A biotechnologist* says the media are misrepresenting the National Academy of Sciences report when it comes to GMO food safety
This week the Spanish newspaper El País published an article called "Science confirms that GMOs are as safe as other foods", commenting on the latest report by the National Academy of Sciences (NAS), titled "Genetically engineered crops: Experiences and prospects”. However, the report paints a very different picture from the one portrayed by the headlines.
First, and most importantly, the NAS doesn't hold an "umbrella" view on the safety of GMOs. Along with many other international bodies, it recognises that such a view is untenable, given that the GM process can give rise to different and unpredictable effects in each case. The committee points out that it has received "impassioned requests" to give the public a simple, general, authoritative answer about GM crops, but that given the complexity of GMO issues, it did not see that as appropriate. Following that, in a 400-page document, the committee develops its view on the issue. That view is undoubtedly more favourable to the use of GMOs in agriculture than that of most environmental groups – but it is also much more conservative than that of many GMO promoters (including El País). Far from stating that "science has confirmed that GMOs are safe", the report could be more accurately summarised as stating, "There’s a lot that we don’t know, which isn’t surprising, because no one is looking."
To judge by the methodology described in the document, the committee has been respectful of the different views on this complex issue, and has made an effort to avoid different kinds of bias. However, one important bias remains, and this is the one emerging from the relationship of several of its members with important biotechnology companies and their associated organisations. For example, the director of the study, Kara Laney, used to work for the International Food & Agricultural Trade Policy Council (funded by Monsanto). Food & Water Watch has documented the links of at least twelve of the twenty-two members of the committee with the main global biotech firms, or organisations funded by them.
The new NAS report, as with previous reports from the same body, has a noticeable "sandwich" structure: several chapters describe risks and perceived problems, limitations of the studies carried out so far, and causes for concern, in between initial and final chapters that show a much more favorable vision. It’s those initial and final chapters that have been mined, without the intervening nuances, by the main media and public relations agencies.
This effect was already marked in the 1989 NAS report, which was also used to justify a scientific consensus on GMO safety, even though it limited its remit to experimental crops and microorganisms in the continental US (not even including Hawaii or Puerto Rico). In addition, this report only considered potential environmental effects, since no studies had been carried out on potential health effects. But these factors did not prevent numerous sources from amplifying the report’s reach to all possible applications of genetic engineering, which have apparently since then been blessed by "science" as "safe".
This artificial amplification of the areas where there is scientific consensus has been happening since the very first applications of genetic engineering. For instance, in the 1970s and 80s, the technology’s proponents would repeatedly declare there was an expert consensus that GMOs are safe, even though there was no evidence that demonstrated safety and many scientists had strongly voiced their concerns. This claim about the supposed scientific consensus regarding GMO safety continues even now, though it is blatantly false. A document signed by hundreds of scientists worldwide shows that the so-called scientific consensus doesn't exist.
The lack of pre-commercialisation studies which could limit unpredicted effects from genetic engineering has been so severe that the first authorisation for unconfined cultivation of a GM maize, in 1980, was granted years before the first GM maize plant was even produced – let alone tested. In the early years of this technology, many of its proponents argued that GMOs could not harm the environment because the genetic engineering would necessarily impair the transformed organisms and render them unfit to survive in the wild. Some even argued that GM crops wouldn't be capable of cross-pollinating or that any GM virus or microorganism would be safe if the parent line was safe. As still happens today, many of these statements were made with no studies whatsoever to back them up – and they have been shown, over time, to be false.
NAS recognizes lack of consensus on GMO safety
The NAS report recognizes and engages with the lack of scientific consensus about GMO safety and the possibility for unexpected changes due to the genetic engineering process. As in their last report, they point out that unexpected effects can also emerge from other techniques considered inside conventional breeding, like radiation- or chemical-induced mutation breeding (also called mutagenesis).
This technique has been used since the mid-20th century, and has been used to generate new elite crop varieties since the Green Revolution. Its possible adverse effects have not been studied. When writing the Directive 2001/18 to regulate GMOs, the EU determined that organisms developed through mutagenesis were considered GMOs, but that it was not necessary to label or test them before commercialisation because of their history of safe use. The questions of whether this decision was right, whether 50 years amounts to a history of safe use, and whether it would be possible to evaluate the possible effects of these crops after this period, are beyond the scope of this article.
In any case, the 2004 NAS report describes how mutagenesis and different techniques used in genetic engineering give rise to unintended changes throughout the genome at much higher levels than other "classical" breeding techniques. So it is confusing that some of the statements in the new NAS report treat all conventional breeding techniques (including mutagenesis, which the EU considers to give rise to GMOs) as the same, since according to the 2004 NAS report there is such a big difference between mutagenesis and all other conventional breeding techniques.
The report describes two sources of unintended differences related to genetic engineering which could affect food safety:
1. Unintended effects of the targeted genetic changes on other characteristics of the food (for example, the intended presence of or increase in one compound in plant cells could result in changes in plant metabolism that affect the abundance of other compounds).
2. Unintended effects associated with the genetic engineering process (for example, DNA changes resulting from plant tissue culture).
This means that the introduced traits could have effects other than the desired ones, and also that the processes of transformation and tissue culture could give rise to changes in other areas of the genome. When referring to the changes derived from tissue culture, the report talks about both genetic and epigenetic changes. Just as some years ago we wouldn't have thought about detecting epigenetic changes (and even today we know surprisingly little about epigenetics, according to the researchers involved), it is possible that whole new levels of gene regulation could exist that we haven't discovered yet. Clearly we cannot measure the possible impacts with current techniques.
Current safety assessment methods not adequate
Since these effects could occur when using these techniques, the next logical question is if our methods to detect and avoid them are good enough. The idea that GM foods are analysed in a very strict and comprehensive way has been widely promoted. However, the report points out that current assessments have deficiencies. Examples given include:
• In a discussion about the evaluation of a Bt crop, the report says it's not the US EPA (a public body) that carries out the assays; they are done by the company, which sends its results to the EPA. The same thing happens in the EU with the European Food Safety Authority (EFSA). Raw data from these studies are not published or available to the scientific community and general public. In fact the committee points out that it too did not have access to these data, which are protected under commercial confidentiality agreements. In other words, for an adverse effect to have been found in a pre-commercialisation test, the very same company that intends to market the product would have had to detect the effect and notify the relevant public body. This makes no sense from a business point of view. The most logical thing, if a product had found an adverse effect in pre-commercialisation assays, would have been for the company not to have pursued the application, so we would have never know about it. Or if the effect was subtle and was not detected by the methods used (or chosen), the product would have passed the evaluation and entered the food supply.
• Internationally accepted protocols use small samples with a limited statistical power, which may not detect differences between treatments, or they might find statistically significant differences that then would not be considered biologically relevant.
• Data obtained by studying cattle for long periods of time, even if they do not reveal adverse effects, cannot be extrapolated to chronic effects in human beings, because of, among other reasons, the young age at which these animals are killed.
• Regarding the a priori evaluation of changes in the levels of "known" toxic substances: The toxic properties of some plant compounds are understood, but most have not been studied.
• Detection of allergies to new proteins (those produced by the introduced gene or by a different gene which has been altered as a result of the GM transformation and/or tissue culture process) cannot be guaranteed with the currently used methods; post-commercialisation studies would be needed.
• The studies which have been carried out have found differences between animals given GMO and non-GMO feed; these differences were statistically significant (i.e. they were not caused by chance but by the treatment), but they weren't considered biologically relevant. However, what would be considered "biologically relevant" was not defined beforehand, and the statistical power of the studies had not been calculated. The differences found could have meant that adverse effects existed yet the methodology wasn't able to detect them. So the report is right to say that even when no adverse effects have been found, this doesn't mean that they don't exist – a point ignored by many media headlines on the report. In one of the cases described in the report, in fact, a feeding study was carried out with a type of rice in which a gene had been introduced to produce a known toxin (as a positive control), and no adverse effects were found. Because of this, the report points out the need to carry out further studies with a corrected methodology: the studies done so far that are said to show no adverse effects cannot offer conclusive data regarding safety.
• Data and studies currently available cannot be used to draw conclusions on possible long-term effects on human health. However, and given the perceived concerns of people surveyed, the committee makes an effort to use the available data (which, they point out, are insufficient and cannot be used to obtain conclusive data) to detect possible changes in the incidence of different chronic diseases. However, as the committee points out, this is a very rough approach to detecting such problems. To reliably detect them, post-commercialisation studies would be needed, which control for a high number of variables so that the only difference between groups would be consuming GMO foods (or a particular GMO food) or not consuming them.
Several of these comments, and especially the last one, take us to what could be one of the key points of disagreement: current methodologies are not appropriate to guarantee safety (by “guarantee”, I mean a similar level of guarantee to the one we have with foods not obtained through techniques with a high probability of generating unintended effects, according to the scale in the NAS’s 2004 report). But guaranteeing this level of safety with the new techniques would be somewhere between very expensive and impossible. At several points, the report talks about "acceptable risks". But who decides what level of risk a given population is willing to accept? Is this necessarily a scientific decision? The decision must logically be supported by scientific data, among other things, but that doesn't mean that the whole decision-making process must confine itself to this area.
Business as usual?
The report's position, and where it deviates from the view of many environmentalists, is that when confronted with this situation, the best thing we can do is to continue marketing these foods as we have done until now, perhaps with some additional technical means to detect possible adverse effects, in the hope that we can withdraw any products which might show adverse effects in the future.
This caution-risk trade-off has been the one followed for years with regard to pesticides, and synthetic chemicals. The twentieth century shows several cases of products and technologies for which scientists had not detected adverse effects (sometimes honestly, sometimes not), and which were thought to be perfectly fine – until they weren’t. With GMOs, this trade-off has taken us to situations like the proliferation of glyphosate-tolerant weeds described in the report, or cases of transgene propagation in nature, something that was claimed could never happen. That same trade-off has also resulted in our failing to recognise global climate change until a very late point.
An alternative view would be one that limited GM techniques to confined areas (with effective levels of confinement, control and information), where scientific research could lead us to a scenario where we know enough about living systems for the "unpredictable and unexpected" effects to stop being so. In these confined areas, the consequences of using a GMO, both good and bad, would affect only those taking the decision to use it (for instance, a patient willing to use insulin produced by a recombinant organism). Meanwhile there are enough alternatives for agriculture to move forward and face the challenges ahead without the need to add the job of continually putting out fires.
Originally Published: GMWatch