RF exposure during mobile phone use is highly attenuated within a few centimeters in the brain, so exposure is largely to the side of the brain, and to the anatomic area, closest to the antenna. It has been reported that on the side of the brain where the phone is used, 50–60% of the total RF energy is absorbed in the temporal lobe, and the average specific absorption rate is highest in the temporal lobe and the cerebellum (Cardis et al. 2008
). Thus, examination of location of the tumor in relation to location of exposure is of interest.
Laterality. If there were a causal association between mobile phone use and brain tumor risk, one would expect an increased risk on the same side of the head as the phone is held and a null finding on the opposite side. On the other hand, if some brain tumor patients believed that mobile phone use had caused their tumor, and consequently overreported use on the affected side, this would result in an apparent risk increase on the same side of the head accompanied by a decreased risk on the opposite side. (The same bias is not possible for controls, who do not have a tumor side.) Furthermore, if there were a causal relationship, one would expect an effect of laterality to occur after a sufficient induction period, not for solely recent use (unless there was a promotional effect of mobile phones that was very rapid and substantial, which presumably would be easily and rapidly detectable from population incidence trends).
ORs for glioma and meningioma in the Interphone study tended to be greater in subjects who reported usual phone use on the same side of the head as their tumor than on the opposite side for most categories of duration of use, cumulative call time, and cumulative number of calls. Most ipsilateral ORs were not above unity, however, and there was no dose–response trend, although the greatest ORs tended to be for the top decile of ipsilateral exposure.
There are currently no validation studies of retrospective self-reported side of use, and there is no evidence of consistency over time in the preferred side of use. Overall, the greater risk for reported ipsilateral than contralateral use would be compatible with causation or bias as an explanation, but the finding that contralateral risks and many of the ipsilateral risks were generally below unity, with no consistent pattern of greater ipsilateral/contralateral ratios with greater exposure (except for cumulative number of calls and risk of glioma), would favor bias as the explanation.
Lobe. The risk of glioma in the temporal lobe for regular use and for most categories of exposure was reduced and did not differ from that in other lobes. ORs for long-term use and highest cumulative call time, however, were somewhat greater in the temporal lobe than in other lobes. This is the pattern one would expect if there were a causal effect, although there was no suggestion of a dose–response effect for temporal tumors, which would also be expected if there were causality. No coherent pattern was observed for meningioma, for which the OR for temporal lobe tumors for regular use was somewhat lower than for other lobes, and there was no evidence of greater risk in the temporal than other lobes in other categories of use.
Exact anatomic location of the tumor.
Interphone collected neuroradiologic information on the exact locations of brain tumors in the study. Although this has not been published for the study overall, it has been published for glioma for many of the study centers and for meningioma for one center. These analyses gave no indication of an association of tumor risk to proximity of the tumor to the exposure source (Larjavaara et al. 2011
; Takebayashi et al. 2008
In summary, among the three types of data on anatomic location, the results for laterality of phone use are the least interpretable. They are compatible with bias, or at least partly with causation, but do not give firm evidence for either. The evidence on lobe of glioma, but not of meningioma, is inconsistently in the direction that would be expected with causality, but not decisively so. The evidence on exact location of the tumor, which one would expect to give the most rigorous analysis because it has greater precision without bias, does not support a causal association.