Overall data are so iffy, however, that researchers remain unsure of cell safety

By Janet Raloff

Web edition: May 17, 2010

Enlarge

What's the risk?

People who talk long and frequently -- and hold cells phones to their ear while they do so -- may face an elevated risk of developing gliomas, a serious type of brain cancer, according to a new study.

Credit: iStockPhoto

A major decade-long international study concludes that, overall, cell-phone users show no increased risk of developing brain tumors. The same study reports that among people who have used cell phones the most and longest — for at least 10 years and on average 30 minutes or more a day — risk of brain tumors is substantially elevated when compared to people who don’t use cell phones.

But the real enigma: For people in each of the lower cell-phone-use categories, tumor risks were substantially lower than those seen in people who used regular, corded phones. In other words, for most people cell-phone use appeared to protect against brain tumors.
 
The generally contradictory findings — apparent protective effects at most doses and elevated cancer risk at the highest exposure — point to the challenge scientists have had in figuring out what to make of data collected as part of the Interphone study. Participants were recruited in 13 countries (all outside of the United States) and included 7,416 tumor patients and almost twice that many controls.

Although the researchers analyzed risks for two types of brain tumors, only data linking heavy cell-phone use to  gliomas appeared due to something other than chance. Moreover, even this association was hardly iron-clad. Based on the reported 95-percent confidence interval, the chance these cancers might have been linked to cell-phone use could be as small as 3 percent or as high as 89 percent.

Here the statistics appeared stronger, with a protective effect for both tumor types — gliomas and meningiomas — in the range of 10 to 25 percent, depending on the exposure category. (And the confidence intervals indicated that the likelihood the effects were real ranged from 2 to 50 percent, again depending on recalled estimates of cell-phone exposures).

Explains Interphone researcher Siegal Sadetzki, a public health physician at Tel Aviv University’s Sackler School of Medicine, “If you look at the overall evidence, this study did not confirm or dismiss the possible association between cell phones and brain tumors. That’s the bottom line.”

Science requires data to meet “very strong criteria before you can say there is an association,” she explains. And the Interphone data that were reported online May 17 in the International Journal of Epidemiology, did not meet those criteria, she says.

“On the other hand,” she adds, “we do we see a few indications of risk. And these indications appear among people who were exposed for the longest duration. We do see an association with ipsilateral use [tumors on the same side of the head that a user holds a cell phone to the ear]. We also see an association with temporal lobe [brain] exposure. So there are some indications of a positive association in these subgroups.”

As a result, she says, “We do have some suspicions.”

Protective effects 'can't be real'
The paper’s authors acknowledge that the apparent brain-tumor protection afforded most of the 21,770 Interphone participants doesn’t make sense. 

David Carpenter, who heads the State University of New York at Albany’s Institute for Health and the Environment in Rensselaer, N.Y., similarly finds “perplexing” that apparent protective effect of cell-phone use for all but the longest, heaviest users. In fact, he says, “This cannot be real and probably is a reflection of some flaw in the design of the study,” one that he says “results in an artificial lowering of the reported risk.”

The study concedes this is a possibility.

If the effect of cell-phone use on tumor risk was zero, it should yield a risk value of 1.0 — equivalent to that assigned to the control group of non-cell users. Any risk number below 1.0 suggests a protective effect of the exposures. The fact that computed tumor risks fell below 1.0 for all cell users except those in the highest-use category “could be taken to indicate an underlying lack of association with mobile phone use, systematic bias from one or more sources, a few random but essentially meaningless increased odds ratios [calculations of risk], or a small effect detectable only in a subset of the data,” the Interphone authors write.

Indeed, Sadetzki says of the below-1.0 risk that Interphone found for virtually all cell-phone users: “We think this is not a true thing. So this would suggest we have an underestimation of the risk.”

Yet “even under these circumstances [the authors] find a clear elevation in risk of brain cancer with prolonged use,” Carpenter points out, especially for gliomas and tumors that occur on the same side of the head as a user typically holds his or her phone. And “this conclusion is exactly what has been reported in the earlier studies,” he observes.

As such, he contends, the paper’s general claim that there is no increased brain-tumor risk among cell users is “certainly cautious, and in my judgment excessively cautious.”

Strong suspicion of hidden biases
The new paper, written by a committee of Interphone researchers across the world, admits that biases and errors may have limited the strength of the study and prevented a causal interpretation.

Those qualifications were “added at the end of the editorial process of revision” and are “both elegant and oracular,” argue Rodolfo Saracci of the National Research Council in Pisa, Italy, and Jonathan Samet of the University of Southern California in Los Angeles. They suggest that Interphone’s authors attempted to finesse their interpretations in a way that would not unduly scare cell-phone users — even if their findings didn’t warrant such caution.

Writing in an editorial that accompanies the new paper, the two point out that “None of today’s established carcinogens, including tobacco, could have been firmly identified as increasing risk in the first 10 years or so since first exposure.” Tumors among the Interphone study’s participants were diagnosed between 2000 and 2004 — even though wide-scale cell-phone use got underway only in the mid-1990s. So fewer than 5 percent of meningiomas and 9 percent of gliomas occurred among people who had used cell phones 10 years or more.

Moreover, Saracci and Samet observe, the apparent protective effect computed in this study is not statistically “plausible.” They argue, therefore, that “bias stands as the most likely explanation of the observed results.”

They probed a few of the types of biases to which the data appeared susceptible and concluded that if these have occurred, they would likely have served to diminish the apparent tumor risk — even amongst people in the highest-exposure group.

For now, Saracci and Samet say, Interphone “tells us that the question as to whether mobile phone use increases risk for brain cancers remains open.”

More studies needed
The study’s authors, too, acknowledge that the jury is still out on cell safety, which is why they recommend further investigation of “possible effects of long-term heavy use of mobile phones.”

That’s a good idea, Saracci and Samet say, since the lower end of the high-use group studied by Interphone were people averaging only a half-hour of calling per day. That’s well below usage patterns for many people in our increasingly cell-dominated society, they say — one that’s populated by an estimated 4.6 billion mobile-phone users globally. Indeed, many people have begun jettisoning land lines for cell-phone-only service.

In a news release, the Mobile Manufacturers Forum, which represents aspects of the cell-phone and wireless industry, said “The mobile industry supports the need for ongoing research,” and pointed to several studies that will be following users in coming decades. Some will even begin investigating risks to children and teens. MMF helped fund the Interphone study.

Until any followup data on heavy users come in, Sadetzki recommends that cell owners adopt "the precautionary principle:" Assume some risk might exist and therefore limit exposures. Tactics might include avoiding long calls, sending text messages instead of voice messages (that require both parties put the phone to their ears) and using a Bluetooth or other hands-free device to keep a phone away from close proximity to the head.