"Study Design

The study was performed as a double-blinded trial with only certain scientists responsible for the electromagnetic-field (EMF) technology

having knowledge of the dosing regimen. Therefore, during the conduct of the study the treatment groups were identified using capital letters and different colors (see Table 1). The code was made known to the study director and pathologist only after all the study data had been provided to the sponsors in PDF format. A total of 1200 rats entered this study. Groups of 65 male and 65 female rats were exposed to GSM and DCS wireless communication signals at three nominal SARs of 4 W/kg (high dose), 1.33 W/kg (mid dose) and 0.44 W/kg (low dose). In addition, there was a sham control group for each signal type. The high dose level was set on the basis of (a) ICNIRP guidelines (5), (b) thermal preliminary studies performed in our laboratory, and (c) the maximum capability of the exposure setup. Consequently the dose of 4 W/kg was considered as a suitable dose level that would avoid overloading of the thermoregulatory capacity of the rat. The mid and low dose levels were reduced by a factor of 3 each following conventional dose-level intervals for toxicology studies. Exposure was regulated for practical considerations to 2 h/day, 5 days/week for 52 consecutive weeks (15 male and 15 female rats/group), or for up to 104 consecutive weeks (50 male and 50 female rats/group). A ninth group of 65 males and 65 females was kept unexposed and unrestrained and served as the cage control for the same intervals. At the start of the exposure period, 15 male and 15 female rats were selected from the reserve pool of animals for disease surveillance. Of these rats, five males and five females were screened at approximately 3-month intervals for rodent pathogens based on the recommendation of Felasa (6). All serological examinations undertaken during the treatment period were negative for the rodent pathogens tested.

Exposure Systems

The exposure of the rats was performed in units, each housing up to 17 animals; i.e., four units were required for each group. The GSM and DCS experiments were conducted in adjacent rooms, each having 16 units installed. The exposure unit was based on a novel concept consisting of a circular cascade of 17 electromagnetically isolated and resonant waveguides. All 17 waveguides were excited by one quarter-loop antenna placed in the center providing uniform circular excitation (maximum deviation of isotropy of _0.2 dB). The waveguides were terminated with regard to the RF field by stainless steel wires, which ensured excellent electromagnetic sealing and allowed laboratory light conditions for the animals. The arrangement enabled placing the rats in optimal H-polarization with the animal’s center of mass at 150 _ 2 mm from the shortcut, which had been found to be the optimal condition that provides reasonable homogeneity of exposure for the varying sizes of rats dependent on their age and sex. During exposure the rats were restrained in the horizontal position in metal-free polycarbonate tubes that were identical to those commonly used for inhalation toxicity testing in rats but without any metallic parts. All metallic material used for the wheel was either stainless steel or gold-plated copper. The synthetic materials were either polycarbonates, PVC or POM (polyacetal), and the rats had direct contact only with polycarbonate material. Air from the animal room was supplied to each rat by a ventilation system (1 liter/min). Environmental parameters such as oxygen, relative humidity and temperature were recorded at all times, even when no exposure was taking place. During exposure, the electric field strength inside the waveguide cavity was monitored and controlled almost continuously at two different places. The RF signal (center of the uplink band, i.e., 902 MHz for GSM and 1747 MHz for DCS) was generated by a digital GSM/DCS signal generator by Rhode & Schwarz (SMIQ02B) combined with a frame generator Digital Control Unit (DCU; Schmid & Partner Engineering AG) that also controlled the switches of the amplifiers (LS Elektronik). The RF signal generator produced GSM/DCS signals and the DCU enabled switching between discontinuous transmission mode (DTX) and non-DTX without interruption. The power control was realized using the power control of the generator. The technical details are described in ref. (7). All applied signals were compliant with the definitions of the GSM or DCS signaling standards and were designed to simulate all exposure conditions (low-frequency power envelope) as occurring during use of GSM/DCS mobile phones at maximized time-averaged exposure. Each exposure session (duration 2 h) was divided into three phases of 40 min each as described in Table 2 and illustrated in Fig. 1. Each slot was modulated with a random code. In the first phase, non-DTX mode (‘‘GSM Basic’’) was applied simulating the exposure conditions during continuous talking, i.e., one active slot per basic frame while each 26th basic frame was idle. The second phase, ‘‘GSM Talk’’, simulated a conversation, i.e., by temporal switching between the non-DTX (average time active: 2/3) and DTX (average time active: 1/3) modes. The third phase, ‘‘GSM Environment’’, simulated exposure during a conversation while moving in the environment. This included GSM features such as non-DTX, DTX, power

control, handovers, etc. according to their statistical occurrence.2 The statistical parameters of power control and handover were derived from measurements performed by France Telecom (8). Those of non-DTX/DTX modes were derived from evaluations of talking and listening phases during phone conversations at the IT’IS Foundations. In all conditions, the target maximum time slot (TS) whole-body average SAR was kept constant. Deviations from the target exposure values as well as all environmental parameters were displayed continuously on the PC screen. The software provided continuous error checking that would automatically terminate the exposure in case of major errors during exposure, e.g., large deviation from target values. All hardware communication was saved on the hard disk of the experimental computer as well as on two servers, enabling full reconstruction of every exposure session. The data were evaluated daily at IT’IS, and the evaluation sheets were sent back to RCC. This information enabled the Study Director to decide whether the systems had operated properly. A new methodology was proposed to obtain comprehensive dosimetric information for whole-body, peak spatial SAR as well as the averaged values for the most important organs. For each of the values, the uncertainty as well as the instant and life-long variations was determined (9).3 During the first year, the position of any one cage of rats on the wheel was moved clockwise by one port on a weekly basis for 13 weeks. The position on the wheel of any one rat from a cage was random within its predefined cage position. In addition, at approximately 13-week intervals, the rats were placed on different wheels according to a predefined plan. With this rotation scheme in place, all rats were positioned on each of the four wheels/group for a similar duration over the course of 1 year and were positioned at most ports of each wheel. Therefore, in terms of the dose received by the rats, any differences in the exposure signal within and between wheels were minimized. During the course of the 1-year interim necropsies (weeks 53 and 54), the rats were placed on the

same wheels as for week 52. From week 55 onward a new rotation plan was used in which the rats were placed on to the different wheels at approximately 4-week intervals. Empty animal ports of the wheels and ports at which the rats had died were loaded with a container that simulated the absorption characteristics of the rats and therefore minimized imbalances of the EMF excitations of the individual waveguides. Two males had to be killed early because they became too obese for their restraint tubes. The effect of these early terminations on the study results is considered inconsequential based on their low incidence relative to the large group size."

Wistar rats, male

HAWC