Childhood cancer and proximity to mobile phone masts

British Medical Journal Editorial on Mobile Phone Masts and Cancer

Epidemiological studies show no increased risk

Radiofrequency fields are now ubiquitous, and several studies have assessed their potential health effects,1 with predominantly negative results. The two main areas of research are exposure associated with the use of mobile phones—for example, the recently published INTERPHONE study2—and risks associated with transmitters, including mobile phone masts.

The linked case-control study by Elliott and colleagues
(doi:10.1136/bmj.c3077), assesses whether proximity to masts during
pregnancy raises the risk of children developing leukaemia or a tumour in the brain or central nervous system.3 The study identified 1397 British children registered with leukaemia or a tumour in the brain or central nervous system between 1999 and 2001, and it compared each of these children with four controls sampled from the national birth registers who were matched for sex and date of birth. The study found no association between the risk of cancer in early childhood and exposure to a mobile phone base station during pregnancy.

The levels of individual exposure from transmitters are much lower than
those from mobile phones,4 although the exposure to mobile phones clearly depends on the extent of usage and may involve different physiological effects. Nevertheless, anxiety about an environmental risk persists, and before the risk is considered to be unfounded it is important to understand why the study may have missed a true effect.

The first, and probably the most important, reason is the size of the
study. Elliott and colleagues’ study is the first to look at phone masts in
Britain as a whole and is the largest of its kind. We can estimate from
table 3, for example, that the study would have had a greater than 90% probability of detecting a doubled risk of brain cancer between the 85th and 15th centiles of modelled power density; for childhood leukaemia (which has a higher incidence) the figure is over 99%.

Secondly, the exposure variables considered may be inadequate surrogates for the true exposure we would ideally measure. Distance from the source of risk has often been used in environmental studies, but this is particularly unsatisfactory in this situation, where the association with exposure is not monotonic. The second measure the authors consider—the total power output of the nearest transmitter—is also a relatively blunt instrument, because it does not distinguish between the highest and lowest risk distances. A third measure models the power density of each transmitter at each given distance, and this comes closer to an ideal measure of the strength of the field in the environment. However, individual exposures could still vary substantially according to building design, lifestyle, and
migration, so the prospects for good estimates of individual exposures are poor.5 Any methodology that permits the measurement of individual exposures would be scientifically valuable.

The third possible reason is case-control bias, but the use of register
data largely eliminates this. Lastly, we have the universal epidemiological problem of confounding. The authors adjusted for certain demographic variables, specifically socioeconomic status and population mixing, both of which have been associated with childhood leukaemia.6 These were defined by imputation from the corresponding “ecological”—or geographical—variables;
although this weakens the observed association between disease and
exposure, it does not invalidate its use in an epidemiological analysis.

The study is the first to focus on the risk of brain tumours and leukaemia after exposure during pregnancy: radiofrequency fields studies have mostly considered broadcast transmitters and adult cancers.1 The highest incidence of childhood leukaemia occurs in the first 5 years of life, so it is less necessary to conduct long term studies. Malignant disease is rare in children, so it is unlikely that more than a few cases could ever be attributed to proximity to phone masts. The risks are dwarfed by the well known dangers of distraction while using mobile phones, especially when driving—even when using hands-free equipment. Studies of cancers in adults would be more useful, because health effects are likely to appear after a long latent period.

Meanwhile, clinicians should reassure patients not to worry about proximity to mobile phone masts. Moving away from a mast, with all its stresses and costs, cannot be justified on health grounds in the light of current evidence. The epidemiological evidence is also supported by experimental evidence, which has so far failed to show any biological effects—in vivo or in vitro—that might lead us to worry about the impact on health.