Weight-of-evidence criteria for mobile phone use and cancer and gaps in knowledge
| Criteria | State of evidence | Gaps in knowledge |
|---|---|---|
| Amount and quality of epidemiological evidence | Nine studies published, five negative and four positive. Eight studies focusing on brain tumours, one about intraocular melanoma, and one additionally investigating salivary gland cancer and one haematopoietic cancers. Studies are lacking good quality data on exposure and have overall too short duration of mobile phone use. | What stage of carcinogenesis is “sensitive” to exposure? Which aspect of exposure is important? Is there an increased risk if exposure starts early in age (in childhood or adolescence), maybe due to exposure of bone marrow of the scull? |
| Strength of association in epidemiological studies | Studies approaching reasonable latencies show moderately increased tumour risk. Relative risks (excluding studies with too short latencies) overall 1.3 to 4.6, increased risk with increasing latency, highest overall risk for acoustic neuroma (3.5) and uveal melanoma (4.2). Significantly increased risk for ipsilateral use overall and for malignant brain tumours. | Is there a lower limit for duration of use below which there is no increased risk? For which types of cancer (and benign tumours) is there a particularly high risk? |
| Consistency of epidemiological findings | Three brain tumour studies show consistently an association to mobile phone use. Inconsistency may be due to insufficient latency and other methodological shortcomings. | Is there a method, feasible for epidemiological studies, that could effectively reduce possible response bias? |
| Dose-response relation | Little evidence for an association between intensity of use and risk to develop brain tumours (for intraocular melanoma only heavy occupational use was investigated), however, some evidence for an association between duration (years) of use and risk. | Is there a meaningful exposure meter measuring the “dose” of exposure to emissions from mobile phones (note that SAR as a rate cannot be used without further consideration, it is not known whether one month exposure to, say, one W/kg is equivalent to 10 months exposure to 0.1 W/kg, a.s.o.)? |
| Amount and quality of in vitro and in vivo laboratory evidence | Few long term animal studies, three on GSM type signals and three on mobile phone standards used in the USA or Japan. Several short and medium term animal studies. In vitro genotoxicity studies of mobile phone frequencies cover different endpoints. Quality of the studies is fair to good, however, their appropriateness cannot be assessed without knowledge of the mechanism of action. | What is a suitable exposure regimen in long term animal studies? Which endpoints are sensitive to exposure? What are the crucial processes in the exposed cells? How is the time course of reactions within cells? Are mechanisms invoked, depending on duration of exposure, that counteract possible damages? |
| Strength of evidence for genotoxicity or epigenetic effects | One long term animal study found significantly increased cancer rates. Several independent studies found increased micronuclei frequencies and some other indications of interference at the genome level. Furthermore, there is some evidence of activation of heat-shock proteins that may have an impact on cell cycle control and apoptosis. | Is there a relation between duration of exposure and effect? Are there differences between continuous and intermittent exposure? Is there a combination effect with coexposure to other agents? |
| Consistency of laboratory evidence | Long term animal studies differ in essential features of experimental methodology. Four long term animal studies found indications of an effect of exposure, however: three a reduced and one an increased risk. Such indications of hormetic effects at low doses are abundant in cancer research. Overall consistency cannot be evaluated at present due to unknown mechanism of action and lack of exact replication studies. | What is the mechanism of action over the whole range of exposures? Is there a change of essential features of the mechanism at the transition between relevant temperature increases and non-thermal levels? |
| Mechanistic explanation of a carcinogenic effect | Energy of the field too low to cause ionisation and signal to noise ratio too low to speculate about an effect relying on equilibrium thermodynamics. Other potential interaction mechanisms not generally accepted. | Are there resonance like phenomena? Are there coherent excitations at the cell membrane that could be responsible for a signal cascade? What is the role of gap junctions between cells? Are there non-linear elements in tissue propagation of the EMF that could act as demodulators? Is the carrier frequency of the EMF important or is only the low frequency modulation effective? |
| Overall | Some evidence of a carcinogenic effect from epidemiological studies weakly supported by long term animal experiments and in vitro investigations. | What is the mechanism of action? Which aspect of exposure is responsible? What is a suitable exposure meter? |