Lonn S, Ahlbom A, Hall P, Feychting M (2004)
This study was part of the INTERPHONE study, an international case-control study of brain tumours, acoustic neuroma, and parotid gland tumours in relation to mobile phone use. It was population-based, and included all persons age 20 to 69 years of 3 geographical areas covered by the Cancer Registries in Stockholm, Göteborg, and Lund.
Cases were those diagnosed with acoustic neuroma during the period September 1 1999 to August 31 2002 in the Lund and Göteborg areas and from 1 January 2000 to 31 August in the Stockholm area. They were identified continuously during the study period, and medical records were used to confirm the diagnosis. The date of diagnosis was the date of the medical examination leading to the diagnosis, which was by histological examination in 39% and by CT or MRI in the others. There were 160 eligible cases.
Controls were randomly selected and stratified on age, sex, and residential area. In total 838 controls were identified.
Data collection began in September 2000, so that cases in the first year (6 months in the Stockholm area) were identified retrospectively. Personal interview was done in most, telephone interview in some that could not participate in a personal interview, and completion of a mailed questionnaire in some others. Details of mobile phone use were obtained. "Regular" users were those who used a mobile phone at least once per week on average during 6 months or more. "Unexposed" subjects were those who did not use a mobile phone, or, if they did, used it irregularly. Calculation of cumulative time of use and of cumulative number of calls was made. To analyze the possible association between laterality of phone use and laterality of tumours, the left and right sides were considered separately. Cases were divided into a left-sided group and a right-sided group depending on the location of the tumour. Controls were randomly assigned to either the left or right side group. For both cases and controls, exposure was defined as ipsilateral use or use of the phone on both sides, whereas contralateral use was defined as unexposed. The results from the two sides were then pooled. Phones were held on the right side 52% of the time and 10% on both sides. Analyses were also made where contralateral use or use on both sides was defined as exposed and ipsilateral use as unexposed. (This was to test for recall bias, but the results are not given in the paper).
Participation rates were 93% for cases (n = 148) and 72% for controls (n = 604). For regular use, regardless of duration, the relative risk was estimated to be 1.0. For those with at least 10 years since first regular use, the odds ratio (OR) was 1.9 (95% CI 0.9-4.1). There were 14 cases in this group. For those with 5-9 years since first regular use, the OR was 1.1.
For ipsilateral use, those with at least 10 years since first regular use had an OR of 3.9 (95% CI 1.6-9.5). For 5-9 years the OR was 1.1. For contralateral use the OR for the group with at least 10 years since first regular use was 0.9.
There were no increased ORs for cumulative use, cumulative number of calls, digital phones (although analyses were limited to less than 10 years), or cordless phones.
The authors point out that their method of laterality analysis was different from other studies, and overcame the difficulty of the cases being unevenly distributed between the two sides of the head - 59% were right-sided tumours.
The authors discuss the potential for bias in their study, including differential misclassification of exposure or of the disease, recall bias, and selection bias. They did not feel that any of these was likely in their study.