A recent paper on rats exposed to cell-phone radiation (text note: for more information on Wifi and Cell Phone radiation, please see our article on our Extraordinary Claims website) has garnered much attention, even though it has not completed peer review and offers inconclusive results.

For some reason the US National Toxicology Program (NTP) decided to publish a paper in May, 2016 about “Cell Phone Radiofrequency Radiation” in rats before the peer review process was completed, presenting a learning opportunity for lay people.  Reviewers have noted several problems with the experimental design and the presentation of the results but most of the popular media has either misrepresented the findings or grossly extrapolated the conclusions. The following is a list of reasons why this paper does not overturn the thousands of other studies that have, so far, failed to find any link between cell phone use and cancer of any kind.

  1. The study is not yet published.  There was no reason given for releasing the paper as a preprint, but it may be related to some public discussions of their findings.

  1. The study is in rats. It’s difficult to overemphasize that rat and mouse physiology is very different to that of humans.   There are hundreds of drugs that looked promising in rodent studies but utterly failed in humans.  Only human studies can tell us about human physiology.

  1. An important anomaly that has yet to be explained is why none of the control rats developed tumours. Background rates for gliomas and schwannomas (the two nervous system tumour types included in the study) are between 1-6% (according to the authors) for the Sprague-Dawley rats they used.  The tumour rates for the test rats (those exposed to the radiation) in this study were in the range normally found in the control groups for other studies.  This represents a huge red flag pointing to the need for replication.

  1. The most troubling problem is one of cherry-picking. There were 3 test groups for each of two cell phone protocols (GSM and CDMA). Rats in each protocol were exposed to 1.5W/kg, 3W/kg or 6W/kg for a total of 9 hours per day and there was one control group. Each of the 7 groups contained 90 rats of each sex.

The raw data show no correlation between radiofrequency dose and tumour incidence.

There were 3 malignant gliomas in the male rats receiving each of the low and middle doses of GSM radiation but only 2 in the highest dose group.  Carcinogens typically show a dose-response curve, meaning that higher rates of carcinogen result in more tumours, but this is the opposite of that. In the CDMA group there were no tumours in the first two groups and 3 in the largest dose group. Three rats is a small number and should not be considered significant for this experimental design.  In fact the authors repeated state that there was “no statistical difference” between any of the groups and the control group.

For female rats there were no tumours in the two low doses for the GSM protocol and only 1 for the highest dose. For the CDMA protocol there were no tumours in the highest dose group and 2 in the lowest dose group. If these differences were statistically significant, which they are not, it would suggest that the radiation prevented cancer, but thankfully nobody jumped to that conclusion.

Schwannoma tumours occurred in 2, 1 & 5 male rats for the GSM protocol and 2, 3 and 6 for the CDMA protocol, in order of increasing dose. For females it was 0, 2 and 0 for GSM and 2, 0 and 2 for CDMA. Clearly there is no trend.

There was one group with a tumour incidence of 6.6% and the rest were below that. These data fall within the normal control rate of 1-6% and should not be considered significant. Reviewers noted that the nature of the experimental design and statistical analysis were such that it would lead to false positives and one bluntly stated “I am unable to accept the authors’ conclusions”. Reading the reviewers’ comments at the end of the paper is highly instructive for those who want to know what it’s like to have a paper reviewed.

Despite the lack of significant results the authors reported that there is a “trend” if you group the results a certain way. This kind of bias is called cherry picking if you want to be polite. If you construct enough groups, by chance alone you will eventually get one that increases but that is not the one you should base your headlines on. These results simply constitute “noise” and are meaningless.

  1. The dosage was in excess of anything a human with a cell phone would experience, by the authors’ admission. Also, the exposure began as the rat pups were gestating, in utero, and continued for two years, the typical lifespan of a lab rat.

What is significant is that a huge National Cancer Institute study showed that, despite a massive increase in cell phones and other devices using electromagnetic radiation over the past few decades, brain cancer has been steadily declining.  Cancers of the ovaries, larynx, esophagus, stomach as well as Hodgkin’s lymphoma have also been decreasing in the US.

Surprisingly the male rats exposed to the radiation lived longer than the control groups for both protocols, but that didn’t make the headlines.


Further Inquiry and Sources

  1. http://biorxiv.org/content/early/2016/05/26/055699
  2. http://seer.cancer.gov/statfacts/html/brain.html
  3. http://progressreport.cancer.gov/diagnosis/incidence
  4. https://www.sciencebasedmedicine.org/no-a-rat-study-with-marginal-results-does-not-prove-that-cell-phones-cause-cancer-no-matter-what-mother-jones-and-consumer-reports-say/
  5. http://microwavenews.com/news-center/ntp-nyt