University of Bielefeld; School of Public Health;
University of Mainz; Institute of Medical Statistics and Documentation; Mainz, Germany
Department of Environmental Epidemiology; German Cancer Research Centre, (DFKZ); Heidelberg, Germany
Department of Epidemiology and Medical Statistics; P.0. Box 10 01 31; 33501 Bielefeld, Germany; (address correspondence to: Maria Blettner)
Epidemiology and society addresses broad-ranging topics of interest to researchers, public-health workers, and other readers. We invite short submissions offering reports on global health issues, insights about epidemiology, policy commentaries, or other pieces of general interest.
On November 12–13, 1999, an international workshop was held in Heidelberg on the relation between mobile telephones and tumors of the brain, head, and neck. Organized by the Department of Environmental Epidemiology of the German Cancer Research Center, the Institute of Medical Statistics and Documentation of the University of Mainz, and the Department of Epidemiology and Medical Statistics of the University of Bielefeld, the workshop aimed to summarize present knowledge of the health effects of mobile telephones, to report on current research activities, and to present results from a pilot-study from Germany to guide further research activities. Scientists from various European countries, the USA, and New Zealand; representatives of district and country authorities; and employees of different telephone companies participated in the workshop.
Kenneth Rothman (USA), the first speaker, summarized results from epidemiologic studies that examined the risk of cancer for amateur radio operators, craftsmen of communication machines, members of the army or police force, and CB radio operators, all of whom were exposed to high-frequency electromagnetic fields (EMF). The results of these studies were inconsistent with respect to cancer risk. One important shortcoming of most of the studies was missing information about length and magnitude of exposure. Cellular telephone users faced health risks apart from tumors: 4% of patients with pacemakers had irregular heart beats, 1 and a recent study discovered a fourfold increase in risk of an automobile collision for mobile telephone users. 2 The effect on collision was observed even for persons using hand-free devices.
Lennart Hardell (Sweden) and Joshua Muscat (USA) presented results from two new case-control studies. The Swedish study is a population-based case-control study that included 209 patients with a brain tumor and 425 controls. Overall, the investigators found no increased occurrence of brain tumor. 3 Some scientists criticized the study for the many subgroup analyses that highlighted only striking results. The American study was a hospital-based case-control study. The study comprised 469 cases from different neurosurgical clinics and 422 controls matched according to age, sex, ethnicity, and region. Exposure was determined on the basis of a “snapshot” assessment of telephone bills. Overall, no increased risk was observed. Neither study is sufficient to assess a potential association between the use of cellular telephones and brain tumors.
Two cohort studies investigating the association between use of mobile telephones and health effects were presented. In 1996 a cohort study started in the United States aiming to compare the mortality of cellular phone users with users of portable phones or hand-free car phones. The exposure was determined via telephone bills. Interference from a lawsuit that was later dismissed by summary judgement prevented the continuation of the study, and as a result data were available for only 1 year of follow-up. There was no indication of higher cancer mortality among cellular telephone users. In addition, the study confirmed that in the United States, telephone bills can be used as a good proxy for the exposure. 4–6 Members of the workshop expressed their regret that the study could not have been continued.
Christianson Johansen presented a cohort study from Denmark that included 500,000 cellular telephone users. The exposure was recorded via records of both incoming and outgoing calls. The implementation of the study was more complicated than expected, as linkage of telephone owners with the death registry was not always possible. Currently the study still is underway, but its continued funding is uncertain.
The main difficulty in investigating the association between EMF and brain tumor is the exposure assessment. Maria Feychting (Sweden) presented results of methodic considerations affecting misclassification when measuring exposure. With low prevalence of the risk factor, a high specificity of exposure measurement is essential to avoid underestimation of the risk. If only the frequency of outgoing calls is reported, the exposure assessment will have a low sensitivity. Specificity, however, will be high as users and non-users can easily be separated, and effects will be estimated with little bias but low power.
Simon Mann (UK) demonstrated problems measuring EMF exposure from base stations. In general, measurements assigned to randomly selected places showed a low energy level, between 0.01–1 mW/m 2, with the maximum being 10 mW/m 2. The measurement closest to the base station did not always show the highest energy level. This phenomenon can be explained by the fact that base stations are usually placed on roofs of high buildings or on towers where they radiate at an angle of 6° with the horizon. Therefore, little exposure occurs directly under or near the base station. Measurements show clearly that the distance to a base station is not a good proxy for the exposure. The geographic surroundings, especially the shielding of radiation from buildings or walls, play a major role. Mann also found a very low association between indoor and outdoor measurements. His conclusion was that at the moment it is impossible to investigate the exposure of base stations in epidemiologic studies.
International Case-Control Study
Elisabeth Cardis presented WHO activities relating to a multi-centered international case-control study on mobile-phone-use and tumors in the head and neck region. The study includes 13 countries yielding nearly 6,000 patients with brain tumors and acoustic-neurinoma, and 6,000 controls. In a feasibility study the prevalence for mobile phone use in different study regions was estimated to 3.1% (range: 0.2–10%) in 1992 and to 12.5% (range: 2–29%) in 1995. The study will include 30- to 59-year-old men and women, predominantly from urban regions. Information about the cellular telephone use will be obtained using a questionnaire, because in most countries cellular telephone records of use are kept for only a short period of time. Furthermore, the bills usually contain information about outgoing calls only and the reconstruction of information is time-consuming and costly. Hence, such information would only be used in a validation study.
Confounders such as extreme low frequency EMF (eg, occupational exposure in electronic companies) were not seen to be a major problem for the study, as they are not considered to be important risk factors for brain tumors nor are such exposures strongly associated with the use of mobile phones. The study will be population-based, and data collection will be done with a computer-assisted personal interview (CAPI).
The study is supported to a large extent by the 5th Framework of the European Union (EU), with some industrial support. In Germany three centers (Bielefeld, Heidelberg, and Mainz) will participate in the international study. A pilot study has been conducted during the last year and has shown that the study is feasible. The complete study will have good power to detect a 50% increased risk for brain tumors.
Although the participants generally regarded the study as necessary and important, a basic critique was that mobile telephone use is mostly recent and duration of exposure is short. In the case of a negative result from the case-control study, it could be argued that the induction time is too short to have observed an effect. On the other hand, postponing the start of the study was not recommended, as cellular telephone use is increasing rapidly. Nevertheless, a negative result would likely only trigger further study.
Several investigators suggested a cohort study to complement the case-control study. The feasibility of such a study in Germany has already been investigated. A population-based sample would include a high proportion of exposed persons, as now some 20% of the population in Germany uses a mobile phone. Exposure assessment was tested with a short questionnaire in the pilot study, revealing a response rate of 40% from a population-based sample. Of the respondents 90% were willing to participate in further investigation. The follow-up can be done easily with the German registration system.
A cohort of about 100,000 persons, including 50% phone users and follow-up for 5 years, would have enough statistical power (80%) to find a 50% increased overall mortality rate or a mortality rate from accidents or other diseases with a high mortality rate if mobile phone users are compared with non-users. The study, however, would have only limited precision to estimate an increased risk for rare diseases, eg, certain cancers. External (with general population) and internal comparisons of the mortality data could be conducted. Owing to the rising number of cellular telephone users, however, the general population cannot be used as a non-exposed group.
The major advantage of the cohort study is that various endpoints can be investigated. A further advantage of a cohort design would be the possibility to contact participants after some years to measure the changes in the use of cellular telephones and in the technology used.
Health Risk from Base Station Exposure
German politicians and those living near base stations have demanded the investigation of the health risks associated with their EMF radiation from base stations. Measurements at randomly chosen places in Mainz showed that the measured field power were smaller than the current regulative limits for the whole RF spectrum as well as for the mobile transmission frequencies. The measurements also showed that even when the fields from base stations yielded strong signals, they constituted only a part of the total EMF radiation. The German investigation confirmed results from the UK (reported by Mann, see above) that a suitable exposure measurement for the exposure from base stations for epidemiologic studies does not yet exist. Distance from a base station does not allow estimation of individual RF-EMF exposure. Measurements are unstable and not representative of the past exposure, in addition to being costly and time-consuming. Suitable dosimeters for measuring individual exposures are unavailable at present and are not expected in the near future, as the threshold for measuring exposure from base stations is too low. Retrospective exposure assessment for case-control studies is nearly impossible as there are no historical data. As mentioned above, high specificity of exposure measurements is essential for estimating the health risks of low exposures, but with available methods only low specificity would be expected. Participants agreed strongly that implementing epidemiologic studies to examine the effect of base station exposure would be fruitless, despite political and other pressures to do so.
The participants recommended that the planned international case-control study should proceed but with tempered enthusiasm because of the restriction to a short induction time. Investigations of base station effects were considered infeasible, as there is no possibility to estimate individual exposure accurately enough. Most participants favored cohort studies, although some criticized the large costs that they would require.
1. Hayes DL, Wang PJ, Reynolds DW, Estes M, Griffith JL, Steffens RA, Carlo GL, Findlay GK, Johnson CM. Interference with cardiac pacemakers by cellular telephones. N Engl J Med 1997; 336: 1473–1479.
2. Redelmeier DA, Tibshirani RJ. Association between cellular telephone calls and motor vehicle collisions. N Engl J Med 1997; 336: 453–458.
3. Hardell L, Nasman A, Pahlson A, Hallquist A, Hansson MK. Use of cellular telephones and the risk for brain tumours: A case-control study. Int J Oncol 1999; 15: 113–116.
4. Rothman KJ, Loughlin JE, Funch DP, Dreyer NA. Overall mortality of cellular telephone customers. Epidemiology 1996; 7: 303–305.
5. Dreyer NA, Laughlin JE, Rothman KJ. Cause specific mortality in Cellular Telephone Users. JAMA 1999; 282: 1814–1816.
6. Funch DP, Rothmann KJ, Loughlin JE, Dreyer NA. Utility of telephone company records for epidemiologic studies of cellular telephones. Epidemiology 1996; 7: 299–302.