Do Cell Phones Cause Brain Tumors? Another Piece of the Puzzle

—You can’t blow an uncertain trumpet.

     Reverend Theodore Hesburgh,  Time magazine

No greater puzzle exists in epidemiology than trying to determine what causes diseases like cancer. It is a puzzle with metaphysical roots that can be traced back to the ancient Greeks. And it is a puzzle with great complexity that tests our ingenuity and skill. We work so hard to solve it because its solutions can lead us to preventive recommendations.

The relationship between cell phones and brain tumors is just such a puzzle. Cell phones have been in use for a scant 40 years, exploding in prevalence from zero to billions of users (1). If their use causes cancer, we need to know. We need to solve this puzzle.

The philosopher Thomas Kuhn, who brought us scientific revolutions, paradigms, and “normal science,” wrote about the puzzle-like nature of scientific problems (2). Kuhn compared the scientist’s problems with jigsaw and crossword puzzles. The commonality is that all puzzles have solutions. There are differences to be sure, but according to Kuhn, all puzzles “test (our) ingenuity or skill in solution (2).” It follows that we devote ourselves to scientific puzzles because we are assured that if we are skillful enough, we will succeed in solving them.

Kuhn was right. We know how to solve the causation puzzle. We have done it hundreds of times using methods that have served us for decades and continue to improve. Just look at the lists of known carcinogens compiled by the National Toxicology Program or the International Agency for Research on Cancer (IARC) (3,4). These lists show that sometimes the solution to our fundamental puzzle is that the evidence is sufficient to determine that an exposure causes a specific cancer. As evidence of this fact, there are more than 180 established exposure–cancer relationships in the IARC list alone.

These same lists also tell us that sometimes the solution is exactly the opposite, that an exposure does not cause cancer. There are many examples in between these extremes, including the relationships between cell phones and brain tumors. Today, the consensus of the scientific community, with a few vocal dissenters, is that it has not been established that cell phones cause brain tumors regardless of how that broad category is defined, whether as gliomas, meningiomas, pituitary tumors, or acoustic neuromas. Perhaps things will change as new research is published. Perhaps not. The study by Schuz et al. (5) published in this issue of the Journal does not change things very much. But then any single study is unlikely to do so given that at least 50 studies of various designs have been published to date. Nevertheless, that fact does not diminish the importance of this update of a large epidemiological study (6,7).

In this latest analysis of the UK Million Women Study, the authors found no statistically significant associations between ever users of cell phones and never users for all brain tumors, glioma, meningioma, pituitary tumors, and acoustic neuroma (5). The authors reasonably conclude that their findings “support the accumulating evidence that cellular telephone use under usual conditions does not increase brain tumor incidence.” These findings confirm those in the earlier report of this same cohort for all brain tumors, gliomas, and meningiomas and contradict the earlier findings for acoustic neuroma. In sum, the results of this study push things a little closer to the consensus view on causation.

A methodologic message to be gleaned here for all those who attempt to predict future “positive” results is that some nonassociations remain intact, and some observed associations disappear as time passes and events accrue in epidemiological studies. Schuz et al. (5) did not fall prey to the temptress who whispers in the ears of those who opine that “more research is needed to confirm” a finding, for example, that cell phones appear to be carcinogenic (8), without providing the alternative possibility that maybe nothing will change or perhaps a refutation such as the one observed by Schuz et al. will emerge.

Not everyone agrees that causal relationships between cell phones and brain tumors remain uncertain (9-12). There is not enough space in this editorial to comment but briefly on the claims of this minority view. It is clear from their published accounts, however, that they have tread indelicately on methodologic norms. These authors have not provided systematic assessments of the available evidence, a methodologic requirement in all aspects of oncology (13), and some have used a legitimate method of causal analysis in contrary ways (14).

On the other side of the aisle are those who write that a causal relationship between cell phones and brain tumors is possible at best, borrowing the “2B” designation from the IARC’s approach to causation, which has no quantitative interpretation and is a far cry from an established carcinogenic relationship (15). At least 2 systematic reviews with meta-analyses and 5 narrative reviews from this camp have been published since IARC made its decision in 2011 (16). Alexiou and Sioka (17) call for further studies, Yang et al. (18) note that the evidence for glioma is of “poor quality” and “limited quantity,” and Ostrom et al. (19) conclude that the evidence does not support an association. Magiera and Solecka (20) conclude that the relationships lack clear evidence. Röösli et al. (21) write that “epidemiological studies do not suggest increased brain or salivary gland tumor risk with MP use,” and Vienne-Jumeau et al. (22) write that “the causal factors of brain tumors remain unknown.” Finally, Choi et al. (23) point out that “there was no significant association between cellular phone use and tumor risk in most (tumor-specific) subgroup analyses.” Choi et al. (23) also note that the studies with an increased tumor risk were the studies authored by the investigators who, it turns out, are in the “we are certain it is causal” camp.

In other words, authors of positive studies are some of those making the strongest pitch for causality. How can this be? Thirty years ago, Wynder (24) warned us that investigators will insist that their evidence is the only true explanation. He called it “wish bias,” but it is also known as “confirmation bias” (25). Alternatively, those who advocate for causality could be an example of “white hat bias,” where investigators distort research-based information in the service of what they perceive to be righteous ends (26). A third possibility is that, in the end, these investigators will prove to be correct. Perhaps someday the scientific community will decide that cell phones cause some form of brain tumor. But predicting future results and conclusions of scientific research is as futile as predicting the next scientific revolution.

The causation puzzle has its promise, its complexities, and uncertainties. Indeed, uncertainty runs through our world like a river. Nevertheless, Schuz et al. have provided us with another piece of a most complex puzzle in a well-designed study with valuable results. As a result, some portion of our inescapable uncertainty has been reduced. We must wait to see if the promise of prevention will ever be fulfilled.

Funding

None.

Notes

Role of the funder: Not applicable.

Disclosures: The author has no disclosures. Douglas L. Weed, MD, MPH, PhD, who is a JNCI Editorial Board member and the author of this editorial, was not involved in the editorial review or decision to publish this editorial.

Data Availability

No new data were generated or used for this editorial.