The following important studies on this subject are introduced below and assessed based on BfS standards.

Study of Fritzeet al. (1997)

Assessment BfS

Study of Tsurita et al. (2000)

Assessment BfS

Study of Finnie et al. (2002)

Assessment of BfS

Studie of Salford et al., 2003

Assessment BfS

Study of Fritze et al. (1997)

In 1997, a comprehensive study on the blood-brain barrier entitled "Effect of global system for mobile communication (GSM) microwave exposure on blood-brain barrier permeability in rat" has been conducted by the Fritze et al.

The cranial area of male rats (weighing 250-300 g) was exposed to radio frequency fields according to the mobile communication standard GMS 900 for a period of 4 hours. The specific absorption rates (SAR) in the heads of the animals were 0.3, 1.5 and 7.5 W/kg and therefore covered the area above and below the partial-body limit value of 2 W/kg. During the exposure with the two lower SAR values the typical mobile communication signals pulse-modulated with 217 Hz were used, but an unmodulated signal was used during the exposure with 7.5 W/kg.

An exposure system was used, in which 10 animals per facility were exposed at the same time. The animals were kept in a carousel of circularly arranged plastic tubes facing the source of the high frequency electromagnetic field. The fluctuation of energy deposit on the head of the test animals (SAR; specific absorption rate) caused by movement of the animals, was minimised by this approach.

To determine the permeability of the blood-brain barrier, the extravasation of albumin, a protein, from the blood to the brain was demonstrated with a specific antibody typical for albumin. The evaluation of albumin extravasations (“albumin spots”) takes place microscopically. This is the usual method to analyse the permeability of the blood-brain barrier. In addition the specimens where examined for the presence of damaged nerve cells. For this purpose, the nerve cells were stained with special dyes. Changes e.g. in shape can then be microscopically identified.

These tests were carried out on half of the animals immediately after the exposure, and 7 days later on the other half.

The aim was both to uncover effects that occur delayed, and to determine whether effects are lasting or of a temporary nature.

Summary of results:

a) Examination immediately after exposure:

Compared to the cage control and sham-exposed animals, some of the animals exposed to 0.3 and 1.5 W/kg showed a small increase in albumin extravasations (“albumin spots”). In the groups (10 animals per group) exposed to 0.3 and to 1.5 W/kg, 7 and 6 “albumin spots” were found respectively. According to the author, this means 0.7 spots/animal at SAR 0.3 W/kg and 0.6 spots/animal at SAR 1.5 W/kg. This effect was statistically not significant. A statistically significant increase in albumin spots was observed only with specific absorption rates of 7.5 W/kg and only in some of the animals.

b) Examination after 7 days:

The number of albumin spots corresponded in all groups to those of the unexposed control animals; therefore also highly exposed animals demonstrated no albumin extravasations.

c) No damaged nerve cells were found in any of the groups.

Conclusion of the authors

The authors conclude that high frequency radiation in the frequency and intensity range of mobile telephony has no or only an insignificant effect on the permeability of the blood-brain barrier.

The BfS assesses the study as follows:

The study is methodologically sound and meticulously conducted, in particular with a view to the exposure of the animals. The experiment is comprehensive and comprehendible described. For the first time, a study examined the aspect of a potential time-delayed occurrence of damages and the reversibility of an immediately occurring effect. The author’s conclusion regarding the research results of this study are comprehensible. If albumin extravasations where observed at all, they were only minimal. Indications on harmful effects on health below the limit values cannot be concluded from this study.

References

Fritze K., Sommer C., Schmitz B., Mies, G., Hossmann, K.-A., Kiessling, M., Wiessner, C. (1997), Effect of global system for mobile communication (GSM) microwave exposure on blood-brain barrier permeability in rat, Acta Neuropathol 94: 465-470

Study of Tsurita et al. (2000)

In 2000, the research team Tsurita et al. conducted studies on the blood-brain barrier of male rats entitled: "Biological and morphological effects on the brain after exposure of rats to a 1439 MHz TDMA field". The cranial area of rats was repeatedly (2 hours per day, 5 days per week for a period of 2 and 4 weeks) exposed to 1.44 GHz TDMA (Time Division Multiple Access) fields as used in Japanese mobile communication. An exposure system was used, whereby 6 animals per facility could be exposed at the same time. The animals were kept in a carousel of circularly arranged plastic tubes facing the source of the high frequency electromagnetic field.

The fluctuation of energy deposit on the head of the test animals (SAR; specific absorption rate) caused by movement of the animals, was minimised by this approach. The SAR value in the head, in other words the power absorbed by the tissue in W/kg body weight amounted 2 W/kg.

To assess the permeability of the blood-brain barrier, two different verification methods were chosen. Some of the animals were injected with a dye (Evans blue) and it was examined to what extent the dye would pass through the blood-brain barrier. In another lot of animals that were examined, the passover of a certain protein, albumin, from the blood to the brain was microscopically established with a specific antibody typical for albumin. Both methods are acknowledged verification methods for the extent of permeability of the blood-brain barrier.

In addition, the animals were examined for changes in the cerebellum. The average body-mass of the animals was determined as well, as it can be influenced by stress, for example.

The results can be summarised as follows:

The exposure showed no effect on the permeability of the blood-brain barrier, no passover of Evans blue or albumin and no effect on the investigated parameters of the cerebellum or the body mass could be determined.

Conclusion of the authors:

The TDMA field used in the study corresponded to the Japanese mobile telephony standards and did not induce any verifiable changes when using intensities which emerge from the use of mobile phones. Neither the permeability of the blood-brain barrier was affected nor was there any evidence of changes in the cerebellum or the body mass of the animals.

The BfS assesses the study as follows:

In view of the exposure of the animals the study was meticulously conducted and the experiment is described comprehensively. An advantage of the study lies in the repeated daily exposure of the animals over two longer periods (2 and 4 weeks). With those exposure conditions, no indication of damage to the blood-brain barrier could be detected. However, as only two animals of each group were tested for Evans blue and only four animals of each group were tested for albumin extravasations, the validity of the study is limited. Not enough animals were tested for an assertive conclusion.

References

Tsurita G., Nagawa H, Ueno S., Watanabe S., Taki, M., (2000) Biological and morphological effects on the brain after exposure of rats to a 1439 MHz TDMA field, Bioelectromagnetics 21: 364-371

Study of Finnie et al. (2002)

In this long-term study entitled "Effect of long-term mobile communication microwave exposure on vascular permeability in mouse brain" the Australian research team Finnie et al investigated the effect of radio frequency fields from mobile communications according to the GSM 900 standard on the blood-brain barrier in female mice. The animals were exposed to the field for one hour per day, five days per week during a duration of 104 week (for nearly two years), starting at the age of eight weeks. An exposure system was used during which several animals per facility could be exposed at the same time. The animals were kept in a carousel of circularly arranged plastic tubes facing the source of the high frequency electromagnetic field. The fluctuation of energy deposit on the head of the animals (SAR; specific absorption rate) caused by movement of the animals, was minimised like this.

The average SAR values, in other words the energy absorbed in the tissue in W/kg bodyweight amounted to 0.25, 1, 2 and 4 W/kg. To assess the permeability of the blood-brain barrier, the passover of the protein albumin from the blood to the brain was established with specific antibody typical for albumin. This is the usual analysis method to check the degree of permeability of the blood-brain barrier.

Free-moving cage control groups, sham-exposed groups as well as positive control groups where compared to the actually exposed animals. The brains of the animals were examined in 3 different sections, to detect albumin extravasations in the different areas of the brain. The number of animals in the selected groups was 21 (cage control), 38 (sham-exposed), 37 (0.25 W/kg), 39 (1 W/kg), 23 (2 W/kg) and 39 (4 W/kg), therefore between 63 and 117 specimens per group were assessed, which allows an assertive conclusion.

The results can be summarised as follows:

Albumin extravasations were minimal in all animals, a maximum of 3 individual capillaries (small blood vessels) were affected in a given brain, whereby no significant difference between the non-exposed controls and the exposed groups was detected. There was no evidence of a dose–related effect (meaning no increase in albumin extravasations with increased specific absorption rate.)

Conclusion of the authors

The conducted long-term exposure under those conditions produced negligible effects of the examined mobile telephone-type radiation fields up to 4 W/kg on the permeability of the blood-brain barrier.

The BfS assesses the study as follows:

A quantitative analysis of the specimens took place (albumin spots were counted). The analysis was undertaken by two pathologists working independently from one another, not knowing whether the preparations they analysed were taken form an exposed or unexposed animal. This ensured an unbiased assessment of the preparations. The study is particularly relevant in the assessment of long-term effects of chronic exposure to radio frequency fields. The study is methodological sound and meticulously conducted, in particular with a view to the exposure of the animals. The experimental conditions are described in detail. The authors’ evaluation regarding the research results are comprehensible. No indications on relevant adverse health effects on the blood-brain barrier below the limit value can be drawn from this study.

Literatur

Finnie J.W., Blumberg, P.C., Manavis J., Utteridge, D., Gebski, V., Davies, R.A., Vernon-Roberts, B., Kuchel, T.R. (2002) Effect of long-term mobile communication microwave exposure on vascular permeability in mouse brain, Pathology 34, 344-347

Study of Salford et al. (2003)

The research team Salford and Persson published several studies in the 1990s, describing an increased permeability of the blood-brain barrier for the protein albumin at partially very low specific absorption rates (Salford et al.). 1993, 1994, Persson et al., 1997). The effects were observed in pulse-modulated and unmodulated RF fields. The current study of the research team attracted the public’s attention in particular and will therefore be discussed here at greater length:

The study of 2003 is entitled "Nerve cell damage in mammalian brain after exposure to microwaves from GSM mobile phones". The effect of radio frequency electromagnetic fields from mobile telephony according to the GSM 900 standard was examined on the blood-brain barrier in rats. The animals were exposed once for two hours, the calculated average SAR values (the energy absorbed in the tissue in W/kg bodyweight) was 0.002, 0.02, and 0.2 W/kg. The two lower values are below the recommended limit value for whole body exposure of 0.08 W/kg. The animals were tested for albumin extravasations and neuronal damage 50 days after exposure. The passover of the protein albumin from the blood to the brain was demonstrated with a specific antibody typical for albumin. This is the usual analysis method for the degree of permeability of the blood-brain barrier. The main focus of the research was to assess damaged nerve cells (so called “dark neurones”) in the brain of the animals. For this purpose nerve cells in the prepared sections were stained with the dye cresyl violet. Changes e.g. in the cell form can then be identified microscopically.

The results can be summarised as follows:

50 days after a single 2-hour exposure, the exposed animals showed albumin passover as well as numerous abnormal nerve cells (“dark neurones”) in all examined brain areas. The occurrence of these “dark neurones” was already detected in the lowest SAR of 0.002 W/kg. At 0.02 W/kg, the maximum number of damage (ca. 2 % “dark neurones”) was reached, exposure with a 10-times higher SAR of 0.2 W/kg showed no further increase of damage. Sham-exposed animals showed either no or only sporadic albumin spots and a moderate amount of dark neurons were observed only in one animal.

Conclusion of the authors

According to the authors the results of the study are a highly significant indication for neuronal damage due to exposure to radiofrequency fields from mobile telephony. Because of the age of the investigated rats, the authors draw parallels to teenage mobile phone users in particular.

The BfS assesses the study as follows

The following comments to the methodology of the experiment have to be noted:

• No cage and positive controls were included in the study; however the control with sham-exposed animals crucial for the assessment of the field effect was carried out.
• The microscopic analysis was carried out blinded, by two pathologists working independently from one another, not knowing whether the preparations they analysed were taken from an exposed or control animal. This ensured an unbiased assessment of the preparations.
• Information was missing on how many sections per animal were analysed. As only a few animals (8 per group) were examined, this information would have been useful in assessing the statistical significance, also in comparison to the other studies mentioned here (Finnie et al., Fritze et al. and Tsurita et al.), which disclose the number of prepared sections per animal.
• The “dark neurons” were analysed semi-quantitatively, therefore the preparations were divided into three damage groups and classified as: 0 (no or occasional dark neurones), 1 (moderate occurrence of dark neurones) or 2 (numerous occurrence of dark neurones). This procedure is more subjective as compared to a quantitative analysis (counting). A quantitative analysis would have increased the significance of the study.
• A quantitative analysis of albumin spots, was not presented. Only two figures are published, but with insufficient descriptions, in particular data on SAR values was missing. One figure that was originally used with regard to albumin passover was exchanged by the authors after international criticism, as it was capable of causing misunderstandings amongst viewers. The figure published originally was taken from a 10 year old study by the same author (Salford et al., 1994, Microscopy Research and Technique 27: 535-542) and showed a preparation immediately after exposure with 3.3 W/kg (40 times above the limit value). Consequently it was unsuitable to document the results of the current study.
• In all other research studies presented here (Finnie et al., Fritze et al. and Tsurita et al.) exposure systems were used which allowed the exposure of several animals at the same time. The animals were kept in a carousel of circularly arranged plastic tubes facing the source of the radio frequency electromagnetic field. By this arrangement the fluctuation of specific absorption rates caused by movements of the animals was minimised. In contrast, the Salford et al research team used so-called TEM cells. A TEM cell is a closed waveguide, in which Transverse Electromagnetic Waves spread out. The rats are placed in plastic trays of 12 x 12 x 7 cm in size within the TEM cell (15 x 15 x 15 cm) and are able to move and turn around in this tray, presumably to a varying extent, depending on their size and weight. Compared to the "carousel set up" described above, the fluctuation range of SAR values is greater here. If two animals were exposed at the same time in a TEM cell, the distance between the head and the radiation source would vary, as well as the energy absorbed by the head of the animals, depending on whether the animals were on a slot above or below the radiation source. The research does not include data on how many animals were exposed at the same time as well as the period during which the exposure of the 32 animals took place.
• Male and female rats of different ages (between 12 and 26 weeks) and different weights (between ca. 191 and 373 g) were analysed. As size and volume of the animals influence the field distribution in the TEM cells and the amount of energy actually absorbed by the animals, a greater uniformity would have been desirable, if only to enable the reproducibility of the study.
• The presented SAR values are indirectly calculated not measured. The authors refer to old studies (Martens et al., 1993 and unpublished references (Malmgren, 1998). It is not comprehensible to what extent the old calculations are relevant to the currently conducted experiments. Furthermore it is not clear to what extent the improvements in numerical and experimental methods developed since 1993 have been taken into account. Overall, a weak point of the research is the fact that the evaluation of the actual exposure of the animals is either not adequately conducted or at least not adequately documented.

The results of the authors are assessed as follows:

Originally, the expression “dark neurones” referred to the dyeability of damaged nerve cells with argentiferous compounds. The cause of this increased dyeability is not clear. Their occurrence is seen as an indication of nerve cell damage. However this can have a variety of causes, including experimental causes such as incomplete fixation of preparations, mechanical impact and posthumous vibrations. Provided that the experimental implementation and processing of the preparations for sham-exposed and exposed animals was exactly the same, these possible sources of error should be of no importance. In this respect, the methods described gave no visible indications that the nerve cells were damaged during processing (preparation-related artifacts).

However, Vohra et al. (2002) describe the occurrence of dark neurones as a "so far unclarified age phenomenon" and point out that the amount of dark neurones in rats aged 6 month is nearly twice the amount of that in 3-month old animals. As the animals examined by Salford et al. comprise this age span, and there is no data on age distribution between controls and exposed animals at the time of the experiment, aging processes cannot be excluded as cause.

Salford et al. hold the view that the nerve cell damage, visible as dark neurones, was caused by albumin uptake. However they do not mention, whether a spatial correlation was detected between albumin spots and dark neurones. An evaluation would have also been desirable as to how far the albumin uptake would have to be increased, compared to normal physiological conditions, in order to cause such massive neuronal damage.

The albumin spots found after 50 days (more than 7 weeks) cannot be caused directly by the albumin extravasations due to the EMF exposure. All proteins are broken down within the cells after a short time, so that after 50 days the original albumin passover could not be detected. Consequently, Salford et al. suggest not further specified "secondary processes“,as a long-term consequence of the original damage. However, in the above mentioned study by Fritze et al., no albumin spots and no signs of neuronal damage could be detected seven days after exposure with 7.5 W/kg. Unfortunately, Salford et al. do not discuss these conflicting results.

The discussion section of the study generally lacks argumentation with findings of other research teams. This discussion would have been especially necessary, as the study confirms previous own studies, but shows findings that are in conflict with other research teams.

In conclusion, it has to be said, that the study describes effects, which if reproduced, would be relevant to health. However, the study is weakened by a number of inaccuracies and ambiguities in its experimental methodology, analysis and assessment of results. The nerve cell damage documented in the study that occurred 50 days after a singular 2 hour exposure with a radio frequency field below the limit values is not compatible with the scientific evidence overall. Despite its shortcomings, the study will be taken into account as an indication for relevant adverse health effects below the limit value, as the observed effects are serious. Consequently, the necessity arises, to test Salford et al.’s results in similar studies by other research teams, avoiding the above mentioned failings, in particular with regard to the exposure conditions.

For this purpose, two studies (in vitro - on a cell culture model and in vivo – animal test) are conducted as part of the German Mobile Telecommunication Research Program (DMF). Their purpose is to assess the influence of radio frequency electro magnetic fields on the blood-brain barrier. In parallel, other countries have commissioned and some have already started replication studies.

References

Salford, L.G., Brun, A.E., Eberhardt, J.L., Malmgren, L., Persson, B. (2003); "Nerve cell damage in mammalian brain after exposure to microwaves from GSM mobile phones", Environmental Health Perspectives, 111, no. 7, 881-883

Martens, L, Van Hese J, De Sutter D, De Wagter C, Malmgren L, Persson BRR (1993); "Electromagnetic field calculations used for exposure experiments on small animals in TEM-cells", Bioelectrochem Bioenerg 30: 73-81

Malmgren L. (1998); "Radio Frequency Systems for NMR Imaging: Coil Development and Studies of Non-Thermal Biological Effects"; (PhD thesis); Lund, Sweden: Department of Applied Electronics, Lund University

Vohra BPS., James TJ, Sharma, SP, Kansal, VK, Chudhary, A., Gupta, SK (2002); "Dark neurons in the ageing cerebellum: their mode of formation and effect of Maharishi Amrit Kalash"; Biogerontology 3: 347-354