Hardell L, Carlberg M, Mild KH. Epidemiological evidence for an association between use of wireless phones and tumor diseases. Pathophys. Mar 5, 2009. Ahead of print.
This paper provides an update of a review of the scientific evidence for mobile phone use and risk of tumors. The manuscript employs several meta-analyses of case–control studies, mostly arising from two groups: those of the author (Hardell) in Sweden and those of the INTERPHONE publications from a variety of nations. Special attention is given to those studies that report exposure periods in excess of ten years, rationalizing that a ten year lag period is minimally acceptable for the potential carcinogenic effects of radiofrequency (RF) exposures. Another focus was ipsilateral and contralateral exposures (meaning same side and different side, respectively) in reference to whether phone use was on the same side of the tumour or not. The meta-analyses were employed with fixed-effects models. A small number of studies (including two cohort studies) were excluded because of methodological limitations.
These studies included typical malignant glioma and highly malignant glioblastoma multiforme. A summary odds ratio (OR) was calculated as 1.0 (with 95% confidence interval of 0.9-1.1, using n=11 studies). Confining results to those studies referring to exposure periods with 10-year of lag, the summary OR was 1.9 (1.4-2.4, n=6) for exposures ipsilateral to the side of the tumour.
These are benign tumours that develop in the vestibulocochlear nerve (CN VIII). The summary OR was 1.0 (0.8-1.1, n=9). For ipsilateral exposures, with a 10-year lag period, the OR was 1.6 (1.1-2.4, n=3).
These are often benign tumours, forming in the membranes surrounding the brain. The summary OR was 0.9 (0.8-0.9, n=9). For ipsilateral exposures, with a 10-year lag period, the OR was 1.3 (0.9-1.8, n=3).
Salivary gland tumours
Referring largely to parotid gland tumours, the meta-analysis summary OR for ipsilateral exposures, lagged 10 years, was 1.7 (1.0-2.9, n=4).
The authors reviewed a number of other tumour types. They concluded that there was “no consistent pattern of an association with use of wireless phones” for Non-Hodgkin lymphoma, testicular cancer, and salivary gland tumours. Review of results for malignant melanoma of the eye, as well as intratemporal facial nerve tumour, was tempered by the methodological limitations in the available studies.
The authors also examined risk in relation to age groupings (based on first use of mobile phones). Highest risk was estimated when first use occurred in those less than 20 years of age, where the OR was 2.7 (1.3-6.0) for malignant tumours and 2.5 (1.1-5.9) for benign tumours. An examination of urban and rural phone use (where rural use typically involves higher power output), with a 5-year lag period, resulted in OR estimates of 0.9 (0.6-1.4) and 3.2 (1.2-8.4), respectively.
The review concludes by discussing possible biases in the various case-control studies, though low numbers of long-term users was identified as an ubiquitous problem at this time. The authors summarize their results as supporting a consistent pattern of increased risk for acoustic neuroma and glioma, but only after 10 years mobile phone use. The authors also suggest that there is an association with use of cordless phones, which have not typically been included in analyses of mobile phone exposure.