ABSTRACT
A range of smaller studies on use of mobile phones and risk of central nervous system (CNS) tumors have been published. Apart from one American cohort study that ended after only 1 year for legal reasons (Rothman et al. 1996; Dreyer et al. 1999), these studies have been case-control studies limited by small numbers, short follow-up, or methodological limitations (Baan et al. 2011). In 2011, the International Agency for Research on Cancer (IARC) classified radiofrequency electromagnetic fields as “possibly carcinogenic to humans” (Baan et al. 2011). This classification was primarily based on results from two questionnaire-based case-control studies, namely, the studies done in Sweden by Hardell and colleagues (2011) and the international INTERPHONE study (Cardis et al. 2007). Both of these studies on mobile phones and brain tumors have been ongoing for many years. In general, the interpretation of the case-control studies is severely hampered by their susceptibility to biases due to differential participation, recall, and reporting, depending on case status. These problems have been further aggravated by the need to use proxy reports
CONTENTS
Introduction ................................................................................................................................. 203 Methods ........................................................................................................................................ 204
Danish Subscriber Cohort ..................................................................................................... 204 Million Women Study ............................................................................................................ 205
Results ........................................................................................................................................... 206 Meningioma ............................................................................................................................ 206 Glioma ...................................................................................................................................... 206 Pituitary Gland ....................................................................................................................... 206 Vestibular Schwannoma ........................................................................................................ 206
Strengths and Limitations of the Cohorts ................................................................................ 208 Strengths .................................................................................................................................. 208 Limitations .............................................................................................................................. 208
Lack of Exposure Details .................................................................................................. 209 Other Sources of Exposure ............................................................................................... 209 Corporate Users/Lenders Borrowers ............................................................................. 209 Late Users ........................................................................................................................... 210
Summary ...................................................................................................................................... 211 References ..................................................................................................................................... 212
for deceased cases. Overall, Interphone found no clear evidence of increased risk of CNS tumors in mobile phone users; there were however increased odds ratios (ORs) for both glioma, 1.40 (95% confidence intervals [CI], 1.03-1.89) (INTERPHONE Study Group 2010), and vestibular schwannoma (also called acoustic neuroma), 1.32 (0.88-1.97) (INTERPHONE Study Group 2011) among subjects within the highest decile of cumulative self-reported use (>1640 min). However, several factors speak against a causal interpretation (INTERPHONE Study Group 2010; Swerdlow et al. 2011): implausible levels of daily use were reported in this stratum, there was no suggestion of dose response over the preceding categories of exposure, and the ninth decile was even among the lowest ORs observed. The Hardell studies (Hardell et al. 2011) have reported elevated risks, already, after a few years of phone use; these results are however incompatible with incidence trends (Deltour et al. 2009, 2012; Inskip et al. 2010; Little et al. 2012). Incidence studies, however, can only discern effects that are strong enough to be seen on population level against a backdrop of other factors influencing incidence. The ideal study would be a prospective cohort study with elaborate and repeated or continuous exposure assessment that would avoid or minimize the issues with incidence and case-control studies (WHO 2006; SCENIHR 2009b). Data are being collected for such a study (Schüz et al. 2011a); the results are however still some time in the future. In the meantime, an existing entirely register-based cohort of subscription holders in Denmark until 1995 has been used for investigation of a range of outcomes (Johansen et al. 2001; Schüz et al. 2006b, 2009, 2011b; Frei et al. 2011; Poulsen et al. 2012, 2013). Much concern has, however, been expressed about both actual and conceived shortcomings of this study (e.g., Ahlbom et al. 2007; Khurana 2011; Philips and Lamburn 2011; Söderqvist et al. 2012). A recently published British study has used an ongoing cohort; the million women study, where participants have been asked basic questions on their use of mobile phones (Benson et al. 2013). An overview of these two studies, with a discussion of their strengths and weaknesses, is given here.
