ID Number		1814
Study Type		In Vitro
Model		Meta analysis of RF exposure and effects on DNA breaks, micronuclei, and chromosome aberrations in in vitro studies.
Details		Meta-analysis of 63 different reports on DNA damage and chromosome aberrations. The authors report some statistically significant increases in genotoxicity endpoints under certain exposure conditions (specifically with occupational / mobile phone use studies and DNA damage, MN formation). They report a collective effect for RF when used with certain chemical initiators. Finally, they report considerable evidence for publication bias, and that the mean indicies for chromosomal aberrations and micronuclei fell within the spontaneous levels reported in the historical literature.
Findings		Effects
Status		Completed With Publication
Principal Investigator		Univeristy of Texas, USA - vijay@uthscsa.edu
Funding Agency		FGF, Germany
Country		UNITED STATES
References		Vijayalaxmi, et al. Radiation Research, (2008) 169:561-574
Comments		The inclusion of occupational / mobile phone use studies with studies of RF exposure to cultured cells in the lab should not have been done. These are entirely different exposure situations and cell environments, and this was one major reason for the statement "at certain RF radiation exposure conditions, there were statistically significant increases in genotoxicity for some endpoints". These mobile phone / occupational use studies have no exposure accuracy and many of them provide questionable results (e.g., Garaj-Vrhovac et al, Maes et al, Lalic et al, Gadhia et al, Gandhi et al). Not surprisingly, when occupational / mobile phone use studies were analyzed alone, this resulted in a significant effect in coment tail length (ss + ds breaks) and MN. The RF + chemical data in Table 3 was difficult to understand. Many of the subset analysis were from a small pool of data (in the case of ss breaks, only 2 publications and these were both were from the same group at Zhejiang University in China - both positive - what use is such analysis ?). Significant findings were reported for MN, chromosome aberrations, and SCE endpoints but it is impossible to tell from this analysis (without a detailed investigation of the literature) exactly what studies contributed to each of these endpoints, what the SAR and exposure conditions were, etc. Although the authors did report data for "all frequencies" and all signal types (CW, PW, etc), the additional stratification of the data by frequency (above / below 2 GHz) as well as signal type could have allowed spurious findings (e.g., Table 1, comet tail, >2GHz). If thermal effects are dominant (per the standards that the authors refer to), it should just be a matter of total energy absorption (i.e., heat). Even if there is a postulated "non-thermal" mechanism, there is still no valid reason to assume it would be frequency dependent. The meta-analysis did not analyze raw data in a consistent way (admittedly because as the raw data is often not available). In relying upon the different approaches used by each study author to calculate (and present) their data, however, they seemed to accept a large degree of error. Some of the studies may have been done poorly and there was no way to make corrections for this (e.g., how did the investigator score MN, how was comet tail measured, what was the ratio of ds : ss breaks in the comet tail, was data from a single or multiple experiments). From what I can see, there was a mixing of studies performed both in vitro and in vivo. This may have diluted some "effects" caused by heating in vitro or animal stress in vivo, so it may not have as significant of a consequence on the final conclusion. However, it is still poor overall design without any attempt for correction. The authors state that SAR was not reported in 21% of the papers, but they did not say (from my read) how they dealt with those studies - apparently they were (somehow) incorporated into the analysis (?). One of the major problems with almost all studies of DNA and chromosome damage in this field is that they enlist a single assay to detect a very subtle effect (at the level of noise), but do not use corroborating assays to confirm / validate their results (either negative or positive). Many (most) of the studies do not even use approrpiate positive and heat controls. A more appropriate design might be to corroborate comet assay data with changes in the cell cycle or expression / activaiton of checkpoint factors (e.g. H2AX, chk-1, chk-2, p53, c-fos), etc. Nobody in this field does this ! This could have been mentioned in the meta-analysis conclusion.