

“This paper provides evidence for a rise in specific malignant brain tumours in England, showing that incidence has more than doubled over the last two decades.  What the analysis does not show is that this rise is caused by mobile phones.  This is for a number of reasons.

“Firstly, this paper does not attempt to link the rise in mobile phone use with a rise in brain tumour incidence directly.  There are statistical techniques available to do this, for instance using time-series analysis, which attempt to link changes in a putative risk factor with changes in the outcome of interest over time.  This was not done here; the authors simply report an increase in brain tumour incidence and speculate about possible reasons for this.

“Second, even if such a link were found, correlation does not imply causation.  There are many environmental and lifestyle factors that have changed over the last two decades and which may show an association.  This does not mean that they actually caused this change, and any analysis would need to control for potential confounding factors.  As an example, over the same time period, there has been a significant rise in the amount of air travel which has been linked to increased exposure to potentially carcinogenic radiation.

“Third, a better test of the specific hypothesis that mobile phones cause brain tumours would be to show a rise in malignant tumours on the right-hand side of the brain over and above the left-hand side, given that most mobile phone users are right-handed.

“Fourth, while mobile phone usage in the population increased from less than 15% to 95% over the time period studied, we do not see the same increase in malignant brain tumours.  This suggests that the strength of any effect, if present, would have to be small.  For comparison, following an increase in smoking rates, lung cancer incidence and mortality increased exponentially over the course of just a few decades.

“While the paper rightly suggests that more research is needed to determine why there has been a rise in this type of malignant brain tumour, it would be premature to conclude that this is due to mobile phones.  These types of tumours are still relatively rare (with 2,500 new cases in 2015 in England compared with 38,000 new cases of lung cancer and 46,000 cases of breast cancer).  Replication of findings in different populations, using more sophisticated statistical analyses that account for confounding and which show a clear dose-response relationship of mobile phone use with location-specific brain tumor incidence, as well as good experimental evidence of biologically plausible mechanisms, would be needed to come to this conclusion.

“For now, linking a rise in malignant brain tumours to mobile phone usage remains speculative and should not detract from encouraging life-style changes which are known to reduce cancer risk, such as adopting a healthy diet, reducing alcohol consumption and stopping smoking.”


Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“On the paper itself: the researchers have indeed found a certain amount of evidence of a rise of incidence in certain types of brain tumour.  There are technical statistical difficulties in assessing the statistical significance of changes if, as in this study, the researchers look at changes across a number of different types and sites for tumours and then concentrate on the biggest changes.  So the true significance of the changes may well be less than the researchers claim.  However, as an exploratory study looking at brain tumour incidence, this research does point to something that may well be worth investigating further.  Other studies in other parts of the world have sometimes found similar increases.

“It’s important, though, to understand that this new paper did not examine any new data at all about potential causes for the increase.  There is a section that describes quite a long list of possible causal factors that have been considered by previous researchers; this list includes X-rays from CT scans, fallout from atomic bomb tests in the atmosphere (because the biggest rises were in the oldest age groups, who would have been alive when such tests occurred), air pollution from traffic, electromagnetic radiation from power lines, and electromagnetic radiation from mobile phones.  The research also quotes a paper by Ostrom et al that lists several more possible causes.  All this emphasises that, if indeed the increase in these tumours is indeed real and not some statistical artefact, nobody really knows what might be causing the increase, and the new research does not change that position.  So the researchers’ recommendation, that more research should be done on the potential causes, does seem reasonable.  That research will not be easy though, because these tumours are rare, and remain rare after the reported increases.  That’s particularly the case in the youngest age group that they considered, where the incidence rate of glioblastomas in the frontal and temporal lobes was found to be 1.5 cases per million people per year in 2015 – that’s about 30 new cases in the whole UK in this age group in a year, out of the 20 million people aged under 30.

“The press release, however, is another kettle of fish.  It was not sent out by the journal that is publishing the paper, or indeed by any of the organisations listed in the paper as potential competing interests by the researchers involved.  The embargo for the press release seems not to apply to the paper, which has been freely available in preliminary form on the journal’s website for some time.  The press release appears to come from, a campaigning organisation.  It concentrates entirely on mobile phone radiation as a possible cause of increases in some brain tumours, even though the paper itself provides no new data on that.  It concentrates on mobile phone use by children, even though the paper found that the biggest increases were in the oldest age groups.  It includes a lot of material that is not in the paper at all and have only tenuous links to it.  It has a list of four “EXAMPLES OF THE LATEST GLOBAL INITIATIVES AND FINDINGS THAT POINT TO A LINK BETWEEN MOBILE PHONE USE AND HUMAN HEALTH” (their capitals), of which two are animal studies from 2016, one is a link to a web page of a US campaigning organisation which quotes the American Academy of Pediatrics (AAP), and the final one is a mention of something from the French ANSES agency (for food, environmental and occupational health and safety).  Neither the AAP nor ANSES say that there is a link between mobile phone use and brain cancers, only that there isn’t much evidence on long-term heavy use of mobile phones and so people should be prudent about their use.  This advice matches current advice from the NHS, see, but again this is precautionary advice, and not based on any firm evidence that there are health effects of the radiation from phones.  (The press release appears to have accidentally left out the link to the ANSES advice, but it seems to be here:  None of this material about mobile phones is in the new paper.  Nor are the three graphs in the press release; the releases says they are taken from the study published in the paper, but they are not in the paper, though one of them is very similar to a diagram in the paper.

“In my view, the paper itself does provide some new information on incidence of certain kinds of brain tumours, that’s worthy of further consideration and possibly more research.  It adds essentially nothing new about mobile phones.  The press release is a campaigning document that is only tenuously related to the paper, and is in my view mostly not based on good scientific evidence.”


Prof Jean Golding, Emeritus Professor of Paediatric and Perinatal Epidemiology, University of Bristol, said:

“This paper highlights data from the English National Cancer Registry showing a rise in a particularly lethal form of brain cancer over the 21 years from 1995 to 2015, but a reduction in other forms of brain cancer.  The paper highlights the need to understand why the increase has occurred.  I would also suggest the need to find reasons why the decrease has happened.

“There are many health trends that result in curiosity about their cause, from dramatic increases in autism to decreases in lung cancer.  Such trends can, and should, initiate research into the reasons for them.  The authors suggest that mobile phone use and/or ionising radiation, including from CT scans, could play a part.  These are worth considering, but it is important to put these theories in context.  There are many other possibilities – including new pesticides, medicines, features of the diet, or air pollutants to name but a few.  All have changed dramatically over time.  Many have experimental evidence in animals of being able to cause cancer.  The question is whether they do in humans.  It is important that epidemiological evidence be examined in order to unravel answers to the important question as to why the rate of this severe type of brain cancer is increasing in this way.”


Prof Malcolm Sperrin, Director of the Department of Medical Physics and Clinical Engineering, Oxford University Hospitals NHS Trust, said:

“It is always useful to receive papers that suggest a link between disease and its cause, but the title of this paper states a proposed link to ‘environmental or lifestyle factors’.  There are many other potential influences on the data reported in the paper and it is crucial to not stretch the data too far since there are other factors that may equally be relevant.

“The identification of a trend in incidence of brain tumours is of fundamental importance if a causal link can be identified, but it is equally important to think about and take into account other factors which might influence the probability of tumour genesis or simply the probability of tumours being identified.  In this latter case this might include the increased availability of improved imaging options such as high-field MRI to detect and diagnose the brain tumours in the first place.  So, the increase in incidence may therefore actually just correlate with the rise in image quality and the subsequent diagnostic indicators.  The link to CT scans is also very tenuous – we would need a study looking at how and whether the increase in CT use, and types of use, correlates to diagnosis of brain tumours or indeed other disease classifications.

“I am not convinced that the statistical analysis in this paper is sufficient to identify any cause of the increase in tumours.  The authors do imply that other tumour types have decreased in incidence and there are a number of ways in which such findings can be explained.

“I suggest that there is a need to fully pick apart potentially confounding factors, but this is very difficult.  The potential alternative here is that incorrect links are proposed that may not actually stand up when the evidence is looked at.”


Prof Andrew Sharrocks, Professor of Molecular Biology, University of Manchester, said:

“This study shows that there has been a rise in one particular type of brain cancer in the past two decades, Glioblastoma Multiforme (GBM), which is found in the frontal and temporal part of the brain.  However, this study does not investigate or uncover any causes for this and merely speculates at possible causes such as radiation emitted from mobile phones.  Further research would be needed before any conclusions linking mobile phone use to the rise in incidence of this cancer could be drawn.  The press release goes way beyond what the paper actually looked at.  There is zero evidence of a link between mobile phones and brain tumours.”


Prof Paul Pharoah, Professor of Cancer Epidemiology, University of Cambridge, said:

“There are several notable issues with this manuscript that makes me cautious in my interpretation of the data and the conclusions of the authors.

“First, the paper is published in a little-known journal, a journal that is published by Hindawi, an open access publisher.  The quality of the peer review process for some Hindawi Journals has been questioned in the past.  If, as the press release says, this is a groundbreaking study it would have been published in a more respectable journal.

“Second, the press release is extremely misleading and written in a rather sensational style.  Some examples are given below.

“Third, while the press release describes the authors as ‘leading academic researchers’, none has a current active academic appointment.

“The manuscript itself reports on incidence of brain cancer over time over the past 20 years.  They use national cancer registration data for their analysis.  They report an increase in the incidence of brain cancer in all age groups and report that this increase is particularly notable for more aggressive forms of brain cancer and cancers in certain parts of the brain (the frontal and temporal lobes).

“It is well recognised that brain cancer has been on the increase since nationwide cancer registration was implemented in 1971.  The authors have presented nothing very new here.

“The authors report that age-standardised brain cancer incidence has increased since 1995.  However, what the age standardised analysis does not show is that the patterns of change in brain cancer incidence vary over a much longer time period than the past 20 years and that these patterns vary in different age groups.  There has been a steady increase in brain tumour incidence in people aged over 75 since 1971, but in middle aged people a steady increase from 1971 to 1990 was followed by incidence rates becoming stable.  This pattern cannot be explained by mobile phone use.

“The authors suggest several possible risk factors that may account for this increase.  However, they do not present any new data to support these hypothetical associations.  The statement in the press release that this study “points to life style factors including excessive mobile phone use as a likely cause” is highly misleading as this does study not point to any specific cause as being more likely than others.

“Since 1971 many things have changed in our environment and lifestlyle.  To select one or two as causes of the incidence patterns seen is simply bad science.  Note that whatever the proposed cause it would need to explain the flattening off of the increase seen in some age groups but not in others.  The age-standardised analysis presented in this paper will have missed such nuances and makes me question the adequacy of the peer review.  If I had reviewed this paper (and I refuse to review for journals published by Hindawi) I would have asked that such age-specific analyses be carried out.

“The authors suggest that most of the increase is due to an increase in the more aggressive types of brain cancer.  If this were true one would expect brain cancer mortality rates to mirror the increase in incidence.  However, they do not.  Brain cancer mortality rates have actually declined in the past 20 years.

“In short, the authors report a well recognised increase in the incidence of brain cancer over the past 20 years.  They present no credible evidence to explain that rise.”


Prof David Coggon FMedSci, Emeritus Professor of Occupational and Environmental Medicine, University of Southampton, said:

“The paper by Philips and colleagues suggests that in England during 1995-2015, there was a rise in one subtype of brain cancer (glioblastoma), which was partially offset by a decline in other subtypes.  Their report is unconventional in its format and terminology, making it difficult to follow in places, but the finding on brain tumour incidence is in line with an earlier study in the United States1.

“The research does not explain the observed trends, or even point strongly to possible causes.  It does not, for example, tell us when the changes in incidence began, although we know from other investigations that there was an apparent rise in the overall incidence of brain cancer during the 1970s and 80s (which may in large part have been due to improved diagnostic techniques).  It would be helpful to know whether, and to what extent, similar trends have occurred in other countries, and over what time periods.  Beyond that, more focused investigation is needed to explore specific causal theories.  Some of the causes that are mooted in the paper, such as mobile phones, have already been researched quite extensively.  However, there is no clear or consistent indication that mobile phones are a hazard.  The explanation may lie in other, as yet unidentified, factors.

“This paper only looks at trends in the incidence of brain cancer, and not at its relationship to possible causes.  The press release goes beyond the content of the paper and is way over the top.”

1 Deorah S et al. ‘Trends in brain cancer incidence and survival in the United States: Surveillance, Epidemiology, and End Results Program, 1973 to 2001’. Neurosurg Focus 20 (4):E1, 2006.


Prof David Spiegelhalter, Winton Professor for the Public Understanding of Risk, University of Cambridge, said:

“It should be kept in mind that many possible trends in incidence of various types of brain tumour were examined and only the most notable highlighted – the increase in Glioblastoma Multiforme.

“Many things have changed over the last 30 years and so the strong causal conclusions made by a campaigning group in the press release seem unjustified.”


Prof Keith Neal, Emeritus Professor of Epidemiology, University of Nottingham, said:

“The authors clearly demonstrate a rise in one type of brain cancer, which is of concern.  The suggestion that mobile phone use is responsible cannot be substantiated as the rise is greatest in the over 55s who use mobile phones much less and there was very little mobile phone use in 1995 when rates are already increasing.  Allowing for a gap of many years from exposure to cancer then other factors other than mobile phones need to be considered.  The press release does not reflect the paper – mobile phone use is hardly mentioned in the paper itself.”


Associate Prof Freddy Sitas, Epidemiologist, University of New South Wales, said:

“The age related data presented by the authors are different from the data presented by UK cancer registries to the WHO-International Agency for Research on Cancer.  Those data show no significant increases in brain and central nervous system cancers in the younger age groups.

“Of course there is work to be done to identify the causes of brain cancer, but the inferences about the role of mobile phones are not supported by their data.  The authors show increases in brain cancer starting from as far back as 1973, 12 years before mobile phones were introduced in the UK (1 Jan 1985).  I suspect much of the discrepancy is in the way brain cancer is being diagnosed, classified and reported by cancer registries over time.”


*‘Brain tumours: rise in Glioblastoma Multiforme incidence in England 1995–2015 suggests an adverse environmental or lifestyle factor’ by Alasdair Philips et al., in press in the Journal of Environment and Public Health on Wednesday 21 March 2018 and published Monday 24 June 2018.



Declared interests

Dr Lion Shahab: “None.”

Prof Kevin McConway: “I am a member of the Advisory Committee of the Science Media Centre.”

Prof Jean Golding: “One of the authors (Denis Henshaw) has been a colleague although we have not published together, and I did receive funding briefly from the charity ‘Children with Cancer UK’ about 20 years ago.”

Prof Malcolm Sperrin: “I have no vested interests in this work.”

Prof Andrew Sharrocks: “I have no conflicts to declare.”

Prof Paul Pharoah: “I have no conflicts of interest to declare.”

Prof David Coggon: “I do not have any competing or conflicting interests.”

Prof David Spiegelhalter: “No conflict of interest.”

Prof Keith Neal: “No conflicts.”

Ass Prof Freddy Sitas: “I have no conflicts of interest to declare.”

in this section

filter RoundUps by year

search by tag