Objective: During the last decade, mobile phone use increased to almost 100% prevalence in many countries of the world. Evidence for potential health hazards accumulated in parallel by epidemiologic investigations has raised controversies about the appropriate interpretation and the degree of bias and confounding responsible for reduced or increased risk estimates.
Data sources: Overall, I identified 33 epidemiologic studies in the peer-reviewed literature, most of which (25) were about brain tumors. Two groups have collected data for >or=10 years of mobile phone use: Hardell and colleagues from Sweden and the Interphone group, an international consortium from 13 countries coordinated by the International Agency for Research on Cancer.
Data synthesis: Combined odds ratios (95% confidence intervals) from these studies for glioma, acoustic neuroma, and meningioma were 1.5 (1.2-1.8); 1.3 (0.95-1.9); and 1.1 (0.8-1.4), respectively.
Conclusions: Methodologic considerations revealed that three important conditions for epidemiologic studies to detect an increased risk are not met: a ) no evidence-based exposure metric is available; b) the observed duration of mobile phone use is generally still too low; c) no evidence-based selection of end points among the grossly different types of neoplasias is possible because of lack of etiologic hypotheses. Concerning risk estimates, selection bias, misclassification bias, and effects of the disease on mobile phone use could have reduced estimates, and recall bias may have led to spuriously increased risks. The overall evidence speaks in favor of an increased risk, but its magnitude cannot be assessed at present because of insufficient information on long-term use.
Keywords: acoustic neuroma; brain tumors; causation; epidemiology; glioma; meningioma; mobile phones; risk assessment.