Mobile phones, magnetic fields, and cancer
Clinical bottom line (2003)
There is no convincing evidence linking mobile phone use to cancers of the head. What evidence we have points to there being no link, though what cannot be excluded is long-term heavy use of mobile phones, with a long latency for cancer development.
There is evidence of a small increased risk of leukaemia with occupational exposure to electromagnetic fields.
Does the use of mobile phones cause cancer? Important and interesting question, even if, like Bandolier , you think the off switch is the most important button on the thing. A number of epidemiological studies are now emerging, so we though it worth having a look at the evidence.
Cancer and cellular telephones - a Danish nationwide survey
C Johansen et al. Cellular telephones and cancer - a nationwide cohort study in Denmark. Journal of the National cancer Institute 2001 93: 203-207.
In 1968 Denmark adopted a unique 10-digit identifier for each citizen, and a central population register to verify personal data and information. From January 1982 when the first mobile phone system began to operate, to 1995, information from mobile phone companies was used to identify all users of mobile phones in Denmark. Excluding corporate users and some others this gave 420,095 persons (357,550 men) in whom information about mobile telephone use could be linked to a Danish cancer registry (operating since 1942), with information on cancer diagnosis.
This linked cancer information was used to calculate, according to sex and age, incidence rates for each type of cancer for Denmark, and for mobile telephone users.
In men, mobile telephone use was associated with a decreased risk of any cancer (standardised incidence ration 0.86, 95% confidence interval -.83 to 0.90). For women the standardised incidence ratio was 1.0. There was no significantly raised risk of any particular cancer associated with mobile phone use in men or women, in particular brain and salivary gland cancers, and leukaemia. Significantly reduced risks were seen for several cancer types, particularly lung cancer (standardised incidence ratio 0.65; 0.58 to 0.73) for men.
For brain and nervous system tumours and leukaemia there was no relationship between time or age at first subscription, or duration. There was no increased risk for any particular type of brain or nervous system disorder.
Brain tumours and salivary gland cancers
A Auvinen et al. Brain tumours and salivary gland cancers among cellular telephone users. Epidemiology 2002 13: 356-359.
This Finnish study identified 398 brain tumours and 34 salivary gland cancers in patients aged 20-69 years in Finland in 1996. Most of the cancers were microscopically confirmed. For each case five age and sex matched controls were selected.
All subscribers to the cellular phone networks were identified and information on cellular phone subscriptions for cases and controls obtained, including type of service (analogue/digital), and start and end dates. Other potential confounding information was gathered, including occupation.
Cellular phone use was not associated with brain or salivary gland tumours, with odds ratios of about 1.3 and with confidence intervals embracing 1.
JE Muscat et al. hand held cellular telephones and risk of acoustic neuroma. Neurology 2002 58: 1304-1306.
In this case control study were 90 patients with histologically confirmed acoustic neuroma diagnosed between 1997 and 1999 in New York. Controls were 86 in-patients with a variety of nonmalignant conditions, matched by age, sex, race and hospital of admission. A structured questionnaire identified use of hand held cellular telephone.
The risk of acoustic neuroma was unrelated to the frequency and duration of mobile phone use, with an odds ratio of 1.7 (0.5 to 5.1). There was no relation with cumulative use, or with increasing levels of exposure.
Malignant melanoma of the eye
C Johansen et al. Mobile phones and malignant melanoma of the eye. British Journal of Cancer 2002 86: 348-349.
This study included all cases of ocular malignant melanomas in Denmark from 1943 to 1986. Age-specific incidence rates were calculated for five year intervals. Information on the number of subscribers to mobile phone networks for the period was also collected.
The mean number of incident cases of ocular malignant melanoma increased, but the age standardised rate remained constant. The number of mobile phone subscribers rose from 13,000 in 1978 to 1,300,000 in 1996 (Figure 1).
Figure 1: Age-standardised rates of ocular malignant melanoma and mobile phone use in Denmark
There is apparently a German case-control study that comes to a different conclusion, and Bandolier will try and obtain it and add it here for comparison.
Brain cancer and magnetic fields
PJ Villeneuve et al. Brain cancer and occupational exposure to magnetic fields among men: results from a Canadian population-based case-control study. International Journal of Epidemiology 2002 31: 210-217.
This study was conducted among men in eight Canadian provinces, for 543 cases of brain cancer confirmed histologically (no benign tumours included). Astrocytoma and glioblastomas accounted for over 400 of these. Population based controls (543) were selected to be of similar age.
Questionnaires were used to obtain information on subject and control residential and occupational histories and other risk factors for cancer, including exposure to occupational carcinogens, and on diet. Subjects were asked to report on all jobs held for more than one year. Occupational magnetic field exposure but through manual inspection for each subject of several key variables through expert review. Occupations were assigned an exposure value based on a time-weighted average magnetic flux density for full time workers.
Cases and controls were well matched. Most jobs (85%) for cases and controls were deemed to have low exposure to magnetic flux.
The risk of all brain cancers was not increased by exposure to higher magnetic fields. For glioblastoma multiforme the odds ratio for association with the highest magnetic flux exposure was higher (odds ratio 5.4 (1.2 to 25), though here there were just 18 cases and 6 controls.
These are interesting studies, and the Canadian study in particular is very detailed. It is the only one showing any relationship between magnetic or radio flux and cancers, and even then just for one cancer at high magnetic flux where there were very few cases and controls (24 in total). It would be premature to make too much of this, given that magnetic flux exposure could not be accurately assessed, and there were occupational and other exposures to carcinogens.
Mobile phones and brain tumours
L Hardell et al. Use of cellular telephones and the risk for brain tumours: a case-control study. International Journal of Oncology 1999 15: 113-116.
This was a case-control study of all cases with a histopathological diagnosis of brain tumour in two regions of Sweden during the mid-1990s. There were 270 cases , with 233 actually participating. Each was age and sex matched to two controls from a population register. The use of mobile (cellular) telephones over preceding years, including type of system and pattern of use was established for each case and each control.
The proportion of mobile phone uses was the same (38%) in each group, and no increased risk of brain tumour was found. Latency, type of tumour, position of tumour and amount of use of mobile phones were all examined for relation to tumour development. None was found.
H Frumkin et al. Cellular phones and risk of brain tumour. CA Cancer Journal for Clinicians. 2001 51: 137-141.
This paper was a description of mobile phone systems and a review of three other case-control studies. They all came up with exactly the same results. None showed any relationship between phone use and brain tumour, nor did any show any dose-response between phone use and risk for developing a tumour.
PD Inskip et al. Cellular-telephone use and brain tumors. Mew England Journal of Medicine 2001 344: 79-86.
This was a prospective case-control study of 792 patients with brain tumours and 799 matched controls in the USA between 1994 and 1998. The patients with brain tumours had glioma (489), meningioma (197) or acoustic neuromas (96).
There was no relationship between any tumour type or all tumours and the use of cellular telephones. There was no relationship between the side on which the tumour occurred and the side on which the cellular telephone was most often used.
L Hardell et al. Ionizing radiation, cellular telephones and the risk for brain tumours. European Journal of Cancer Prevention 2001 10: 523-529.
This retrospective case-control study in Sweden included 209 cases and 425 controls. It examined not only cellular phone use, but also occupation, and exposure to ionising radiation through occupation, or through diagnostic or therapeutic exposure.
Two occupations had a significantly raised incidences. Workers in the chemical industry had an odds ratio of 4.1 (1.3 to 13) and laboratory workers an odds ratio of 3.2 (1.2 to 8.9). These represented 13 and 16 cases plus controls respectively.
X-ray investigations of head and neck gave an odds ratio of 1.6 (1.04 to 2.6), but with no increase for X-rays to other parts of the body. Increased risk for cellular telephone use was found only for tumours in the temporal, tempoparietal or occipital lobe on the same side as the use of a mobile phone, and only with conditional logistic analysis with 13 exposed cases.
JE Muscat et al. Handheld cellular telephone use and risk of brain cancer. JAMA 200 284: 3001-3007.
Between 1994 and 1998, 469 men with primary brain cancer were diagnosed in five US medical centres. They were studied with 422 matched controls who were inpatients at the same hospital as the cases and who had benign conditions. The risk of brain cancer was compared according to the use of handheld cellular telephones in hours per month and years of use.
Median monthly hours of use were about 2.5. Compared with controls who never used cellular telephones there was no increased risk of brain cancer, with an odds ratio of 0.85 (0.6 to 1.2). There was no increased risk for heavy users (more than 10 hours a month) compared with light users (less than 0.7 hours a month). There was no difference for side of head (compared with usual side phone was used). The odds ratio was less than one for all but uncommon neuroepithelialiomatous cancers, where it was above one, but not significantly so.
Occupational electric and magnetic field exposure and leukaemia
LI Kheifets et al. Occupational electric and magnetic field exposure and leukaemia: A meta-analysis. Journal of Occupational Medicine 1997 39: 1074-1091.
This meta-analysis was conduced on 39 of 70 studies identified though a literature search (not described). The 39 studies all reported results about occupational electromagnetic field exposure and leukaemia.
The overall result was a small increase in overall risk, with a relative risk of 1.18 (1.12 to 1.24) among individuals employed in electrical occupations. There may be a higher risk for more specific leukaemias subtypes.