Tarone, Robert E.; Inskip, Peter D.
International Epidemiology Institute, Rockville, Maryland, firstname.lastname@example.org (Tarone)
Radiation Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland (Inskip)
To the Editor:
Lönn et al1 state that a method of laterality analysis used in previous studies of cellular telephones and brain tumors “assumes that tumors are equally likely to occur on the left and right side in the absence of mobile phone exposure.” This is incorrect. The significance test for the 2 × 2 case-only laterality table used in prior studies2–4 is valid even if tumors are more likely to occur on 1 side of the brain in the absence of exposure.2 The square-root of the odds ratio from the 2 × 2 case-only laterality table estimates the ratio of the acoustic neuroma risk on the telephone-exposed side of the brain to that on the unexposed side. This laterality risk ratio is a measure of the tumorigenic potential of cellular telephone use. The assumption that tumors are equally likely on the left and right in the absence of exposure is only required to transform this laterality risk ratio to the relative risk in cellular telephone users compared with nonusers.2
The laterality analysis in the Swedish study1 begins by dividing cases into a left-sided group and a right-sided group based on the acoustic neuroma location. Each control is then randomly assigned to a case group, with ipsilateral telephone exposure defined with respect to the tumor side in each group. Lönn et al report that 59% of the acoustic neuromas were right-sided tumors; however, they do not report the laterality distribution assumed in their control assignment. Nine cases were excluded from their laterality analysis without explanation, possibly because there were cases with bilateral acoustic neuromas. In the National Cancer Institute study,2 there were 4 cases with bilateral tumors (all neurofibromatosis patients); 3 had never used a cellular telephone, and 1 was a regular cellular telephone user.
The reporting by Lönn et al of only the numbers of cases with long-term exposure to summarize study size in their Table 4 exaggerates the difference between the Swedish and Danish Interphone studies. In the Danish study,4 there were 15 controls who had used a cellular telephone for 10 or more years compared with 26 controls in the Swedish study.1
Table 4 of Lönn et al highlights an odds ratio estimate from the National Cancer Institute study of 1.9 for more than 5 years of cellular phone use; of the 5 cases who had used cellular telephones for more than 5 years, 2 had tumors on the usual side of cellular telephone use, 2 had tumors on the opposite side of usual cellular telephone use, and 1 had bilateral tumors and was a right-sided cellular phone user. Thus, consideration of laterality of cell phone use does not lend support to a causal interpretation of the elevated odds ratio estimate.
Robert E. Tarone
International Epidemiology Institute, Rockville, Maryland, email@example.com
Peter D. Inskip
Radiation Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland
1. Lönn S, Ahlbom A, Hall P, et al. Mobile phone use and the risk of acoustic neuroma. Epidemiology. 2004;15:653–659.
2. Inskip PD, Tarone RE, Hatch EE, et al. Cellular-telephone use and brain tumors. N Engl J Med. 2001;344:79–86.
3. Muscat JE, Malkin MG, Shore RE, et al. Handheld cellular telephones and risk of acoustic neuroma. Neurology. 2002;58:1304–1306.
4. Christensen HC, Schüz J, Kosteljanetz M, et al. Cellular telephone use and risk of acoustic neuroma. Am J Epidemiol. 2004;159:277–283.
© 2005 Lippincott Williams & Wilkins, Inc.