Genetic damage in human cells exposed to non-ionizing radiofrequency fields: a meta-analysis of the data from 88 publications (1990-2011).

Based on the 'limited' evidence suggesting an association between exposure to radiofrequency fields (RF) emitted from mobile phones and two types of brain cancer, glioma and acoustic neuroma, the International Agency for Research on Cancer has classified RF as 'possibly carcinogenic to humans' in group 2B. In view of this classification and the positive correlation between increased genetic damage and carcinogenesis, a meta-analysis was conducted to determine whether a significant increase in genetic damage in human cells exposed to RF provides a potential mechanism for its carcinogenic potential. The extent of genetic damage in human cells, assessed from various end-points, viz., single-/double-strand breaks in the DNA, incidence of chromosomal aberrations, micronuclei and sister chromatid exchanges, reported in a total of 88 peer-reviewed scientific publications during 1990-2011 was considered in the meta-analysis. Among the several variables in the experimental protocols used, the influence of five specific variables related to RF exposure characteristics was investigated: (i) frequency, (ii) specific absorption rate, (iii) exposure as continuous wave, pulsed wave and occupationally exposed/mobile phone users, (iv) duration of exposure, and (v) different cell types. The data indicated the following. (1) The magnitude of difference between RF-exposed and sham-/un-exposed controls was small with some exceptions. (2) In certain RF exposure conditions there was a statistically significant increase in genotoxicity assessed from some end-points: the effect was observed in studies with small sample size and was largely influenced by publication bias. Studies conducted within the generally recommended RF exposure guidelines showed a smaller effect. (3) The multiple regression analyses and heterogeneity goodness of fit data indicated that factors other than the above five variables as well as the quality of publications have contributed to the overall results. (4) More importantly, the mean indices for chromosomal aberrations, micronuclei and sister chromatid exchange end-points in RF-exposed and sham-/un-exposed controls were within the spontaneous levels reported in a large data-base. Thus, the classification of RF as possibly carcinogenic to humans in group 2B was not supported by genotoxicity-based mechanistic evidence.