Are these data real? Comments on “No effects of intermittent 50 Hz EMF on cytoplasmic free calcium and on the mitochondrial membrane potential in human diploid fibroblasts.” by Pilger et al. (Radiat Environ Biophys 43:203–7 (2004))

Abstract

Sir, Two papers (Diem et al. 2005; Schwarz et al. 2008) by the group of Hugo Rudiger, former head of the Division of Occupational Medicine, Medical University Vienna, on the effects of high-frequency mobile phone electromagnetic fields on DNA molecules have been questioned with regard to the origin of the data (Vogel 2008; Tuffs 2008). Although these papers have not yet been retracted, the editors of the International Archives of Occupational and Environmental Health (IAOEH) published an ‘‘Expression of Concern’’ (Drexler and Schaller 2009) in which they ‘‘apologize to the readers of IAOEH for publishing this paper’’. The editors of Mutation Research published an editorial (Baan 2009) in which they stated not to retract this paper, but to ‘‘...leave this issue open for scientific debate’’. COPE (Committee on Publication Ethics) considered a complaint against both journals and asked Elsevier to alter the title of the editorial to ‘‘Letter of concern’’ which has been published meanwhile (Baan 2010). This organization has more than 4,000 members (mainly Editors-in-Chief, but also Elsevier and Springer, which are the publishers of Mutation Research and IAOEH, respectively). The clear regulations, to which COPE members must adhere, also define the procedures once readers of published articles approach Editors with concerns about the validity of research data. The group in Vienna published not only data on the effects of electromagnetic fields on DNA but also a series of papers describing the same sort of effects caused by lowfrequency magnetic fields which are produced, for example, by household appliances or high-voltage power transmission lines (Ivancsits et al. 2002, 2003a, b, 2005; Pilger et al. 2004). Also here, the extent of the effects and the low variability of the data, respectively, were remarkable. In this respect, the data on DNA damage in the paper by Pilger et al. (2004) in this journal are record-breaking with regard to the magnitude of effect and the extremely low standard variations, respectively. For the sake of clarity, the figure (Fig. 1) was redrawn from the data given in the original paper (in the text and in Fig. 1). The standard deviations of the mean comet tail factors are barely visible. When calculating the coefficients of variation (standard deviation divided by the mean), these are below 1% for the sham-exposed group and below 0.5% for the exposed group. In other words, the individual data (n = 5) of each group are virtually identical. These extremely low standard deviations of ‘‘five independent series of exposure’’ are suspicious for a number of reasons. It was argued earlier that the very low standard deviations shown in two of the publications (Diem et al. 2005; Ivancsits et al. 2005) would be extremely sensitive to counting one single cell more or less in the respective categories (Vijayalaxmi et al. 2006). In addition, all other inevitable variations, such as cell cycle, exposure conditions, biochemical analysis, and time course of effect, should result in a much bigger variability in the data. To further clarify this important point on the incomprehensibly low standard deviations, some data are given in the following section. The authors (Pilger et al. 2004) refer in ‘‘Materials and Methods’’ to a paper from the same group (Ivancsits et al. 2002). In that paper, a detailed description of the comet assay and the calculations to get the tail factors are given. First, the cells’ comets were estimated in the following way: for each individual data point, 1,000 DNA spots were microscopically inspected, manually classified into five categories (A–E), and the A. Lerchl (&) School of Engineering and Science, Jacobs University Bremen, Campus Ring 6, 28759 Bremen, Germany e-mail: a.lerchl@jacobs-university.de