GERoNiMO: An interview with project leader, Professor Elisabeth Cardis

I would like to cordially thank Elisabeth Cardis for taking time and answering a set of 11 questions.

Dariusz Leszczynski: In the past epidemiological studies there were serious problems with the exposure data. Problems were not only with what persons remember. There were also problems of serious misclassifications of exposures caused by scientists designing studies (e.g. Danish Cohort, Million Women Study). Will GERoNiMO be void of such problems?

Elisabeth Cardis: While no study is completely void of problems, a considerable amount of work is foreseen to improve exposure assessment for existing epidemiological studies included within GERoNiMO (as well as for health impact assessment as appropriate data for that purpose is sorely missing). As you know, GERoNiMO builds upon birth and childhood/adolescent cohorts (in five countries: Spain, Denmark, Norway, Switzerland and the Netherlands) as well as on two multinational case-control studies Mobi-Kids and INTEROCC.

To improve exposure assessment, specifically, we are extending Mobi-Expo to a much broader population. This will provide us with the information to not only describe patterns of use but also validate questionnaire responses (how well participants recall their phone use) in studies such as Mobi-Kids.

We will collate, critically review, and summarize all available measurement data on occupational RF and IF sources so that estimates of exposure level by source (and associated uncertainty) can be assigned to all sources reported in INTEROCC and Mobi-Kids, thus developing individual estimates of occupational exposures to these frequency bands and characterizing related uncertainties.

Environmental RF exposure from RF emitters in the general environment (base stations, radio and TV emitters, etc.) will be estimated using a 3D propagation model already applied and validated in Switzerland and the Netherlands, which will be extended to (and validated in) other countries, thus providing spatial estimates of RF field strength for use in birth cohorts.

Surveys of personal exposure to RF will be conducted, nested within birth cohorts to facilitate linkage with information already collected on sources of RF for cohort subjects. For all present RF sources (also those not covered by personal exposure meters) and for IF, spot measurements and 48-hour logs of RF and IF will be taken in shops, supermarkets, public transport, and libraries to characterise exposures in public places. A sample of volunteer homes will also be surveyed to provide residential data, with measurements in bedrooms and other places where people spend most of their times and where exposure-relevant appliances are present. The information from these measurements will be crucial in validating the models used to assign exposure estimates to study subjects in the birth and childhood cohorts and MOBI-Kids.

We will bring all this information together with the information collected in the epidemiological investigations (e.g. birth cohorts and MOBI-Kids) to derive exposure estimation algorithms for estimating individual exposures and related uncertainties for all subjects in the epidemiological studies in WPs 1-3. As not all information on relevant exposure sources is available for each subject in each study, imputations and sensitivity analyses will be performed to study the impact of more limited information being available in the epidemiological investigations.

Thus, in summary, the exposure work in GERoNiMO will be much more sophisticated than what was done in the studies you mention and we are making every effort to classify subjects correctly

Dariusz Leszczynski: Predictive value of the potential health risk by the past epidemiological studies is very limited. This is caused by both poor exposure data and/or by low numbers of cases. It is stated that databases from different countries will be pooled together to get sufficient numbers of cases for statistical evaluation. However, how probable it is that databases set for different purposes will be possible to unify into a single large and reliable database?

Elisabeth Cardis: There has been a lot of work already within the European network of birth cohorts (through the ENRIECO and CHICOS projects) to review information collected and harmonise procedures across cohorts (1,2). These projects have also pooled data from many countries successfully for topics other than RF(3).

An important part of WP1 – and of the discussions at the kick-off meeting this week – is task 1.1 to standardize the assessment of outcomes between the studies. As indicated in that task, CREAL has gained experience through these and other projects in creating harmonized databases and will follow standard definitions and protocols developed in these previous projects.

In the other epidemiological studies included in GERoNiMO – both Mobi-Kids and INTEROCC were set-up from the start as collaborative projects using standardized core protocols and procedures. INTERPHONE is the parent study for both, so although the age range is different, the methods are quite similar.

Dariusz Leszczynski: How reliable will be results obtained from cohorts where exposure data is collected by asking parents? Will this cause similar recall bias as this hampering Interphone data?

Elisabeth Cardis: The advantage of the cohort studies in GERoNiMO is that we are talking about prospective follow-up. Questions have already been asked in the birth and childhood/adolescents cohorts and the cognitive and neurodevelopment testing will be done years after that information was collected, thus avoiding the potential recall bias which could affect, for example, the Divan studies.

Dariusz Leszczynski: It seems that, in the description of the background of epidemiology, studies from Sweden by Hardell group were omitted. Why?  (or did I miss something?)

Elisabeth Cardis: We are not specifically referring to any study on RF and brain tumours, just to the RF evaluations of IARC and EFHRAN, which were based on the INTERPHONE and Hardell studies.

Dariusz Leszczynski: Animal studies feel like waste of time and money in situation when we do not have solid data showing that RF or IF affect human physiology. Why not to perform instead human volunteer studies and find out whether exposures affect normal physiology of human body?

Elisabeth Cardis: The call asked for the study of a number of endpoints which we cannot at present study in epidemiological or human volunteer studies – aging, neurodegenerative diseases, reproductive effects. The use of existing animal models is therefore an important complementary approach to epidemiology in order to study these endpoints (which were listed in the EU call).

Animal studies provide information that is highly valuable in human health risk assessment. An important strength (in comparison to epidemiology) is that causality can be easily shown. Human volunteer studies cannot be used to study adverse health effects (e.g., neurodegenerative diseases, reproductive and developmental effects…)

Dariusz Leszczynski: Would it not be better to wait for the results of NTP animal study in USA before spending time and money on yet another animal study of limited value for predicting human health risk of low level exposures?

Elisabeth Cardis: The animal studies planned in GERoNiMO focus on different end-points than the NTP study – aging, neurodegenerative diseases, reproduction. Again, animal studies provide information that is highly valuable in human health risk assessment.

Dariusz Leszczynski: Metabolomics is included in the project even though none among the GERoNiMO scientists published in this area before. It means that there is no real hands-on expertise to properly analyze and explain the results. How confident you are that use of metabolomics will bring reliable results by “non-experts”?

Elisabeth Cardis: The GERoNiMO description of work and website currently only mention the names of Work Package and Module leaders. There will be other researchers involved and the project will rely on top level equipment and expertise available at the Metabolomics Centre of the University of Eastern Finland. More details about the participating scientists and details of the work will be provided in the coming months on the GERoNiMO website.

Dariusz Leszczynski: Proteomics and transcriptomics are completely omitted, even though there is data suggesting the RF affects gene and protein expression/activity. Why, in spite of early plans to use them, proteomics and transcriptomics were excluded?

Elisabeth Cardis: The EU call text specifically asked for novel approaches. We therefore offered to do metabonomics as this is one emerging and promising approach in environmental toxicology that has not been used so far in bioelectromagnetics, by contrast to transcriptomics and proteomics. Moreover, one can consider that metabonomics is integrating the activation of at least a subfamily of proteins, i.e. the proteins involved in metabolism.

Dariusz Leszczynski: Description of the systems biology approach is difficult to comprehend. It might be for the reasons of limited space or because of inexperience (no publications) in this area of the scientists leading this part of the GERoNiMO project. How reliable results will be?

Elisabeth Cardis: The text had to made very short, so it is understandable that it is difficult to comprehend. We have collaborators that are experienced (and have published) in using this approach. Again, more detailed descriptions of the approaches and partners will be published in the coming months on the GERoNiMO website.

Dariusz Leszczynski: Why in vitro experiments use only rodent cell cultures. It would be much more informative to use human primary cells that are available for purchase? Why settle on rodent cells? Fact that the rodent cells  are available in labs performing this part of GERoNiMO is not good scientific justification why not to use human cells.

Elisabeth Cardis: The format of the document sent for evaluation did not allow giving too many details.

The objective of the in vitro studies is to screen for the effect of RF and IF on a number of primary cell types and thus test for (i) the response of normal cells to RF and IF and (ii) the possibility of different cell type sensibility. The use of primary cells is of upmost importance when looking at induced genomic instability for example, since cell lines are known to be hyperdiploid.

The rationale to use rodent primary cells is that (i) previous results were obtained using rodent primary astrocytes, (ii) human primary brain cells are barely or even not commercially available, and (iii) inter-individual differences are much more pronounced in human primary cells than in rodent primary cells.

We do, however, plan to use some human cells, such as the SH-5YSY neuroblastoma cell line. Human primary skin cells (fibroblasts, keratinocytes) can also be easily tested.

Finally, to our knowledge, risk assessment is using laboratory work as a whole, with rodent and human cells as cellular models and there is no scientific evidence that rodent and human cells can respond differently to RF and IF fields.

Dariusz Leszczynski: Exposure equipment is not ready for the experiments and some exposure chambers are scheduled to be built and tested in the project.  In the past such arrangements caused time-table problems when the  building and testing dosimetry of exposure chambers was delayed for  technical problems. How confident you are that biological experiments will not be delayed by technical problems with building exposure set-ups?

Elisabeth Cardis: The Radio frequency exposure setup to be used within the GERoNiMO already exists and only requires reconfiguration for the new frequency and experimental configuration. For this reason we bypass the stages of development that have been known in the past to provide delay in system delivery, so we can be confident concerning timely delivery and operation within this project.

The set-up for IF exposure in the frequency bands that will be used for the experimental work is also available.

I am sure that different persons would like to get answers to different questions. I hope that this set of answers, together with the earlier comments on GERoNiMO work-packages, will help to understand better the advantages and the limitations of this new European EMF project.

Advertisement