A lesson in how easy it is to be misled Little is known about the causes of brain cancer and brain tumors. Exposure to ionizing radiation (X-rays) and certain rare genetic syndromes have been identified as causes, but these account for only a small proportion of cases. Since the early 1990s the possibility that exposure to radiofrequency (RF) radiation from wireless communications devices may be causing an increase in brain cancer has become a public concern, and this possibility was given credence in 2011 when the International Agency for Research on Cancer (IARC), an arm of the World Health Organization, classified cell phone RF energy as a “possible carcinogen.” In addition, a research group in Sweden has been publicizing results that appear to indicate that long-term cell phone use is associated with a 2-3-fold increased risk of brain cancer. These results contrast with the vast majority of epidemiologic studies, which show little evidence of an association. However, the frightening results from this single group tend to get much more attention than the much larger body of studies which show no association. In this situation, a crucial question is: Has the incidence of brain cancers – and of specific types – been increasing over recent decades – as the use of cell phones has surged? This would appear to be a simple question, but, as I explain below, it is anything but. In response to a column I wrote regarding the California Department of Public Health’s recent advisory on cell phone safety, I received a number of letters/tweets asserting that the incidence of brain cancer, and particularly of glioma and glioblastoma – the most deadly type – has been increasing in Denmark, Sweden, the Netherlands, UK, USA, and Australia. In view of the concern regarding cell phone use and brain cancer, determining whether there is, in fact, a clear increase in brain cancers over recent decades is an important question, with implications both for public health (if answered in the affirmative) and for the public’s peace-of-mind (if answered in the negative). To attempt to answer this question, I went into PubMed and Google to find any articles I could in medical journals addressing change in brain cancer incidence rates in advanced countries in recent years. Before getting to what the articles show, a few basic points need to be kept in mind. 1) Brain cancers are rare. In the US, the incidence (that is, the number of newly diagnosed cases per year per 100,00 population) of all brain cancers is 6 cases per 100,000. (By contrast, the incidence of breast cancer is 125 per 100,000, and that of prostate cancer is 120 per 100,000). 2) Approximately 80 percent of brain cancers are gliomas. There are a number of distinct cell types within gliomas, but the main distinction is between low-grade gliomas and glioblastomas -- the most common and most deadly type in adults. 3) Many different organizations track the incidence of gliomas by means of surveillance (country-wide, state-wide) or health system records. However, the lack of a consistent definition of glioma as well as differences in data collection techniques make it difficult to compare rates from different sources (Ostrom, 2014). 4) People who work with cancer statistics are well aware that improvements in diagnosis and changes in the classification of cell types can result in artifactual changes in cancer rates. So, the first question that specialists ask when they notice a significant change in rates is, could this be an artifact of change in diagnostic practice? More about this in a moment. 5) Since brain cancers are very rare and gliomas are even rarer, in small populations the rates are likely to be unstable, and one may see fluctuations that are merely due to chance. 6) Finally, the time period over which change in incidence is assessed will influence the results. Looking at the relevant articles, one finds that the incidence of all brain cancers has been remarkably stable over a number of decades in various countries: U.S. (slight decrease – 1992 to 2014), Australia (stable - 1982-2014), Ireland (stable - 1994-2013), New Zealand (stable – 1995 to 2010), and Taiwan (decrease from 1999 to 2012). A similar point is made regarding gliomas in a 2014 article in Neuro-Oncology entitled “The epidemiology of glioma in adults: a ‘state of the science’ review,” by Quinn Ostrom, a cancer epidemiologist at Case Western Comprehensive Cancer Center, and coauthors. The authors wrote: “Many analyses have examined the incidence rates of glioma to assess whether rates are increasing. The results of these have generally shown the incidence of glioma overall and glioma subtypes to be fairly stable over the time periods assessed. An examination of the annual age-adjusted incidence in Nordic countries between 1979 and 2008 found no clear trend in glioma incidence rates during this period, though there as a slight increase in brain tumor incidence rates overall. In an analysis of data from 12 Surveillance, Epidemiology, and End Results cancer registries [of the National Cancer Institute in the U.S.] between 1997 and 2008, no significant trend in incidence rates of all gliomas was found overall, although a slight decrease in incidence of low-grade gliomas was observed. An analysis of Israeli brain tumor incidence found a significantly decreasing trend in incidence rates of low-grade gliomas between 1980 and 2009.” One of the countries in which correspondents asserted that there was an increase in glioblastoma is the Netherlands. Dutch researchers reportedthat the incidence of astrocytomas (the most common type of glioma) has remained fairly stable from 1989 to 2010. However, the incidence of glioblastoma has more than doubled – from 1.5 per 100,000 to roughly 3.4 per 100,000. But before we jump to interpret this change as evidence of a real increase in incidence of glioblastoma over this time period, we need to consider changes in the classification of these entities. The authors point out that in 2000 the World Health Organization, which publishes the handbook for the classification of diseases, revised its definition of glioblastoma by including the category of anaplastic astrocytoma, which formerly had been a separate category. Another change was made in 2007, further expanding the definition of glioblastoma to include anaplastic oligoastrocytoma. Thus, rather than providing evidence of an increase in the incidence of this most fatal type of brain cancer, it appears that the observed increase in glioblastoma may be due to changes in the classification scheme. And this change could clearly affect the rates reported in other countries as well. If one believes that RF radiation from cell phones is promoting the development of brain cancer, one’s interpretation of the evidence – whether it be studies in humans or in experimental animals, or changes in cancer rates over time - - can be affected. We are all susceptible to accepting evidence that supports our beliefs. For this reason, findings such as those contained in the Dutch paper, can get latched on to and passed around as providing “hard evidence” that confirms our fears. The scientific evidence is often inconclusive and cannot provide definitive answers regarding factors that may affect our health. However, as the example of brain cancer rates shows, where there is high-quality data available (as exists tabulating cancer rates in advanced countries), care is needed to understand the source of the data and how it is to be interpreted. The fact that the incidence of gliomas or other brain cancers shows no increase in recent decades does not prove that RF has no effect on this disease; however, it does provide important information that weighs against the hypothesis. Source
