Mobiles phones and health

Abstract

Mobile phones came on the market in the 1980s and people in the Nordic countries were among the first to use the technology. We are now among the most frequent mobile phone users in the world. Mobile phone technology has changed from analogue to digital signals, and signal frequencies have changed from 450 to 2200 MHz. Exposure to this non-ionizing radiation depends on the distance from base stations, the type of phone, where the antenna is placed, and of course on how frequently the phone is used by you and by those around you. A possible brain cancer risk was among the first concerns for no particular good reason. An anecdotal report on the Larry King Show in 1993 linking phone use to brain cancer had substantial significance for later funding, in spite of the fact that this type of radiation is not expected to cause DNA damage. The exposure may at worst promote a cancer risk, although we do not know how. It may, in fact, be more plausible that the exposure could impact brain functions and lead to a change in cognitive or behavioural problems. The largest research funds have by far been given to cancer studies, and a number of large and small studies have been carried out showing anything from no association to a possible protective effect, or that mobile phones are a possible cause of some types of brain cancer [1–9]. The only conclusion we can draw at present is that the use of mobile phones is probably not a strong cause of cancer. This conclusion is based on evidence from specific epidemiologic studies as well as on the general epidemiologic monitoring of brain cancer incidence over time. Some brain cancers have displayed an increasing incidence over time, but this can partly, or perhaps fully, be explained by improvements in diagnostic tools. But why is it so painful and difficult to reach consensus? One reason is that these studies are difficult to do, and perhaps because funding agencies jumped into large-scale case-control studies at an early stage without doing their homework properly. The casecontrol study is a valid design, as good as any observational design, if a number of conditions are fulfilled. In this case, they are not. First, you need to be able to obtain valid exposure data retrospectively, and in most of these studies exposure assessment is based on recall over years or even decades. This recall for cases is given by a brain with a cancer that may have been present for many years before it was diagnosed. How is it possible under these conditions to get valid exposure data or at least to maintain symmetry in the quality of exposure assessment among cases and controls? Most likely, it is not. The method studies done by the Interphone group have clearly shown that it is difficult to get valid exposure data and to avoid selection bias.