Earlier this year, Australia saw a whirlwind tour from the electromagnetic radiation from mobile phones alarmist Devra Davis. Davis is an international champion of the belief that populations bathed in radiation emitted by mobile phones face epidemics of disease – particularly brain cancer.
Davis’ concerns were the focus of an ABC Catalyst program which attracted widespread criticism, including from me and Media Watch. The Catalyst presenter Maryanne Demasi was nominated for the Australian Skeptics bent spoon award.
At the time of the Catalyst program for which I declined to be interviewed, I had my hands tied behind my back because, with colleagues in cancer research, I had a paper in preparation examining the possible association between the incidence of brain cancer in Australia and the inexorable rise of mobile phone use here over the last three decades. Releasing our findings would have jeopardised publication, we could say nothing about what we had concluded.
Today the paper is published in early view in Cancer Epidemiology. Here’s what we set out to examine and what we found.
We examined the association between age and gender-specific incidence rates of 19,858 men and 14,222 women diagnosed with brain cancer in Australia between 1982-2012, and national mobile phone usage data from 1987-2012.
In summary, with extremely high proportions of the population having used mobile phones across some 20-plus years (from about 9% in 1993 to about 90% today), we found that age-adjusted brain cancer incidence rates (in those aged 20-84 years, per 100,000 people) had risen only slightly in males but were stable over 30 years in females.
There were significant increases in brain cancer incidence only in those aged 70 years or more. But the increase in incidence in this age group began from 1982, before the introduction of mobile phones in 1987 and so could not be explained by it. Here, the most likely explanation of the rise in this older age group was improved diagnosis.
Computed tomography (CT), magnetic resonance imaging (MRI) and related techniques, introduced in Australia in the late 1970s, are able to discern brain tumours which could have otherwise remained undiagnosed without this equipment. It has long been recognised that brain tumours mimic several seemingly unrelated symptoms in the elderly including stroke and dementia, and so it is likely that their diagnosis had been previously overlooked.
Next, we also compared the actual incidence of brain cancer over this time with the numbers of new cases of brain cancer that would be expected if the “mobile phones cause brain cancer” hypothesis was true. Here, our testing model assumed a ten-year lag period from mobile phone use commencement to evidence of a rise in brain cancer cases.
Our model assumed that mobile phones would cause a 50% increase in incidence over the background incidence of brain cancer. This was a conservative estimate that we took from a study by Lennart Hardell and colleagues (who reported even higher rates from two studies). The expected number of cases in 2012 (had the phone hypothesis been true) was 1,866 cases, while the number recorded was 1,435.
Using a recent paper that had Davis as an author we also modelled a 150% increase in brain cancer incidence among heavy users. We assumed that 19% of the Australian population fell into this category, based on data from the INTERPHONE study an international pooled analysis of studies on the association between mobile phone use and the brain. This would have predicted 2,038 expected cases in 2012, but only 1,435 were recorded.
Our study follows those published about the United States, England, the Nordic countries and New Zealand where confirmation of the “mobile phones cause brain cancer” hypothesis was also not found.
In Australia, all cancer is notifiable. At diagnosis, all cases must by law be registered with state registries tasked with collecting this information. It has been this way for decades. So we have excellent information about the incidence of all cancers on a national basis.
The telecommunications industry of course also has information on the number of people with mobile phone accounts.
While touring Australia, Davis was confronted with the “flatline” incidence data on brain cancer. Her stock response was that it was far too early to see any rise in these cancers. She was here to warn us about the future.
However, prominent Sydney neurosurgeon Dr Charlie Teo would appear to disagree about it being too early. He told Andrew Denton on ABC-TV’s Enough Rope in 2008:
If you look at the science on mobile phones and the link with brain cancer, it is quite compelling … we know that radiation causes cancer, but it takes about ten years for it to develop, so we know that EMR electromagnetic radiation is going to take at least ten years to create brain tumours and possibly longer fifteen, twenty years.
In cancer epidemiology, the concept of the latency (or lag) period is well known. This refers to the time that it takes between initial exposure to a potentially carcinogenic agent (like cigarette smoke, asbestos, or nuclear radiation) and excess cases of cancers of interest to appear.
Davis would appear to be arguing that we would see a sudden rise many years later. That is not what we see with cancer; we see gradual rises moving toward peak incidence, which can be as late as 30-40 years (as with lung cancer and smoking).
For example, as I showed in a recent Conversation piece, this paper also reports on central nervous system cancers (including brain cancers) in those exposed to atomic bomb radiation in Japan in 1945. This graph shows 110 of 187 cases (58.8%) were diagnosed in the first 40 years (before 1985) (so before 40 years).
The incidence and type of cancers of those exposed to atomic bomb radiation varied over the years. And this quote from the methods section shows that there were another 27 who died before 1958 from central nervous system cancers, within 13 years of the bombs.
We excluded 73 tumours in individuals who were not in Hiroshima or Nagasaki at the time of the bombings, 35 individuals who did not have available organ dose estimates, and 27 individuals who died or were diagnosed before January 1, 1958.
Note here that A-bomb survivors were affected by ionising radiation (that is, radiation of sufficient energy to produce ionisation). This is where the energy is strong enough to remove electrons off their atoms or molecules, including causing DNA damage. Mobile phones produce non-ionising radiation which is low energy, sufficient only to ‘excite’ the electrons enough to make them just heat up.
We have had mobiles in Australia since 1987. Some 90% of the population use them today and many of these have used them for a lot longer than 20 years. But we are seeing no rise in the incidence of brain cancer against the background rate.