The basic population for the cohort study were all Danes (n=723,421) who had signed a mobile phone contract with a phone company for the first time between 1982 and 1995. After having excluded all subjects who did not fulfil the inclusion criteria of the study among which were 200,507 corporate contracts that could not be allocated to a single person, the cohort included 420,095 individuals. For these, data from the Central Population Register were compared to those of the Danish Cancer Register and the number of individuals who had developed cancer before the fixed date of 31.12.2002 was determined. Individuals who had developed cancer before signing the first contract were excluded from the analysis. Next, the cancer incidence in the cohort of the mobile phone users was compared to that in the general Danish population. Data on the incomes of the cohort were obtained from the Danish Statistics Office.

The so-called standardised incidence ratio (SIR) that indicates the factor by which the incidence rate observed in the cohort differs from the incidence rate expected in the general population for the same age and gender distribution, was calculated for the comparison of cancer incidences. In order to obtain a comparative population consisting of non-mobile phone users – as far as possible –, the number of cancer cases occurring in the cohort as well as the associated person-years were subtracted from the corresponding data of the general population in the analyses.

A SIR greater than 1 means an increased risk, a SIR smaller than 1 a reduced risk. The 95 % confidence interval indicates the range in which the actual risk lies with 95 % certainty. This means that an increase or a reduction in risk is statistically significant if 1 is not within the confidence interval.

The observation period was then extended by five years until the end of 2007 and specific analyses on brain tumours were performed. In order to obtain individual information on socio-economic factors, the data of the mobile phone user cohort were compared to that of the Danish CANULI cohort. This cohort included all Danish residents, aged 30 or older and who were born in Denmark after 1925. The association between socioeconomic status and cancer was investigated in the cohort. By restricting the study to the CANULI cohort, the data sets were reduced to 358,403 individuals.

• The cohort initially included 420,095 individuals, of which 357,553 were men and 62,542 were women. The average length of the mobile phone contracts was 8.5 years, the maximum was 21 years.
• During the follow-up period, 14,249 cancer cases occurred within the cohort of the mobile phone users, according to the control group 15,001 would have been expected. This results in a statistically significantly reduced risk of cancer in mobile phone users (SIR=0.95; 95 % confidence interval: 0.93 – 0.97).
• When viewing the individual tumour localisations, the reduced risk of cancer in men is mainly attributed to the reduced risks of tumours associated with tobacco smoking (for example lung cancer). Comparing the incomes of the mobile phone users to those of the general population revealed significantly higher incomes for mobile phone users, irrespective of age and gender. As there are fewer smokers in higher income groups than in lower ones, it was concluded that perhaps lower smoker prevalence and a healthier lifestyle may prevail within the cohort in comparison to the general population. This would explain the lower rate of tumours associated with tobacco smoking.
• No increased risks were observed for tumour localisations under discussion in the context of the use of mobile phones. These included brain tumours (SIR=0.97), acoustic neuromas (SIR=0.73), salivary gland cancer (SIR=0.77), eye tumours (SIR=0.96) or leukaemias (SIR=1.00). Also for long-term users (more than 10 years of mobile phone contract length, n=28 observed cases) no increased risk of brain tumours (SIR=0.66; 95% confidence interval: 0.44 – 0.95) was observed. Investigations on the risk of brain tumour depending on the length of time since the first contract did not show any statistically significant trend. The same applied to leukaemias.
• Analyses of the subgroups of brain tumours did not show any increased risk of cancer either. These contain gliomas (SIR=1.01), meningiomas (SIR=0.86) or cranial nerve sheath tumours including acoustic neuromas (SIR=0.73).
• When viewing the individual tumour localisations according to gender, a statistically significantly increased risk of kidney cancer and cervical cancer in women was found. However, the authors do not assume an association with the use of mobile phones, but that there are possible differences between mobile phone users and the general population regarding potential risk factors for these cancer types or that this is a random result due to multiple testing.

The results were published in the Journal of the National Cancer Institute in 2006 (Schüz et al. 2006).

Analyses on brain tumours were performed during the follow-up period extended until the end of 2007. The relevant findings were published in the British Medical Journal in 2011 (Frei et al. 2011).

Furthermore, the association between mobile phone use and the risk of skin cancer was analysed. The results were published in a further publication in the American Journal of Epidemiology in 2013 (Poulsen et al. 2013).

During the follow-up period 1990 to 2007, 10,729 new cases of brain tumours occurred. In comparison with the general population, mobile phone users in the cohort had no increase in the risk of brain tumours. This also applied to male and female long-term users (more than 13 years). No exposure-response relationships were found neither for analyses according to time in years after the first use nor for analyses according to the anatomical position of the tumour. Socioeconomic status was not rated to be a significant confounder. On the whole, the new results confirmed earlier findings from shorter observation periods.

The results from the study on the risk of skin cancer also indicate no association between mobile phone use and an increased risk of disease.

Strengths and weaknesses of the study

• The strengths of the study include the size of the cohort, the long observation period, the – compared to other studies – relatively long period that elapsed since the first use of the mobile phone and the collection of objective data on exposure (years of contract). Furthermore, cohort studies – in comparison with case-control studies – are generally not prone to bias issues such as recall bias, selection bias due to non-participation etc.
• One of the potential weaknesses of the study is that the selection of the participants may lead to a possible underestimation of risk. A large proportion of mobile phone users with corporate contracts were excluded. However, exactly those could be heavy users. Individuals who had signed mobile phone contracts after 1995 were not included in the cohort of mobile phone users. This means that the comparative population was not entirely made up of “non-mobile phone users”. Incorrect exposure classification cannot be excluded either. This is the case when individuals who have signed a contract (supposedly exposed individuals) do not use the mobile phones themselves but have passed them on to other individuals (supposedly non-exposed individuals) or vice versa. Data on the actual phone use of a subgroup of 822 individuals within the cohort were available: sensitivity analyses revealed that such exposure classification leads to a slight underestimation of risk. A further limitation of the study is the missing information on potential confounders.

Limited power of the study

The results of the Danish mobile phone cohort study show no increased risk of cancer in mobile phone users. This applies to brain tumours, salivary gland cancer, acoustic neuromas, eye tumours and leukaemias and also to short-term as well as long-term users (more than ten years).

A significantly reduced risk was found for brain tumours in long-term users. There is no biological explanation for this. As only a small number of cancer cases was observed, this result can be explained by random chance or by negative confounders that have not been taken into account.

On account of the methodical weaknesses mentioned above, on the whole, an underestimation of risk cannot be excluded. The validity of this study is therefore limited and does not provide any useful new information for radiation protection.

• Johansen, C., Boice, J. D., McLaughlin, J. K., Olsen, J. H. (2001). Cellular telephones and cancer—a nationwide cohort study in Denmark. Journal of the National Cancer Institute, 93(3), 203-207.
• Schüz, J., Jacobsen, R., Olsen, J. H., Boice, J. D., McLaughlin, J. K., Johansen, C. (2006). Cellular telephone use and cancer risk: update of a nationwide Danish cohort. Journal of the National Cancer Institute, 98(23), 1707-1713.
• Frei, P., Poulsen, A. H., Johansen, C., Olsen, J. H., Steding-Jessen, M., Schüz, J. (2011). Use of mobile phones and risk of brain tumours: update of Danish cohort study. British Medical Journal, 343.
• Poulsen, A. H., Friis, S., Johansen, C., Jensen, A., Frei, P., Kjær, S. K., Dalton, S.O., Schüz, J. (2013). Mobile phone use and the risk of skin cancer: a nationwide cohort study in Denmark. American Journal of epidemiology, 15; 178(2), 190-197