Feb 15 2013
by Kenneth R. Foster & Lorne Trottier
Science-based medicine is great, but it all depends on how you evaluate the scientific evidence. A bad example is the BioInitiative Report (BIR), an egregiously slanted review of health and biological effects of electromagnetic fields (EMF) of the sort that are produced by power lines, cellular telephones, Wi-Fi, and other mainstays of modern life. When first released in 2007, the BIR quickly became a key document used by anti-EMF activists in their various campaigns. Early in January 2013, the BIR appeared in a major update, to extensive media coverage.
The BIR concerns possible biological effects and health hazards of electromagnetic fields in two very different frequency ranges: at extremely low frequencies ELF’s of the sort emitted by power lines and appliances, and at radiofrequencies (RFs) of the sort that are transmitted by mobile phones, Wi-Fi and a host of other technologies. Both ELF and RF fields (which are subsumed under the more general EMF) are part of the electromagnetic spectrum, which includes infrared energy, light, ultraviolet energy, as well as X-rays.
ELF and RF fields are nonionizing, in that the energy of their photons is far too low to break chemical bonds, an effect that makes ionizing radiation such as X-rays so hazardous. Fields from power lines are at 50 or 60 Hz or cycles per second; those from mobile phones and other RF communications and broadcasting systems are in the range of hundreds or thousands of MHz (megahertz or million cycles per second). Simple physics tell us that a photon of 1GHz frequency has an energy of 6 millionths of an electron volt (eV), while the average thermal energy of a molecule is 0.03 eV and the ionization energy of a chemical bond is on the order of 1 eV
There are, of course, well-established hazards from excessive exposures to ELF and RF fields, which are mainly associated with electric shock (ELF) and excessive heating of tissue (RF). Such problems, however, require exposure to fields at vastly higher levels than anything that would be encountered in ordinary life. Most countries around the world have adopted roughly similar exposure limits that are designed to protect against these known hazards.
The possibility that the electromagnetic fields at much lower exposure levels can be bad for you has been a matter of public concern for many years. Countless public, scientific, and legal battles have been waged about possible health hazards produced by fields from power lines, cellular base stations, broadcasting facilities, and other technologies, despite the fact that public exposures from such technologies are invariably far below government exposure limits.
In response to such concerns, government and other agencies have funded many studies over the years. Thousands of scientific papers have accumulated on biological and possible health effects of ELF and RF fields, going back to the mid 20th Century and even before.
In many respects this scientific literature is uneven and confused. The studies vary widely in quality, biological endpoint, and relevance to health. The literature is filled with low-quality fishing expeditions in search of effects (as opposed to studies that tested hypotheses). Many of these studies were one-shot experiments, that were not followed up or even repeated by the investigators themselves. Many studies have obvious technical flaws, typically poor dosimetry (determining how much exposure the preparation actually received in an experiment) or poor temperature control (heating is a necessary consequence of RF exposure and most biological reactions are sensitive to temperature). Many of the reported effects were small, close to the level of background variability and small compared to potential artifacts (and hence difficult to identify reliably), with no particular relevance to health. The literature suffers badly from publication bias — researchers are more likely to report having found an “effect” and less likely to publish no effect studies. As might be expected, the literature abounds with reports of “effects”, many of which are simply artifacts from poorly conducted experiments.
At the opposite end of the quality spectrum are a number of well-done, massively funded studies that follow the lines of standard toxicological assays or epidemiology studies, which were designed to provide reliable evidence in assessing possible risks from ELF or RF fields. The results of these studies have been overwhelmingly negative, failing to document adverse (or any) effects of exposures at levels below current safety limits that are in effect throughout most of the world.
This massive literature has been reviewed by numerous expert groups, who, with the exception of the BIR as described below, have consistently failed to find clear evidence for health hazards from ELF or RF fields at levels below international and U.S. limits. Luc Verschaeve (Univ of Antwerp), a noted Belgian health expert in this field, has reviewed more than 30 recent expert reviews. His review is available online.1 Links to over 30 of these expert reviews by the health agencies of virtually every industrialized country can also be found at EMFandHealth.com. Readers of this article are invited to review the findings of these reports.
But the science is hardly clear-cut. Some effects have been reported at levels below international safety limits from specialized exposure conditions, that health agencies regard as having some level of consistency. For example, several independent studies have reported minor changes in brain wave activity associated with use of a mobile phone handset. The effect has no clear health significance, and it is still an open question whether it is directly caused by exposure to RF fields or some other factor associated with exposure.
The second, and politically more inflammable, issue is whether EMF exposure is linked to cancer. The International Agency for Research on Cancer (IARC, a part of the World Health Organization) has classified powerline magnetic fields and RF energy from mobile telephone handsets as “possible carcinogens” (class 2B in the IARC terminology). This is based on weak epidemiological evidence that children living in homes near power lines, or long-term users of mobile phones, have a slightly increased risk of, respectively, leukemia and brain tumors. Thus, in IARC’s scheme of things, the evidence is sufficient to raise suspicions that a problem might possibly exist, but not enough for the agency to conclude that the fields actually (or even probably) do cause cancer.
While acknowledging the epidemiology data, health agencies have generally found them to be unpersuasive for several reasons. The studies report small increases in risk, close to statistical variability. While the increases may be “statistically significant” (unlikely to be due to chance), it is difficult to rule out possible errors or biases in the studies, of which there are many in epidemiology. The general lack of supporting evidence from the animal cancer tests showing no effect, and lack of generally accepted mechanism by which RF or ELF fields can cause any biological effects at the low level exposures considered here (apart from heating) are other stumbling blocks.
Two additional comments are in order. First, the epidemiology studies were simply not adequate to reliably detect small increases in risk after long term exposures to cell phones or powerline fields. Most of the cell phone-brain cancer studies determined exposure simply by asking subjects about their previous use of cell phones. (What reader can say for sure how much he or she used a mobile phone a year ago, much less in the distant past?).
Furthermore as we indicated in a previous article on SBM, since the IARC ruling on cell phones, three recent studies on brain cancer incidence rates have appeared. The studies show no change in incidence rates in the U.K., U.S., and Israel over the past decade despite the tremendous increase in use of cell phones. While one might argue that there is not enough time for a real effect to appear (cancer can take years to develop), the fact remains that the data are inconsistent with epidemiology studies by a Swedish group that supposedly gave the strongest indication of a link between brain cancer and use of cell phones and strongly influenced IARC’s 2B classification in the first place. In the long run we will all be dead, but there is little indication so far that it will be from use of mobile phones. It is for all these reasons that the World Health Organization issued Fact Sheet 193 following IARC’s ruling which stated the following: “A large number of studies have been performed over the last two decades to assess whether mobile phones pose a potential health risk. To date, no adverse health effects have been established as being caused by mobile phone use.”
The same scenario has been in play for the 40 years in which one of us (KRF) has been involved with the issue. Activists on the issue typically present long lists of effects from exposure to ELF or RF fields. They imply that these fields are biologically active at all levels of exposure, and by presumption are hazardous to your health at even very low exposures. Health agencies, in response, sponsor massive reviews that find no convincing evidence for any health hazard at exposure levels below international exposure limits, even as they point to inconsistencies in the data and argue for more research. Making lists of reported effects, and conducting a proper weight-of-evidence assessment of potential health risks, are two very different things.
Scientific research on health, either concerning the effectiveness of treatments or possible health risks, is invariably subject to differing interpretations even by well-intentioned reviewers. To improve the reliability of assessments, agencies generally insist on detailed protocols for admitting and evaluating evidence. The definitive Cochrane Reviews, for example, has a lengthy handbook that guide its assessments of the effectiveness of medical interventions. The manual points out, for example, “in order to minimize the potential for bias in the review process, these judgments [about effectiveness of medical interventions] should be made in ways that do not depend on the findings of the studies included in the review”.
For its own part, the World Health Organization has published a set of guidelines for the assessment of the health risks of EMF: “All studies, with either positive or negative effects, need to be evaluated and judged on their own merit, and then all together in a weight-of-evidence approach. It is important to determine how much a set of evidence changes the probability that exposure causes an outcome. Generally, studies must be replicated or be in agreement with similar studies. The evidence for an effect is further strengthened if the results from different types of studies (epidemiology or laboratory) point to the same conclusion”.
Enter the BioInitiative Report (BIR)
The latest (2012) version was released on the BioInitiative website early in January 2013. This massive 1479 page report contains 28 sections, incorporating in toto the original 2007 version together with new sections. It was prepared by a group of 29 individuals, most of them scientists who have long held controversial positions on health effects of electromagnetic fields. Two individuals (Carpenter and Sage) wrote the introductory and concluding sections. One of them, Cindy Sage, is not a scientist but a long time activist on the issue who runs a consulting firm on hazards of EMF exposure. Sage, the organizer of the BioInitiative project, also contributed sections on the inadequacy of current exposure limits and other topics.
As might be expected from a compilation of separately authored chapters done at two points in time, the BIR is inconsistent in approach and content. The 2007 chapter by David Gee (European Environment Agency) says almost nothing about EMF but is largely taken from a separate paper that discusses “late lessons from early warnings” about, for example, the tragedy of Diethylstilbestrol (a drug that was supposed to prevent miscarriage but later was found to cause birth defects in children whose mothers had used the drug). Many pages consist simply of abstracts copied from the original papers or from Pubmed (which may raise fair-use issues). Several chapters discuss the idiosyncratic theory of Blank (retired from Columbia University) that DNA is a “fractal antenna”, a theory that has received no independent scientific support, much less general acceptance by the scientific community. Nevertheless, his theory is repeatedly cited throughout the BIR as being somehow related to a mechanism by which weak RF or ELF fields can damage DNA. A relatively balanced section by Fragopoulou and Margaritis (University of Athens) offers a detailed review of “omics” studies that search for effects of ELF or RF exposures using techniques of modern molecular biology. This section emphasizes the difficulty of drawing reliable conclusions from such work, and concludes merely that changes in protein expression that some studies report following EMF exposure “might potentially explain human health hazards”. But here again, the evidence is mixed and inconsistent among studies, and the endpoints studied are difficult to relate to health.
Given the structure of the BIR, there is no way to tell how many of the 29 authors of the various sections agree with the conclusions and recommendations of the report itself — or with each other for that matter. Indeed, Henry Lai, one of the authors in the BIR, recently wrote:
I don’t think the BioInitiative Report came up with any unanimous conclusion. Each author wrote his/her chapter and the opinion in each chapter is that of the authors alone. There was no communication and discussion among the authors on the preparation of the Report. As a matter of fact, I don’t personally know some of them.
Moreover, Sage and Carpenter, authors of the introductory and concluding sections, clearly have their own political axes to grind. In a recent letter they emotionally attacked the World Health Organization and a major standards setting group (The International Commission on Nonionizing Radiation Protection, ICNIRP). The overall impression is that the BIR has been structured to give scientific support to Sage’s activist ideas.
Indeed, the BIR presents many alarming health claims. EMFs at ordinary environmental levels, the reader is told, are linked to autism, Alzheimer’s disease, several forms of cancer, genetic damage, neurological problems in children whose mothers used mobile phones, hypersensitivity reactions to EMF (with symptoms similar to those of allergies), among many others. Readers who are looking for reasons to fear electromagnetic fields will find plenty of material in the BIR to justify their concerns.
Cautionary limits and the end of wireless communications
The latest BIR (or at least the two individuals who wrote the concluding sections) proposes “cautionary” limits of 0.3 to 0.6 nanowatts (billionths of a watt) per square centimeter (nW/cm2) for exposure to RF fields – this is over 100X lower than the 2007 edition. This is roughly a million times below limits in effect in the US and most other countries at frequencies used by mobile telephone systems, Wi-Fi and other technologies.
As an example of another source of radiation, ordinary sunlight (one of the many forms of EMF) has an energy density of about 100 mW (milliwatt or a thousandth of a watt)/cm2, which is more than 100 million times stronger than the proposed BIR limit for RF energy. Sunlight contains little RF energy, but about half of the solar energy that hits the earth is in the infrared part of the spectrum, which is just above the radiofrequency region considered by the BIR.
Depending on how these limits would be implemented (a matter not discussed in the BIR), their implications might be profound. All urban areas have many places where RF signals from cellular base stations, television and radio broadcasting facilities, public safety communications systems, and other useful technologies will exceed these limits – sometimes by a very large factor. Signals from Wi-Fi devices, mobile phone handsets, cordless phones, and many other useful devices would also exceed the limits, as would transmissions from the police car driving through your neighborhood, and energy leaking from your kitchen microwave oven – the strongest RF source in most households. Radar for air traffic control would be ruled out by the recommendations. Assessing compliance with the proposed BIR limits would certainly create a lot of work for Sage Associates Environmental Consultants, and undoubtedly for legions of lawyers as well.
Selective attention to data, colloquially known as “cherry-picking” or more technically as confirmation bias, is a failure in reasoning that affects all aspects of life. Recent U.S. presidential candidate Mitt Romney recently fell victim to confirmation bias by allowing himself to be convinced by the results of (evidently) skewed Republican polls that he was about to win the election.
To guard against confirmation bias, good expert reviews incorporate safeguards to ensure that all relevant data, supportive or not of the hypothesis being tested. Thus the Cochrane Review, cited above, says “judgments [about effectiveness of medical interventions] should be made in ways that do not depend on the findings of the studies included in the review”. A review that focused only on studies that report positive outcomes of a drug (for example) and ignored no-effect studies would surely have biased and unreliable conclusions.
The authors of the BIR commit exactly this error with EMF bioeffects studies, by speculating at length about possible implications of studies reporting effects of EMF while saying little about studies that failed to find effects. Rather than taking a “weight-of-evidence approach” to put all the studies together in a coherent picture, most authors simply listed numbers of studies reporting effects (of whatever nature at whatever exposure level) in comparison with those that found none.
An egregious example is connected with a long table near the beginning that lists reported biological effects from RF energy at low-intensity levels. The authors (at least the two authors who wrote the summary sections) based their cautionary recommendations on the lowest exposure levels used in studies that reported effects – regardless of the health significance of the effects, the scientific credibility of the studies, and presence of contradictory evidence.
And here is where the cherry picking comes in. The table only includes lists of studies reporting effects, some at vanishingly small exposure levels. Studies that did not report effects, or which could not confirm studies that earlier had reported effects, are conspicuously missing.
For example, one of the effects at the lowest exposure levels was reported in 2000 by David de Pomerai (University of Nottingham) and colleagues2 (see p. 106 of the PDF). In that study, exposure to low-level microwave radiation caused nematodes (a kind of worm) to express heat shock proteins. (Heat shock proteins are “expressed” or produced by the body as a way of adapting to temperature changes, an effect that can be observed at even slight temperature increases). Not mentioned is the fact that de Pomerai retracted the paper in 2006 after he had discovered that the earlier results were an artifact due to inadequately controlled temperature.3
The BIR also fails to discuss the high quality follow up studies (including one by de Pomerai and colleagues4) that found that RF exposure levels far above those used in the earlier studies did not induce heat shock proteins in a different nematode. Health agencies in their reviews have paid little attention to the expression (or non-expression) of heat shock proteins induced by RF exposure, in part because of lack of a robust and repeatable effect and in part because of the difficulty in separating the effects of simple temperature change from any specific effect of RF. Also, one might question the relevance of a small biological effect reported in nematodes in response to mild heating to human health.
A further example: the BIR concludes: “One study reports that RF at levels equivalent to the vicinity of base stations and RF- transmission towers is genotoxic and could cause DNA damage” (see P 73 of the PDF), citing a 1998 paper by Phillips (University of Colorado) et al.5 In fact, this study reported both increases and decreases in damage to DNA after very low-level exposures. One might equally cite the study as showing a protective effect of RF exposure at low exposure levels.
More importantly, this 15-year old study is not supported by later work. A careful review in 2012 by Vijayalaxmi (Univ of Texas) of 88 studies found no clear evidence of genetic damage in cells produced by RF exposure, even though the exposure levels varied by more than a million.6 Some studies reported effects, and some did not. Most of the reported effects were within the range of “spontaneous levels reported in a large data-base” and may have been unrelated to the RF exposure. By focusing on the one positive (actually, mixed) report by Phillips et al., the authors imply that RF exposures at levels produced by wireless base stations are genotoxic, a conclusion that is not supported by many later studies and a more careful analysis of the literature.
The BIR dismisses the apparent lack of consistency among studies with the rhetorical statement: “some experts keep saying that all studies have to be consistent (turn out the same way every time) before they are comfortable saying an effect exists” (see P20 of the PDF).
To our knowledge, no expert report in this field requires such absolute consistency, which would be impossible to achieve in real experiments in any event. Instead, health agencies look for different lines of evidence to point to the same phenomena across studies, together with other criteria such as potential health relevance.
For example, the authoritative World Health Organization (WHO) Environment Health Criteria document for ELF Fields7 says, with respect to cancer risk assessment, “For laboratory studies, priority should be given to reported responses (i) for which there is at least some evidence of replication or confirmation,(ii) that are potentially relevant to carcinogenesis (for example, genotoxicity), (iii) that are strong enough to allow mechanistic analysis and (iv) that occur in mammalian or human systems.”
This concern for confirmation is related to a fundamental requirement of science, that knowledge be generalizable. A study whose results cannot be generalized to other situations or yields no predictions that can be successfully tested by others, lacks external validity and is useless scientifically.
The BIR tries to have it both ways: it extrapolates from unconfirmed (or unconfirmable) laboratory studies to make dire predictions of health significance of RF exposures to humans. In effect it assumes that the results are generalizable from laboratory studies in cells or animals to human health. At the same time it dismisses the reluctance of health agencies to consider reports of biological effects that cannot be independently confirmed, which is to say that they cannot even predict results in similar laboratory experiments by other scientists.
The BIR has long been criticized by health agencies for slant. In its devastating review of the original 2007 version,8 the Health Council of the Netherlands concluded:
“In view of the way the BioInitiative report was compiled, the selective use of scientific data and the other shortcomings mentioned above, the Committee concludes that the BioInitiative report is not an objective and balanced reflection of the current state of scientific knowledge. Therefore, the report does not provide any grounds for revising the current views as to the risks of exposure to electromagnetic fields.”
The same weaknesses are still present in the 2012 version, which moreover does not address the criticisms of the Health Council of the Netherlands or other expert groups.
It takes only a glance at Verschaeve’s article1 to realize how far out of line the BIR is with assessments of the issue by mainstream agencies. Of the more than 30 reviews that he considered, all but one did not “consider that there is a demonstrated health risk from RF-exposure from mobile telephones and other wireless communication devices.” The single exception was the BioInitiative report, which Verschaeve judged to be by far the weakest of the group of reports that he considered.
Ultimately, the question reduces to the quality of judgments on what is, by all accounts, a mixed and in places contradictory scientific literature. This confusion may conceivably be the result of a small “signal” (a real but weak biological activity of ELF or RF electromagnetic fields at typical environmental exposure levels). But alternatively, it may also reflect the fact that science is difficult and often unreliable when searching for small effects in the face of a multitude of potential experimental artifacts.
It would require a more careful and balanced assessment than provided by the BIR to distinguish between these possibilities, or to fairly assess the potential health consequences of any effects from low-level exposures to electromagnetic fields in the environment, assuming that any can be definitely established.
The web page for the latest edition of the Bio-Initiative Report has been completely redesigned. The page is dominated by pictures of natural landscapes: retreating glaciers, wetlands, coral reefs, etc. that have nothing to do with the subject matter of EMF fields. They appear to be trying to associate themselves with legitimate concerns about the environment and climate change. But the analogy with climate change is the exact opposite of the impression they are trying to make. The consensus of the vast majority of climate scientists is that human caused climate change is real. Only a small minority of climate scientists is opposed to this consensus. In the case of EMF and health, the overwhelming majority of scientists see no good evidence for health effects. The BIR represents the views of a small minority.
The first edition of the BIR was widely quoted by activist groups, but had no significant effect on public policy. The “cautionary” recommendations of the latest 2012 edition of the BIR, which are more than 100 times lower than the previous one, are made without clear scientific justification and at levels that would all but eliminate broadcasting and wireless technology. Perhaps they are hoping to gain more attention with such an extreme position. It will certainly excite the activists but it is unlikely to influence public policy any more than the first edition did.
Individuals can choose in whom to place their trust. However, governments, including health agencies, have an obligation to use the best available advice about matters of importance to the health of their populations, and the BIR falls short by a huge margin. As in U.S. presidential elections, cherry picking can lead to disastrously bad judgment.
About the authors
Kenneth R. Foster is a professor of Bioengineering at the University of Pennsylvania: http://www.seas.upenn.edu/~kfoster/kfoster.htm
Lorne Trottier is co-founder of Matrox, a manufacturer of specialized video and graphics boards www.matrox.com. He maintains the website EMF and health (www.emfandhealth.com), which provides more information on this issue
- L. Verschaeve, Evaluations of International Expert Group Reports on the Biological Effects of Radiofrequency Fields. (2012). In Wireless Communications and Networks – Recent Advances, Dr. Ali Eksim (Ed.), ISBN: 978-953-51-0189-5, InTech, Available from: http://www.intechopen.com/books/wireless-communications-and-networks-recent-advances/evaluations-of-international-expert-group-reports-on-the-biological-effects-of-radiofrequency-fields
- de Pomerai D, Daniells C, David H, Allan J, Duce I, Mutwakil M, Thomas D, Sewell P, Tattersall J, Jones D, Candido P. Cell biology – Non-thermal heat-shock response to microwaves Nature 405; 2000. PubMed
- Retraction notice. Nature 440: pg 437, 2006.
- Dawe AS, Nylund R, Leszczynski D, Kuster N, Reader T, De Pomerail DI. Continuous wave and simulated GSM exposure at 1.8 W/kg and 1.8 GHz do not induce hsp16-1 heat-shock gene expression in Caenorhabditis elegans. Bioelectromagnetics 29:92-99; 2008. PubMed
- Phillips JL, Ivaschuk O, Ishida-Jones T, Jones RA, Campbell-Beachler M, Haggren W. DNA damage in Molt-4 T-lymphoblastoid cells exposed to cellular telephone radiofrequency fields in vitro. Bioelectrochem Bioenerget 45; 1998. PubMed
- Vijayalaxmi, Prihoda TJ. Genetic damage in human cells exposed to non-ionizing radiofrequency fields: A meta-analysis of the data from 88 publications (1990-2011). Mutat Res -Genet Toxicol Environ Mutag 749: 1-16; 2012. PubMed
- WHO, Extremely Low Frequency Fields Environmental Health Criteria Monograph No.238, Geneva 2010. http://www.who.int/peh-emf/publications/elf_ehc/en/index.html
- Critical Review of the BioInitiative Report, from the Health Council of the Netherlands http://www.gezondheidsraad.nl/sites/default/files/200817E_0.pdf
20 Responses to “Picking Cherries in Science: The Bio-Initiative Report”