May 30 2016

Underwhelming Cell Phone Rat Study

cell-phone-rat-studyMother Jones headline declares: “Game-Changing” Study Links Cellphone Radiation to Cancer.” NBCNews was similar: A Possible Cellphone Link to Cancer? A Rat Study Launches New Debate.

Any evidence that might link cell phone use to cancer is of legitimate concern, but this is a classic situation in which such evidence needs to put into proper context. I will start with some reassuring clinical context – human epidemiology data has failed to show any consistent association between cell phones and cancer. Further, brain cancer rates have not been increasing overall in the last 20 years when cell phone use skyrocketed. Therefore, any real world effect of cell phones on humans must be tiny to nonexistent.

Toxicology science, however, looks at questions several ways. The most definitive evidence would be placebo-controlled trials, but we almost never have this because it is unethical to expose a subject to a possible toxin just to see if it has a negative effect. (You can do this as part of a therapeutic trial where there is a greater chance of benefit to the subject, but not just to test toxicity.)

So, the best evidence we actually can get is epidemiological or ecological – looking at people out in the world to see if there is an association. You can follow people with varying levels of exposure and track possible negative outcomes. You can also look at people with negative outcomes and ask them about prior exposures. Further, you can look at the overall incidence of the negative outcome to see if it is increasing after a potential risk is introduced, or decreased after it is removed.

The strength of this type of data is that it is in humans and it is a real-world assessment of risk. The weakness is that this data is never fully controlled and so we have to extrapolate probable causation from correlation.

We can also do toxicology studies in cells or in animals. The strength of these studies is that they can be highly controlled. The weakness is that they are not conducted in whole people, and still we are left with extrapolating the data to actual effects on people. Also, it is easy to generate positive results by using unrealistic exposures.

Often these two types of data conflict, usually because toxicology data in cells or animals shows a possible effect which is not detected in people epidemiologically. I think the best way to look at the data is this – the in vitro and animal data is used to establish plausibility; is there a biological effect that is potentially toxic to humans? Epidemiological data is used to ask if any toxicity has actually occurred.

Toxicology data involving cell phone radiation is therefore not very concerning to me when large scale epidemiological data has already failed to show a consistent correlation.

The New Rat Study

With that in mind, let’s take a look at this new rat study to see what it shows. Here is the quote being passed around making it seem like this data is concerning, from Christopher Portier, a retired head of the NTP who helped launch the study and still sometimes works for the federal government as a consultant scientist.

“There will have to be a lot of work after this to assess if it causes problems in humans, but the fact that you can do it in rats will be a big issue. It actually has me concerned, and I’m an expert.”

The first thing I note is that the study exposed rats to 9 hours a day of total radiofrequency (RF) radiation to the whole body in a reverberation chamber. This exposure started in utero. It’s a stretch to say that this mimics RF exposure from cell phone use. I guess if you spend half your waking life in a phone booth where your whole body is exposed to high levels of RF, you might be concerned.

The study also notes that rats of exposed mothers had decreased body weight. This to me sounds like a confounding factor – why was their body weight reduced, and what other negative effects could this have?

These two points aside, the results themselves are not very impressive. In male rats there was a statistically significant increased rate of brain gliomas and heart schwannomas. The authors claim there is a dose response effect, and maybe you can tease out a statistical effect, but looking at the numbers there does not appear to be a clear dose-response effect.


You might also note that the absolute numbers are very small – just 1-2 cases total in each affected group. These are marginal results at best.

The numbers are slightly more significant for the heart schwannoma results, but those results have other problems. This was whole-body RF exposure, and yet only the heart showed a significant result with respect to schwannomas. Schwann cell exist throughout the body. Yet, when other tissues were examined there was no effect. Also, when you take all the schwannoma data together, there is no statistically significant effect – only when you pull out the heart data by itself.

Even more significant – all of this only applies to the male rats. There was no significant effect in the female rats. So, this effect in male rats does not even extend to females of the same species, what are the odds it will extend all the way to humans?


For all these reasons I am, to say the least, unimpressed. The results are marginal and inconsistent. There is also the possible confounder of reduced birth weight. The exposure is also a lot greater than what humans get exposed to from cell phone use.

At best this study shows that RF radiation can potentially have a biological effect. That is news, because this has not been clearly established. I don’t think this study clearly establishes an effect, however. The results are easily consistent with noise plus possible unknown confounders.

To highlight this, the study also showed that the exposed rats lived longer than the control rats. Headlines, however, did not declare that cell phone use makes you live longer. This is probably just as quirky and spurious a result as the higher tumor rates.

This data is certainly not enough to call into question the negative epidemiological data, and I don’t think it warrants scary headline. I would also note these are preliminary results released to the media prior to peer-review and publication. In the long run, I don’t think these results are going to change the conversation on cell phones and cancer, as the headlines indicate.


Note: David Gorski also comments on this study at SBM


12 responses so far