Cell Phones and Brain Tumors: No Link, But Is Study Flawed?

From Medscape Medical News > Oncology Roxanne Nelson October 21, 2011 — The latest study on cancer and cell phones — the largest to date — has found no evidence of an overall increase in brain tumors or any cancers over an 18-year period. However, a group of experts says that the study is seriously flawed, and declares that it should be "condemned as misleading spin." The study was published online October 20 in BMJ, and the fierce rebuttal comes from ElectromagneticHealth.org. Updated Results The study is an update of a nationwide Danish study that compared the cancer risk for all 420,095 Danish cell-phone subscribers with that for nonusers from 1982 to 1995, with follow-up to 1996 (J Natl Cancer Inst. 2001;93:203-207) and then 2002 (J Natl Cancer Inst. 2006;98:1707-1713). This study has found no evidence of any increased risk for brain or central nervous system (CNS) tumors or any cancer among cell-phone subscribers, and the latest results — which now span a period of 18 years — confirm this finding of no evidence. The researchers, led by Patrizia Frei, PhD, a postdoctoral research fellow from the Institute of Cancer Epidemiology, Danish Cancer Society, Copenhagen, extended the follow-up of the Danish cohort to 2007, and focused on CNS tumors in 358,403 cell-phone subscribers. Overall, 10,729 CNS tumors occurred from 1990 to 2007 in 5111 men and 5618 women. The incidence rate ratio for CNS tumors was consistently close to 1 in women and men, both overall and when stratified. However, when analyzed by morphologic subtype of intracranial CNS tumor, there was a slight but nonsignificantly elevated incidence rate ratio of glioma in men (1.08; 95% confidence interval [CI], 0.96 to 1.22). The ratio was highest in the shortest-term users (1 to 4 years); after 5 years of use, the numbers were only slightly elevated. For other and unspecified types intracranial tumors of the CNS, incidence rate ratios were nonsignificantly higher in both men (1.12; 95% CI, 0.95 to 1.33) and women (1.19; 95% CI, 0.85 to 1.67), but the authors did not see a "clear indication of a dose–response effect." The authors note that this "extended follow-up allowed us to investigate effects in people who had used [cell] phones for 10 years or more, and this long-term use was not associated with higher risks of cancer." However, they point out that because "a small to moderate increase in risk for subgroups of heavy users or after even longer induction periods than 10 to 15 years cannot be ruled out, further studies with large study populations, where the potential for misclassification of exposure and selection bias is minimized, are warranted." More Research Needed? The study has 2 important methodologic advantages over most other studies, write Anders Ahlbom, PhD, and Maria Feychting, MD, PhD, both from the Karolinska Institute in Stockholm, Sweden, in an accompanying editorial. First, "it was based on a computerized cohort that was followed passively in registries, so it avoided the need to contact people," they point out. Thus, the problems of nonresponse and selection bias, which have been an issue in other studies, were eliminated. Second, the researchers used digitized subscriber data that were obtained from the operators, rather than from retrospective questionnaires or interviews with users. "This circumvented the recall bias that is present in other studies," the editorialists write, although they point out that "having a [cell] phone subscription is not equivalent to using a [cell] phone and, conversely, some users will be nonsubscribers." Although the study used a Danish cohort, the editorialists point out that in Sweden, where handheld cell phones were introduced almost 25 years ago, the incidence rates for glioma have not risen since 1970. The use of cell phones has spread quickly, and 87% of people 16 to 75 years of age were using them in 2002. The proportion that has been using them for 10 or even 15 years must have been substantial, note Drs. Ahlbom and Feychting. Thus, "the absence of a trend in the incidence of brain tumors in national statistics is reassuring." The research that has been conducted evaluating the safety of cell phones is now extensive, and "the question is how much more research is needed," they write. "Continued monitoring of health registers and prospective cohorts is warranted, but more case–control or other studies with built in selection and recall bias are not needed," the editorialists conclude. Fierce Rebuttal However, a group of experts from several countries have joined together and issued a fierce rebuttal of this study in a document posted on ElectromagneticHealth.org, a health education and advocacy group based in the United States. "This seriously flawed study misleads the public and decision makers about the safety of [cell] phone use. I consider that their claims are worthless," Denis L. Henshaw, PhD, emeritus professor of human radiation effects, University of Bristol, United Kingdom, states in the document. "From the way it was set up originally, this deeply flawed study was designed to fail to find an increased risk of brain tumors tied [to] cell phone use. In order for any study of a relatively rare disease like brain tumors to find a change in risk, millions must be followed for decades. By extending an earlier analysis on the same group of cell phone users, this new report provides unsurprising, biased, and misleading conclusions," explains Devra Davis, PhD, MPH, cancer epidemiologist and president of Environmental Health Trust, in the document. A serious concern about this study is the choice of individuals in the control group, the group of experts asserts. The Danish researchers compared the rates of brain tumors that occurred from 1990 to 2007 in those who began using cell phones after 1987 with the rates in those who were nonsubscribers when the study started. "This understates risk, because most of those who began as 'nonsubscribers' to cell phone service (i.e., the 'controls' at the time the cohort was collected) became cell phone users later on, and accumulated almost as many years (on average per person) as the 'exposed' subscribers. Hence, the comparison to the population not contained in the subscriber sample is a comparison between 2 exposed groups. When Michael Kundi and colleagues from the Medical University of Vienna mathematically corrected for this concern in an earlier report from this Danish study, they found a significantly increased risk for brain tumors," the group writes. This concern about the control group is raised in the editorial, which describes it as a "weakness" of the study. The misclassification of subscribers and nonsubscribers "would dilute any association between [cell-]phone use and cancer risk, and this is important for a negative study like the current one," they note. "However, for long-term users, this misclassification would have only a small effect: long-term users who did not hold personal subscriptions would make up a small proportion of the reference population," they assert. In their rebuttal, the experts declare that this updated Danish study "in fact did find increased risk, even though the study is currently being promoted to the media as if it did not." "Statistical significance tests are tools used in science to help understand the chance that a finding is real. In fact, the article reports a significant increased risk of a very rare form of glioma of the cerebral ventricle based on 8 cases (2.58; 95% CI,1.08 to 6.10), but the authors chose to make no mention of this significant finding. In this instance, despite the small number, the finding is significant," they write. "The authors reject all other findings of borderline significance completely. In a study of relatively rare diseases, such as brain tumors, the failure to obtain statistical significance should not be confused with a lack of public health importance. In fact, most of the reported numbers of brain tumors in this article give estimated risks where the result goes from below 1 (a negative result meaning no increased risk), to above 1 (a positive result indicating in some instances a doubled or greater risk)," they add. "All of the few well-designed case–control studies of this issue have found significantly increased risk. Thus, these borderline findings of increased risk may well signal an important association," they add. In the document, Alasdair Philips, an expert in electromagnetic health from Powerwatch in the United Kingdom, says: "This study only looks at 7% of the Danish population who had a personal cell-phone subscription for at least 1 year during the period 1987 to 1995. It ignores corporate subscribers (the heaviest users then), and the researchers have no data at all on cell-phone use since 1995, so the extra 86% of the population who started to use a cell phone since 1996 were left in the 'nonsubscriber' part of the population. This study uses seriously flawed data to make a flawed analysis and should be condemned as misleading spin." A Mixed Bag Previous studies on the link between cell phones and cancer have come to various conclusions. Approximately 30 epidemiologic studies have attempted to evaluate a possible association between cell-phone use and the risk for brain and salivary gland tumors. There have also been a number of experimental studies involving cell cultures and animal models. One meta-analysis found evidence linking cell-phone use to an increased risk for tumors, whereas another study showed that cell-phone use for as little as 50 minutes at a time appears to affect brain glucose metabolism in the region closest to the phone's antenna. Yet another study pointed to evidence that exposure to cell phones prenatally and early in life increases a child's risk of developing behavioral problems. Conversely, a recent study reported that children and adolescents who use cell phones do not appear to be at a higher risk for brain cancer. In 2009, a Scandinavian study failed to find substantial changes in brain tumor incidence among adults 5 to 10 years after the use of cell phones sharply increased. Results from the 13-country INTERPHONE project, the largest study ever conducted on cell-phone use and cancer risk, reported no increase in risk for glioma or meningioma with the use of cell phones. However, there were suggestions of an increased risk for glioma in people with the highest levels of exposure. Some countries have begun to take measures to limit cell-phone exposure in children. In June, the World Health Organization's International Agency for Research on Cancer announced that radiofrequency electromagnetic fields have been classified as possibly carcinogenic to humans (group 2B) on the basis of an increased risk for glioma that some studies have associated with the use of wireless phones. This study was funded by the Danish Strategic Research Council to cover costs for data linkage. The authors and editorialists have disclosed no relevant financial relationships. BMJ. Published online October 20, 2011. Abstract, Editorial