• How reliable is the epidemiological evidence on mobile phones and cancer?

This blog was first published on July 24, 2009 on http://rock-and-hard-place.blogspot.com/

This blog was updated on July 25th, 2009 with a post scriptum.

The opening lecture of the recent BioEM 2009, the Joint Meeting of the Bioelectromagnetics Society and the European BioElectromagnetics Association in Davos, Switzerland (http://bioem2009.org/) was given by Anssi Auvinen. He reviewed all to date published epidemiological studies dealing with mobile phones and cancer. In the conclusion of his talk Auvinen said that we have performed sufficiently many epidemiological case-control studies but that their scientific evidence is of insufficient quality to reliably draw any health-risk-related estimates. Responding to the question from the audience, Auvinen admitted that execution of more of the case-control studies will not improve the situation, because of the unavoidable biases, and will be only waste of time and money.

I do agree with Auvinen’s position. At this stage of the research, the extensive list of various limitations of the executed case-control studies is casting a strong doubt over the reliability of the available epidemiological evidence as a basis for any human health risk estimate. The insufficient quality of dosimetry, selection and misclassification bias, low sensitivity in detection of health risk within the population are causing that drawing any health-related conclusions (no-risk or risk) is like “flipping-a-coin”.

Another recent development is the publication of two different reviews by the ICNIRP Standing Committee I – Epidemiology. One of the reviews (available already on-line) will be published in September 2009 issue of journal Epidemiology (http://journals.lww.com/epidem/toc/publishahead). The other review is published as part of the ICNIRP Report (http://www.icnirp.de/documents/RFReview.pdf).

There are two major surprises associated with ICNIRP SC-I reviews. First is the timing of publication and the second, even bigger surprise, is the difference in conclusions of both reviews.

The first surprise is the publication time of the reviews. The obvious question is why ICNIRP SC-I did not wait for the publication of the summary analysis of Interphone project, which has been already submitted and undergoes peer-review before publication? Without inclusion of this “nearly ready” summary analysis of Interphone, the ICNIRP SC-I reviews are incomplete and already “outdated”.

The second surprise is that the health risk conclusions are different in both reviews.

In the review published in journal Epidemiology it is clear that the authors “want to have it both ways”. On the one hand they diligently present scientific results of the published studies with all their limitations, shortcomings and unreliability:

  • …“some key methodologic problems remain, particularly with regard to selective nonresponse and inaccuracy and bias in recall of phone use”…
  • …“all studies based on phone use are affected by exposure misclassification, which (if nondifferential) could dilute risk estimates”…
  • …“data on exposures more than 10 years before cancer diagnosis are still limited. Most types of cancer occur many years, or even decades, after initial exposure to known carcinogens. A widely expressed view has been that it is therefore too soon to know whether mobile phones have an effect on cancer risk”…

The review of the results of published scientific studies is comprehensive and it takes into consideration all published studies, including these by Swedish group of Lennart Hardell. The Hardell studies are considered as problematic because their results differ from the results of other studies. Therefore, the meta-analysis results are presented in two formats as data with and without Hardell studies. Unfortunately, the authors of the review did not put sufficient effort to determine what might be the cause of the differences between “Hardell and the rest of the world”. They satisfied themselves with a statement indicating that in fact they do not know what is going on:

…“In our view, the series of decisions in methods, analysis, and presentation provide the most plausible explanation for the deviation of the findings of the Hardell studies from those of other investigators.”…

Immediately after this statement comes another puzzling sentence:

…“This does not address the other positive reports, but they seem to fit more in the distribution of results expected given random error across studies.”…

Is it so that the authors of the review consider (any) other positive studies as caused by a “random error” only? Just for the sake of dispute: could it be that the negative studies are caused by a “random error”?

But then, as if all the methodological limitations and inconsistency of the results of various published studies would not matter, the authors of the review reach a surprising conclusion:

…“Overall the studies published to date do not demonstrate an increased risk within approximately 10 years of use for any tumor of the brain or any other head tumor. Despite the methodologic shortcomings and the limited data on long latency and long-term use, the available data do not suggest a causal association between mobile phone use and fast-growing tumors such as malignant glioma in adults (at least for tumors with short induction periods). For slow-growing tumors such as meningioma and acoustic neuroma, as well as for glioma among long-term users, the absence of association reported thus far is less conclusive because the observation period has been too short.”.

It seems like all limitations of epidemiological evidence were put aside in order to reach the no-risk conclusion.

In the same issue of journal Epidemiology is published a commentary, by Kenneth J. Rothman (Rothman Commentary), about the ICNIRP’s review. He also presents a list of severe limitations of epidemiological studies on mobile phones and cancer that seriously undermine the usefulness of the evidence:

  • …“If using mobile telephones causes cancer with a long induction time, use of today’s mobile telephone may be inducing many future cancers but relatively few of today’s cancers, given the comparatively small number of users until recent years.”…
  • …“By far the greatest obstacle is exposure assessment.”…
  • …“the study of tumors occurring now might reflect the effects, if any, of first- and second-generation telephones that are no longer used”…
  • …“without an elaborate metering effort that would be impractical for any but a small subsample of users, a person’s actual tissue exposure to microwaves can only be vaguely estimated, even from telephone call records.”…

Rothman states that the authors of the ICNIRP SC-I review felt very confident about their interpretation of the science but he also noted that skeptics might rightly disagree with the conclusions of the review. He wrote:

…“Despite these problems, the reviewers were confident enough to conclude that “Overall the studies published to date do not demonstrate an increased risk within approximately 10 years of use for any tumor of the brain or any other head tumor.” Skeptics might rightly take this as only mild reassurance, because induction times for radiation caused tumors often exceed 10 years.”…

Indeed, some of the questions that are not at all addressed by epidemiological studies, which also skeptics might pose, are:
Assuming that the mobile phone radiation causes brain cancer

  1. What will happen to the user that stops using the phone within the 10 years of “no-risk” period? Will he be at risk later or will the effects of exposure simply disappear?
  2. How many years of prior exposure to mobile phone radiation are sufficient to cause effects even if phone use is later terminated? Cancer might not appear within the first 10 years due to latency period but is one, or more, years of use of mobile phone sufficient to trigger the cascade of events leading to development of cancer at later time?

The same authors, ICNIRP SC-I, wrote another review that is the part of the just released ICNIRP Report. Also in this review are duly listed the major methodological limitations of epidemiological studies:

  • …“A key concern across all studies is the quality of assessment of RF exposure, including the question of whether such exposure was present at all.”…
  • …“Another general concern in mobile phone studies is that the lag periods that have been examined to date are necessarily short. The implication is that if a longer lag period is required for a health effect to occur, the effect could not be detected in these studies.”…
  • …“The majority of research has focused on brain tumors and to some extent on leukemia. However, because the RF research questions are not driven by a specific biophysical hypothesis but rather by a general concern that there are unknown or misunderstood effects of RF fields, studies on other health effects may be equally justified.”…

However, the health risk conclusion is very different:

…“Results of epidemiological studies to date give no consistent or convincing evidence of a causal relation between RF exposure and any adverse health effect. On the other hand, these studies have too many deficiencies to rule out an association.”…

Why such a dramatic difference in conclusions published at the same time by the same authors but in two different publications? Another puzzle… To me, the present message of ICNIRP, as expressed in conclusions of the two above presented reviews, is clearly confusing. Depending on the reader, it can be freely interpreted to suit any needs.

What is next? It is the long overdue summary analysis of Interphone Project. Great hopes for the answers whether there is a link between brain cancer and mobile phone radiation exposure were with the EU 5th Framework Programme-funded “Interphone Project”. This project is a multinational case-control study that was set-up to investigate whether mobile phone radiation increases the risk of cancer and whether it is carcinogenic. The study focused on tumors development in brain and head tissues that are the most exposed to mobile phone radiation and can generate glioma, meningioma, acoustic neurinoma and partoid gland tumors. The study was conducted in 13 countries: Australia, Canada, Denmark, Finland, France, Germany, Israel, Italy, Japan, New Zealand, Norway, Sweden and the United Kingdom (divided in South and North). The hallmark of this project was supposed to be the common core protocol that would allow to pool data obtained in all 13 countries and to perform one statistical analysis that would consist of sufficient numbers of cases and controls to be meaningful and informative. However, as it appears there were numerous exceptions from the common core protocol that put in doubt the reliability and the validity of the yet unpublished pooled-analysis (for details see Eur. J. Epidemiol. 2007; 22: 647–664).

Furthermore, the definition of the regular user is, in my opinion, another major problem of the Interphone protocol. As a regular user it is defined a person who makes at least one phone call per week for the period of at least 6 months. This means that person who makes 24-25 calls over 6 months period (one call per week) is put into the same category of regular user as a person who makes 24-25 calls per day – as e.g. many business people do. This means that the category of regular users consists of those who are only occasionally using mobile phones and are very little exposed to mobile phone radiation and those who use them on daily basis and frequently and are very heavily exposed. It means that, if we assume that the heavy exposure to mobile phone radiation could cause any health effects, then these effects will be diluted by analyzing in the same group the very low- and the highly-exposed study subjects.

The problem of the so-called “regular user” has been well compared by Lloyd L. Morgan (Am. J. Epidemiol. 2006; 164: 292–296) to tobacco use:

…“Would we expect to find a risk of lung cancer (which is a high risk among smokers) for smokers who had smoked once a week for 6 months”…

To me, the Interphone’s definition of “regular user” describes rather an “occasional user”. Use of the definition of “regular user” has also other consequences. In the conclusions of some of the already published Interphone studies we can find sentences stating that the study did not find any increased risk of brain cancer among the “regular users”. Such conclusions are clearly misleading because no lay person would ever assume that making one phone call per week for 6 months period constitutes regular use of mobile phone.

However, when setting definition of “regular user” in such way as Interphone did it, it is surprising that there are Interphone studies suggesting the possibility of the increase in brain cancer incidence among the long-term “regular users” (>10-years). These results are still statistically uncertain and might either be showing a real link or might be just a statistical “glitch” due to small sample numbers. However, these results were obtained using definition of “regular user” as specified above. It means that the effects induced by mobile phone radiation could be seen even when they are “diluted” by the presence of very low exposed users among the so-called “regular users” population.

Finally, the so far executed epidemiological studies examining health effects of mobile phone exposures were focused only on cancer. At the same time, the discussion continues whether this radiation could cause effects that, although not able to develop into life-threatening disease, could become detrimental to the quality of life. These life non-threatening effects might include e.g. sleep disorders, headaches or various allergy-like symptoms. Therefore, independently of their outcome and/or interpretation, the so far executed epidemiological studies provide information only about the cancer and are unable to give mobile phone radiation “a clean bill of health”.

In conclusion, to answer the question presented in the title of this blog, I think that the epidemiological evidence is not sufficiently reliable to conclude that human health neither is nor is not at risk. I think that at this time any statements suggesting that there “is health risk” or that there “is no health risk”, based on the epidemiological evidence, are premature and not reliably supported by the available science. To me, the case-control studies have failed to provide useful and reliable evidence. Will the ongoing cohort study provide such? Only time will show… But I have doubts because we are dealing with a very weak stimulus and epidemiological approaches are very slow and might have too low sensitivity to reliably detect changes in the huge population of mobile phone users. We should focus on human volunteer studies using different subpopulations of users to determine whether mobile phone radiation has any significant effect on human physiology. We did not study it enough, mistakenly hoping that epidemiology will answer the questions and solve the problem…

Post scriptum added on July 25th, 2009

In response to Cindy Sage comments, I still stand by my statement that “…at this time any statements suggesting that there “is health risk” or that there “is no health risk”, based on the epidemiological evidence, are premature and not reliably supported by the available science…”. However, even though the data are of insufficient quality to draw reliable conclusions, the existence of data suggesting the possibility of increased risk of brain cancer is, in my opinion, sufficient to advice precaution and to request further research to clarify this issue. At the present, one can not “yes” or “no” to the risk question but we can advice precaution because there are indications that risk might exist.


Comments copy/pasted on August 6, 2009 from http://rock-and-hard-place.blogspot.com/


Elaine Pratt said…

I find the subject of the brain fascinating. I wonder about the radiation effect of the right hemisphere of the brain vs. the left hemisphere of the brain. Most people being right handed and holding the phone against the right side of their head. Will that cellular change eventually have a different effect on left vs. right handed people; physically, behaviorally? What about the effect on newer cells such as those within the eye?

July 24, 2009 3:18 PM


Cindy Sage said…

Several questions and comments:

1) You say Auvinen profiled ‘all published epidemiological studies to date dealing with mobile phones and cancer’ at the BEMS meeting. Did he include the Kan et al., (2007) meta-analysis? For brain tumors, it reported a pooled OR for long-term users of 10 years and longer (5 studies) at 1.25 (CI = 1.01-1.54). This is important. The meta-analysis EXCLUDED any Hardell et al studies. It still found increased risk of brain tumor. J Neurooncol. 2007 July 1:(Epub). Kan P Simonsen SE Lyon JL Kestle JR. Division of Pediatric Neurosurgery, Department of Neurosurgery, University of Utah, 100 N. Medical Drive, Salt Lake City, Utah. john.kestle@hsc.utah.edu

2) You say Auvinen indicated “the scientific evidence is of insufficient quality to reliably draw any health-risk-related estimates.” Well, this cannot be true when all the published studies to date with data on 10 year and longer latency report increased risk of malignant brain tumor. It is enough to say there is early evidence of some risk, but the risk cannot yet be quantified precisely. So, it is sufficient for early precautionary advice. That should be the signal from the evidence to date.

3) By agreeing with Auvinen’s position that “more case-control studies will not improve the situation because of unavoidable biases, and will only be a waste of time and money”, are you saying that Hardell et al studies are a waste? That they contribute no useful information?

4) You say that “case-control studies” are of such insufficient quality that drawing any health-related conclusions is like ‘flipping-a-coin”. In fact, you could and should say also that cohort studies suffer from the same methodological deficiencies and are equally or more unreliable. There are several published articles reporting the specific deficiencies of other types of epidemiological studies, including exclusion of heavy users, diluting effects by including improper definition of regular user, drawing conclusions from latency periods too short to show effects and finding none, concluding there is no risk, etc. Deficiencies that have led to ORs consistently under 1 across many of the published studies indicate something is wrong with their research design, or execution.

5) I think the results of the paper in Epidemiology and the ICNIRP RF Review paper do have consistent conclusions (you do not). ICNIRP concludes that “Results of epidemiological studies to date give no consistent or convincing evidence of a causal relationship between RF exposure and any adverse health effect. On the other hand, these studies have too many deficiencies to rule out an association.” A conclusion of the Epidemiology paper is “For slow-growing tumors such as meningioma and acoustic neuroma, as well as glioma among long-term users, the absence of an association reported thus far is inconclusive because the observation period has been too short.” It cannot rule out effects either. And because the Epidemiology paper omits reference to the Kan et al., 2007 finding of increased risk at 10 or more years, excluding all the Hardell et al papers, and because it dismisses its own finding that combined results of studies with 10 years or more of cell phone use show elevated risks as ‘deviant results’, fixing these two glaring errors would perhaps require a shift to “there is some evidence” or “there is a real risk”.

6) You conclude that “at this time, any statements suggesting that ‘there is a health risk’ or ‘there is no health risk’, based on the epidemiological evidence, are premature and not reliably supported by the available science.” Dariusz, you forget that it is not black and white, nor are we given only two choices – yes it does, or no it does not conclusively cause brain tumors. There is the important middle ground which would say “we have some evidence… and that the risks of doing nothing with it, of waiting until we know definitively one way or the other, will result in decades of unnecessary exposures that could have been prevented by some sensible public health cautions based on the data showing some risk at 10 years or longer use. There is no excuse for demanding an ‘either or” scenario. Thanks for the forum to discuss these ideas. You are doing us all a service by providing a place for exchange of viewpoints.

July 24, 2009 9:50 PM


Dariusz Leszczynski said…

Response to Cindy Sage comments

Dear Cindy,
In all three articles cited in my blog (two from ICNIRP SC-I, one from Rothman) and in Auvinen’s lecture at BioEM2009, were listed numerous methodological problems with the case-control studies conducted so far. The major problem is that there are compared good quality data on cancer with very poor quality data on radiation exposures. This causes that the data are of “insufficient quality”. Continuation of research without substantial improvement of the quality of data on radiation exposure will only add more of “insufficient quality” data. Having more and more of “insufficient quality” data will not miraculously change them to “quality data” by sheer piling of numbers. These will be still comparisons of good data on cancer and poor data on exposures. That is why continuation of case-control studies without substantial change of exposure assessment will be waste of time and funding.

I stand by my statement that “…at this time any statements suggesting that there “is health risk” or that there “is no health risk”, based on the epidemiological evidence, are premature and not reliably supported by the available science…”. However, even though the data are of insufficient quality to draw reliable conclusions, the existence of data suggesting the possibility of increased risk of brain cancer is sufficient to advice precaution and request further research to clarify this issue. So, we can not say yes or no to the risk question but we can advice precaution because there are indications that risk may exist.

As to the “confusing” ICNIRP’s messages – yes, they are confusing. In one article the experts say that the studies have not shown causal relationship and that they, the experts, think that it is unlikely that such causal relationship exist. What it tells the non-experts is that any suspicions can be disregarded because the experts think them unlikely. On the other hand in another article the same experts can not make up their mind and say that the results could still go both ways and they do not assure us expertly that it is unlikely to be that case. Hence – confusion.

July 25, 2009 9:44 AM


Beti Dejanova said…

I absolutely agree with your opinion, although I believe that this “risk factor” might show as a real one for brain cancer appearance. More studies related to cell culture experiments should be done to confirm it. Up to now epidemiological studies can’t be realable. Beti Dejanova, Macedonia

July 25, 2009 1:34 PM


Anonymous said…

It is a nonsence that people have do be argumented 100% about risks of mobile phones exposure, as it effects their own health. The final evidence due to noted reasons will probably not be gained in the near future. All the wireless communications can be replaced by wired, specialy by optical cables, so there is no need to take a chance with exposure. As mobile phones use intensity is mostly a matter of personal decision, I found the most problem from the mobile masts, as they present involountary exposure.

July 27, 2009 12:18 AM


Anonymous said…

Report from ICNIRP is very insufficient as it does not rely on other possible risks. Induced currents pass thru the body can effect chronic stress that is not initialy observed and effect human behaviour. It is noted in the Hallberg Johansson 2005. research an incidence of melanoma regarding the number of masts, as well as dr. Magda Havas research of effects of exposure on diabetes.

July 27, 2009 12:35 AM


Stelios A. Zinelis  said…


With a great interest I read that “…Auvinen admitted …more of the case-control studies will not improve the situation … and will be waist of time and money”. Also you “…do agree with Auvinen s position.” Furthermore, “…health-related conclusions(no-risk or risk) is like “flipping a coin””.
The public would not agree with this approach, since they are exposed to electromagnetic radiation without their permission and their knowledge. What good does it do to the individuals who would be affected with this radiation and in 30-50-100 years measures will be taking to eliminate this risk?
The past has taught us many lessons for example asbestos. The mining began in 1879. Nineteen years later was reported about the dust harmful effects. However, no actions were taking to protect the public. If a ban was done back in 1965 when the cause for adverse effects was plausible but unproven only Holland would have avoided 52,000 victims and 30 billion euro in costs for the period 1969-2030 and an estimation of 250,000-400,000 deaths from mesothelioma, lung cancer and asbestosis will occur the next 35 years in European Union from the past asbestos exposure.
With the current status of the epidemiology studies: 1) Problems in design of the Interphone studies, 2) Excluding Hardell s studies and 3) The Danish study-where the criteria of inclusion was if someone “ever” used phone between the period 1982-1995 and also by excluding the corporate phone users(most likely heavy users), it is not possible to resolve the issue and this may have major consequences in public health.
However, objective, well design, independent epidemiological studies can contribute information of EMF and biological effects.
Some of the Interphone studies have shown an adverse effect by the EMF. This is very powerful(with a such design) and along with Hardell s findings, we should apply immediately the precautionary principle and minimize the human exposure. We would do more good than harm.
Thank you
Stelios A Zinelis MD
Hellenic cancer Society
Cefalonia, Greece

August 3, 2009 5:04 PM


Dariusz Leszczynski said…

Dear Dr. Zinelis,
As I have stated in the post scriptum, I agree with the notion that the existence of studies indicating (eventhough not proving) possible health effects, is sufficient to advice preacution to the mobile phone users.

August 3, 2009 5:16 PM


Erja Tamminen said…


Thank you for the analysis and insightful questions concerning the research on cancer risk. It is nice that you decided to continue conversation on this issue privately even after STUK decided to close down your blog in their web-site. (The board of STUK decided that the contents of the blog was not in line with the official policy STUK is conducting….)

1. There is something about professor Anssi Auvinen that I would like to comment. The TV-news made an interview (2000) with Auvinen when the Interphone started in Finland. Auvinen stated: “In case the results of the Interphone project would show any cancer risk in the future, that would have remarkable consequences for the mobile industry and for the whole society.” In the EBEA -seminar two years later Auvinen presented some preliminary Interphone results of his own but totally dismissed findings. Two years later in a seminar at the University of Helsinki Auvinen presented all the epidemiological evidence, including the Interphone results available by 2002, confirming the audience: “By now only the Hardell group from Sweden has found some evidence for tumour risk but the study of Hardell is of poor quality due to methodological problems”.

I made an interview with Auvinen in 2007 for my book (EMF) and inquired about the methodological problems of Interphone. He had no time to answer my question. Professor Lennart Hardell can be very proud of his research.

 2. What concerns me now are the cancer statistics from UK and US indicating an increase in the brain tumour incidence with children and young people. (Neurosurgeons Charlie Teo, Keith Goh and Vini Khurana have faced this in their practises. The results by Lennart Hardell support these alarming observations.)

In Finland nearly 99,99 per cent of children have mobiles. Hands free or headset are not commonly used. The legislation obliges the mobile companies not to provide their clients with ordinary landline phones which would be preferable for children. Fibre cable is being built but will not cover the entire country. In January 2009 the Finnish Radiation and Nuclear Safety Authority (STUK) adviced children to follow the precautionary principle with their mobiles. It is hard to believe that simultaneously the professors of the Finnish Cancer Association (Risto Sankila and Harri Vertio) encourage children to use mobiles and neglect science reports (and the vulnerability of children). In the medical journals they refer to Interphone. According to Sankila & Vertio, the Finnish brain tumour statistics indicate nothing to be worried about. The question is, how much do society and science nowadays respect the value of human life?

According to statistics, Finland is a country free from corruption. However, in the Finnish society we have wide and open networks supporting the wireless technologies and especially Nokia which has an influence on the government and threatens to establish its headquarters abroad in case of unfavourable regulations. In France a consensus at the governmental level has been emerged and as a result the national RF research programs should cut its ties to mobile industry. After Interphone it is time to make the same decision internationally. I suggest this should be taken seriously into consideration in the planning of the programme of the hearing conference in the US senate in September.

3. Professor Ken Rothman surprises me. He was given the opportunity to comment the Epidemiology article. In my opinion more criticism could have been expected from an experienced epidemiologist like him. For example Rothman agrees with Ahlbom by saying: “Even for carcinogens that induce cancer with very long average induction times, a causal effect should produce some increase in risk much earlier than the average induction time. Therefore, the absence of an effect for the first ten years after exposure should have implications that extend beyond that time.” How can this view be supported when Rothman himself admits there are serious methodological deficiences in the Interphone?

Erja Tamminen

August 4, 2009 5:53 PM


Mikko Ahonen said…

Shouldn’t we look at history and chronic exposure to microwave/RF-radiation?
I have tried to collect all positive and negative findings related to:
Mobile phone base stations:
FM / AM Radio, television and radar transmitters:
Occupational exposure:
(this OE area is still under development)
Several epidemiological studies indicate risk of leukemia and lymphomas. Based on that, I see that we be should be very cautious with especially children’s use of mobile phones.

August 4, 2009 6:54 PM



2 thoughts on “• How reliable is the epidemiological evidence on mobile phones and cancer?

  1. Pingback: • “A job worth doing is a job worth doing well” « Between A Rock and A Hard Place Blog

  2. Pingback: • Interphone afterthoughts and question about compliance of mobile phones with safety standards « Between A Rock and A Hard Place Blog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s