The authors respond.

Abstract

We thank Lennart Hardell and Michael Carlberg for their comments,1 which gave us the opportunity to expand on our findings. Our study2 was initiated after publication of the Swedish Interphone study,3 which found an increased risk of acoustic neuroma after 10 years of mobile phone use, confined to ipsilateral use, but with uncertainties regarding recall bias when reporting side of phone use, and no association with cordless phones. In the current study, cordless phone use was included as a secondary objective because the output power is considerably lower than from mobile phones of technologies used during the relevant time period.4 Therefore, fewer details were collected about cordless phones, to keep the number of questions to a reasonable amount, and these did not include such details as laterality. The definition of the unexposed category “non-regular mobile phone use” was decided a priori based on the Interphone study,5 and it is what most previous studies of mobile phones have used.4 We have now conducted the analyses with the unexposed category defined as “non-regular use of mobile and cordless phones” as suggested by Hardell and Carlberg. For regular mobile phone use, the odds ratio was 1.04 (95% confidence interval = 0.68–1.60), and for regular use of cordless phones, 1.22 (0.80–1.85). Thus, both ORs became lower than with the a priori-decided definition of the unexposed category. The matched design excludes incomplete case-control sets from the analyses, which in stratified analyses makes number of cases and controls apparently fewer than in the overall analyses, as is seen also in other studies, for example, the Interphone study.5 This was explained in a footnote in Table 2, which should also have been added to Table 3. In addition, a small number (n = 9) of the regular mobile phone users did not report their preferred side of mobile phone use. In previous studies, laterality of phone use was asked about only once, and it is likely that cases arbitrarily chose a time point some time before diagnosis, as the disease impairs hearing. Table 4 shows odds ratios between acoustic neuroma and ever-use of regular mobile phones according to laterality of phone use, where the full laterality history was used to determine laterality, instead of the cases’ own decisions. Changes of the preferred side of phone use do not date further back in time than their start of mobile phone use, which would also have been the situation if cases had determined the time period for reporting laterality themselves. David Pettersson Institute of Environmental Medicine Karolinska Institutet Stockholm, Sweden Maria Feychting Institute of Environmental Medicine Karolinska Institutet Stockholm, Sweden [email protected]