Saturday, October 18, 2008

Deadly Devices? The disputed malignancy risk of mobile phones

Last month, over the course of three days and two emergency department shifts, I diagnosed two patients with brain cancer. In and of itself, this was unusual; brain cancer is rare, occurring only in approximately 6 out of 100,000 people. But what made these diagnoses particularly striking was that both of these patients were in their 40's. Stan* was 49, a Fairfield-area beer distributor who was perfectly well, other than drinking a little more of his product than he should have. A week earlier, he had noticed some tingling in his left arm and hand and slight clumsiness – symptoms he attributed to having a crick in his neck. A few days later, his arm started shaking uncontrollably. A Cat scan (CT) of his brain showed a mushroom-top sized lesion in the right side of his brain, with a smaller satellite lesion nearby. Each lesion was surrounded by a dark ring of swelling (edema), and I guessed that this was probably what was causing Stan's seizure activity. On the phone, I reviewed the scan with our neurosurgeon. “Looks like a glioblastoma,” he said, adding “with a crappy prognosis.” I broke the news to Stan as gently and thoroughly as I could and offered him some short-term relief – a prescription for steroids that would decrease the swelling in his brain and hopefully halt the seizures. I answered his questions as best I could, although I didn’t know how to address his final one. “Heh doc, I spend a lot of time on my cell. Do you think that could have caused this?”

A day later, I treated Matt*, a 42-year-old father of three, who was completely healthy until he suddenly lost the ability to formulate speech. By the time I saw him, his speech had returned and he was ready to blow off the entire incident – he’d smoked marijuana just before his symptoms set in and figured he’d tripped out. I noticed that the right side of his face looked a little asymmetric and that his smile seemed somewhat crooked. I asked his wife about this, but she was nonplussed – he’d looked like that for as long as she’d known him. I talked Matt into getting a CT scan, and was shocked by the images – he had a large egg-sized mass in his left brain. “I’ve got another ugly one,” I told our neurosurgeon and he agreed. “Hard to say what type that is, but the prognosis is not good.” When I walked in to tell Matt the awful news, I saw his kindergarten-aged daughter, with shiny black hair and overalls, and I almost lost my composure. This girl was not much older than my own daughter and I was about to tell her father that he had a large brain tumor. At my request, the young girl left the room with her mother, skipping contentedly, oblivious to the fact that her life was about to drastically change. Matt took the news stoically and asked all the appropriate questions. It wasn't until I showed him the images from the CT scan on a hallway monitor that he began to tear up and shake. An hour later, as we were discharging him, Matt asked me if I thought there was anything specific that could have caused this. "Well, probably not," I replied, "but let me ask you this, do you spend much time on your cell phone?" "Yes," he said, "all day long."

Now, as a physician and part-time medical researcher, I am well aware of the danger of extrapolating large-scale causality from the circumstances of one or two patients. Intellectually, I knew that just because Stan and Matt were both heavy cell phone users, and both developed brain cancer at an early age, did not mean that cell phone use causes cancer. But, it seemed like too much of a coincidence not to at least investigate the question.

For over a decade, researchers have searched for just this sort of connection: the chance to say that brain cancer can be caused by exposure to increased levels of electromagnetic radiation (EMR) from cellular phones. It is well known that high levels of electromagnetic radiation can cause severe symptoms such as blindness, burns, and even death. Lower levels of EMR are associated with headaches, auditory and sleeping disturbances, and short-term memory loss. EMR has also been implicated in cellular and DNA changes that may be precursors to cancer. And while multiple studies have failed to find a connection between cell phone use and brain cancer, research continues, and at least one respected oncologist has sounded the alarm.

Last June, Ronald B. Herberman, head of the prestigious University of Pittsburgh Cancer Institute, sent a memo to 3,000 faculty and staff warning them of a risk of cancer from the low levels of EMR emitted by cellular phones. Citing early and unpublished data, Herberman called for across-the-board caution with cell phone use, especially in children. According to a recent article in The Economist, it is thought that Herberman was basing his recommendations on the preliminary findings of the multi-national "Interphone study". This investigation, involving 13 countries, 14,000 subjects and spanning 2000-2006, has yet to formally release its overall findings. To date, however, nine of the thirteen individual national research groups have published their work - without evidence of a cell phone-cancer link. On the other hand, an outside researcher from Sweden (Lennart Hardell), who grouped the analysis of the Interphone study with nine other studies found an increased risk of cancer in people using their cell phones for more than an hour a day for ten years or more.

Having read a bit about the potential methodological weaknesses of the Interphone study (and similar studies), I am skeptical about whether these results are meaningful. Interphone's design suffers from an excess of bias. Bias, which might be healthy in some circumstances (such as parenthood), is not good for research projects. Interphone identified patients with the outcome of interest (brain cancer) and quizzed them, after the fact, about cell phone use. This group was compared to another group of healthy, cancer-free patients who also answered questions about cell phone use. It turns out that the cancer-free patients selected had a much higher cell phone use (nearly twice that) of comparables in the general public. This flaw is called selection bias and in the case of Interphone would be expected to (artificially) deflate the observed cancer risk of cell phone users. A second Interphone bias, common to many retrospective (backwards looking) trials is recall bias. One might expect that subjects diagnosed with brain cancer would be more likely to over-report previous cell phone use. Finally, there is a generalization bias; many study participants were asked about cell phone use in the 1990s - at a time when most phones were analogue rather than digital and therefore resulted in different types of EMR exposure (analogue phones usually have higher wattage than digital). Any results in these patients could not be reliably applied to the current population of over 3 billion digital cell phone users.

So it makes one wonder, what exactly does Dr. Herberman know that we don't? If he is right about the potential health risks of cell phone usage, especially long-term cell phone usage, we may have a big ol' brain cancer epidemic on our hands in a few years. If he is wrong, this scare will join a long list of other associations, such as immunizations and autism, that never panned out. In the meantime, I am hedging my bets. I will limit my own cell phone use to less than an hour a day and keep my mobile as far away as possible from my daughter's developing brain. And I will hope that by the time she asks, or perhaps demands, a phone of her own, quality research will have settled the question.

*Names and indentifying information altered to protect patient privacy