The partial findings from a recent animal experiment conducted by the National Toxicology Program (NTP) on rats and mice exposed to cellular telephone radiofrequency (RF) waves are too limited to interpret. Further, protection guidance relies primarily on human population-based studies, not animal data. The high-quality epidemiologic studies are providing increasing evidence that mobile phone use does not cause brain tumors in humans.

The NTP animal experiment is the best yet conducted, but has serious limitations. The comprehensive reviews by scientists from the National Institutes of Health, attached to the report, are informative with one concluding: “I am unable to accept the authors’ conclusions”.

Consistency (or replication) is the key to interpreting findings. In these data, the results were not internally consistent in that the significant excesses occurred only in male rats and not in the females. Further, it appears that there were no similar effects in mice of either sex. Thus the findings are not replicated in studies conducted by the authors.

Another peculiarity is that the control rats did not live as long (survival rate 28 percent) as male rats exposed to RF nor with NTP control rats in other studies (47 percent). The tumors presented, gliomas and schwannomas, occur in later life and apparently control mice did not live long enough to develop these tumors. In fact, none of the control rats developed even one of these tumors! This was not the case for control rats in other NTP studies who lived longer. Do the findings in male rats reflect a RF effect or is it partially influenced by a low rate of tumors in controls related to shorter survival?

The study was severely underpowered. What does that mean? The number of rats was not large enough to detect even a high risk, raising concern that the findings "might" reflect an increased likelihood of false positives. This concern could be alleviated by releasing all of the data, i.e. was there an equal number of significantly low findings? Tumors of interest not reported include meningiomas, salivary gland tumors, and acoustic neuromas.

Only selective results were reported for brain tumors (gliomas and hyperplasia) and schwannomas of the heart. Schwannomas are tumors of tissues that cover the nerves, and the most common are acoustic neuromas in the ear. Heart schwannomas have not been linked to cellular phone use in epidemiologic studies. To account for the possibility of false positive findings and reporting bias, there is a need to know all the other finding, in mice and in other tumors. For example, many schwannomas occur in other tissues (nerves in the ear, leg, arm, lower back), but there was no increased risk of schwannomas overall or for any other tissues. This raises concern about reporting bias of positive results where chance may have played a role, i.e., when you make many comparisons, by chance alone some might be significantly high and others significantly low.

Generalizing this study to humans is problematic. The rats were exposed to 900 MHz of different modulating frequencies for 9 hours per day, every day of their lives, starting as a fetus. The whole-body nature of the exposure differs from the exposure received when holding your phone to your ear, especially for "heart" exposures. The findings in only one sex and in only one species adds concern on how to interpret these results, even if borne out when a full report is released. How might these findings be incorporated into current recommendations? Or what should you be concerned about?

Exposure limits are currently set based on thermal properties; the RF waves might jiggle molecules and thus increase heat—when your cell phone gets hot, however, it’s because of the battery, not the RF. Expert agencies and nonionizing radiation committees conclude that the epidemiological evidence to date does not convincingly or consistently link RF waves to cancer in humans.

The relevance to humans might be what? Usually high quality population-based studies take precedence over animal studies in setting guidance for the nation and the world. This is most notable for ionizing radiation where human cohort (prospective) studies have been used for over 40 years in setting guidance to protect the public and workers.

The high quality human studies of population exposures to RF waves are the most relevant and they are uniformly "negative," i.e., little to no evidence of brain tumors linked to mobile phone use. The studies include those in Denmark of 400,000 cell phone subscribers and the U.K. of one million women. Studies in children are also “negative”, and there is no evidence that the occurrence of brain tumor has increased during the past 20 years in the United States, U.K. and Nordic countries.

Finally, the "possible" carcinogenic classification by the IARC, which was not unanimous, is not the same as a "probable" carcinogen, and I suspect this study would have minor impact on upgrading this classification. "Possible" classification is based on very weak and potentially biased evidence, and also includes coffee and pickled vegetables.

This is the best experiment yet conducted and NTP certainly should be congratulated on such a massive endeavor. Nonetheless, serious uncertainties exist that limit interpretation and the applicability to humans. We should stay tuned for the fuller reports yet to come.