Savitz, David A.
From the Department of Epidemiology, University of North Carolina School of Public Health, Chapel Hill, North Carolina.
Correspondence: David A. Savitz, Department of Epidemiology, CB #7435, University of North Carolina School of Public Health, Chapel Hill, NC 27599-7435. E-mail: email@example.com
Epidemiology sometimes pursues hypotheses with strong prior probability. Some researchers might even suggest that studies would ideally proceed from firm biologic findings through increasingly complex levels of organization to free-living human populations. In practice the evolution of knowledge is rarely—if ever—so orderly.
When epidemiologists contend with a question driven by public health concerns, the foundation for a causal association is often fragmentary, and the likelihood of an effect is typically very small. This certainly applies to the potential carcinogenicity of emissions from cell phones (or mobile telephones as they are known in Europe). The initial epidemiologic data on this question have been reassuringly negative.1–3 The only demonstrated health risks are the obvious ones associated with inattention while driving.4 Aside from a few studies using questionable methods,5–6 the evidence on cell phones and cancer largely followed this script. That is, until the publication in this issue by Lönn and colleagues.7
These authors identify an increased risk of acoustic neuroma among persons in Sweden with 10 or more years of cell phone use. There was no association with use less than 10 years (and thus with the use of newer digital technology), consistent with a recent Danish study.8 Given the size and quality of the Lönn study, these results provide perhaps the strongest negative evidence for recent cell phone exposure and acoustic neuroma generated thus far. With regard to long-term use and an extended latency period, there is much less evidence; the Swedish study stands essentially alone. These data, at minimum, do not reassure us that there is no effect.
One striking aspect of their findings is that risk for long-term users was confined entirely to the side of the head on which the phone was most often used. This finding merits special attention—what distinguishes the side of the head used for listening to the telephone? At least 3 candidate explanations come to mind: 1) It is the side of the head that receives, by a large margin, more radiofrequency radiation from the antenna, consistent with a causal hypothesis; 2) It is the side that has been chosen for listening, perhaps because hearing is best on that side; 3) Because it is the side that is relied upon for hearing, subtle losses would more likely lead to medical attention.
As acknowledged by the authors, subtle losses in hearing (an early indication of acoustic neuroma), would be more readily detected on the side that is used for the cell phone. Depending on the prevalence of undetected tumors, such detection bias would result in a positive association for the preferred side. To the extent the side with superior hearing is used for listening, detection bias would instead tend to inflate the association for the nonpreferred side, counter to these results. In assessing the scenario of detection bias, it is not clear why this would be an issue for long-term but not short-term use. Perhaps only prolonged subtle hearing impairment leads to medical attention, tolerated for shorter periods of time. The completeness of diagnosis and sequence of events leading to diagnosis would need to be evaluated to address the plausibility of exposure-driven detection.
Taking into account all the evidence, how likely is it that long-term cellular telephone use increases the risk of acoustic neuroma? The prior likelihood was quite low, given the limited biologic support and the negative epidemiologic evidence regarding short-term use and brain tumors. Furthermore, the absence of association with short-term use does not appear to be consistent with a dose-response gradient based on duration. Even so, the plausibility that long-term cellular telephone use increases risk of acoustic neuroma has been increased by this study, albeit with considerable uncertainty. Studies in progress will (not might) add markedly to the evidence within a few years. Meanwhile, researchers have the challenge of conveying to a simultaneously skeptical and panicky public the nuances of evidence shifting from “very unlikely but highly uncertain” to “slightly more likely but still highly uncertain” –a concept that is difficult to appreciate even for researchers. This uncertainty regarding long-term use should not distract from the growing evidence, enhanced by this study, that neither acoustic neuroma nor brain tumors is associated with cell phone use of less than 10 years.
1. Inskip PD, Tarone RE, Hatch EE, et al. Cellular telephone use and brain tumors. N Engl J Med
2. Johansen C, Boice JD, McLaughlin JK, Olsen JH. Cellular telephones and cancer—a nationwide cohort study in Denmark. J Natl Cancer Inst
3. Muscat JE, Malkin MG, Shore RE, et al. Handheld cellular telephones and risk of acoustic neuroma. Neurology
4. Cohen JT, Graham JD. A revised economic analysis of restrictions on the use of cell phones while driving. Risk Anal
5. Hardell L, Nilsman A, Pahlson A, Hallquist A, Mild KH. Use of cellular telephones and the risk for brain tumours: A case-control study. Int J Oncol
6. Hardell L, Hallquist A, Mild KH, Carlberg M, Pahlson A, Lilja A. Cellular and cordless telephones and the risk for brain tumours. Eur J Cancer Prev
7. Lönn S, Ahlbom A, Hall P, Feychting M. Mobile phone use and the risk of acoustic neuroma. Epidemiology
8. Christensen HC, Schüz J, Kosteljanetz M, Poulsen HS, Thomsen J, Johansen C. Cellular telephone use and risk of acoustic neuroma. Am J Epidemiol
© 2004 Lippincott Williams & Wilkins, Inc.