There is a
scientific consensus It concluded that "the combination of existing test methods provides a sound test-regime to assess the safety of GM crops." In 2010, the European Commission Directorate-General for Research and Innovation reported that "The main conclusion to be drawn from the efforts of more than 130 research projects, covering a period of more than 25 years of involving more than 500 independent research groups, is that biotechnology, and in particular GMOs, are not per se more risky than e.g. conventional plant breeding technologies." Consensus among scientists and regulators pointed to the need for improved testing technologies and protocols.
Transgenic and
cisgenic organisms are treated similarly when assessed. However, in 2012 the European Food Safety Authority (EFSA) GMO Panel said that "novel hazards" could be associated with transgenic strains. In a 2016 review, Domingo concluded that studies in recent years had established that GM soybeans, rice, corn, and wheat do not differ from the corresponding conventional crops in terms of short-term human health effects, but recommended that further studies of long-term effects be conducted.
Substantial equivalence Most conventional agricultural products are the products of genetic manipulation via traditional cross-breeding and hybridization. Governments manage the marketing and
release of GM foods on a case-by-case basis. Countries differ in their
risk assessments and regulations. Marked differences distinguish the US from Europe. Crops not intended as foods are generally not reviewed for food safety. GM foods are not tested in humans before marketing because they are not a single chemical, nor are they intended to be ingested using specific doses and intervals, which complicate
clinical study design. Regulators examine the genetic modification, related protein products and any changes that those proteins make to the food. Regulators check that GM foods are "
substantially equivalent" to their conventional counterparts, to detect any negative unintended consequences. New protein(s) that differ from conventional food proteins or anomalies that arise in the substantial equivalence comparison require further
toxicological analysis. In 1999, Andrew Chesson of the
Rowett Research Institute warned that substantial equivalence testing "could be flawed in some cases" and that current safety tests could allow harmful substances to enter the human food supply. The same year Millstone, Brunner and Mayer argued that the standard was a pseudo-scientific product of politics and lobbying that was created to reassure consumers and aid biotechnology companies to reduce the time and cost of safety testing. They suggested that GM foods have extensive biological,
toxicological and
immunological tests and that substantial equivalence should be abandoned. This commentary was criticized for misrepresenting history, for distorting existing data and poor logic. Kuiper claimed that it oversimplified safety assessments and that equivalence testing involves more than chemical tests, possibly including toxicity testing. Keler and Lappe supported Congressional legislation to replace the substantial equivalence standard with safety studies. In a 2016 review, Domingo criticized the use of the "substantial equivalence" concept as a measure of the safety of GM crops. Kuiper examined this process further in 2002, finding that substantial equivalence does not measure absolute risks, but instead identifies differences between new and existing products. He claimed that characterizing differences is properly a starting point for a safety assessment and "the concept of substantial equivalence is an adequate tool in order to identify safety issues related to genetically modified products that have a traditional counterpart". Kuiper noted practical difficulties in applying this standard, including the fact that traditional foods contain many toxic or
carcinogenic chemicals and that existing diets were never proven to be safe. This lack of knowledge re conventional food means that modified foods may differ in anti-nutrients and natural toxins that have never been identified in the original plant, possibly allowing harmful changes to be missed. In turn, positive modifications may also be missed. For example, corn damaged by insects often contains high levels of
fumonisins, carcinogenic toxins made by fungi that travel on insects' backs and that grow in the wounds of damaged corn. Studies show that most Bt corn has lower levels of fumonisins than conventional insect-damaged corn. Workshops and consultations organized by the OECD, WHO, and FAO have worked to acquire data and develop better understanding of conventional foods, for use in assessing GM foods. A survey of publications comparing the intrinsic qualities of modified and conventional crop lines (examining
genomes,
proteomes and
metabolomes) concluded that GM crops had less impact on
gene expression or on protein and
metabolite levels than the variability generated by conventional breeding. In a 2013 review, Herman (
Dow AgroSciences) and Price (FDA, retired) argued that transgenesis is less disruptive than traditional breeding techniques because the latter routinely involve more changes (mutations, deletions, insertions and rearrangements) than the relatively limited changes (often single gene) in genetic engineering. The FDA found that all of the 148 transgenic events that they evaluated to be substantially equivalent to their conventional counterparts, as have Japanese regulators for 189 submissions including combined-trait products. This equivalence was confirmed by more than 80 peer-reviewed publications. Hence, the authors argue, compositional equivalence studies uniquely required for GM food crops may no longer be justified on the basis of scientific uncertainty.
Allergenicity A well-known risk of genetic modification is the introduction of an
allergen. Allergen testing is routine for products intended for food, and passing those tests is part of the regulatory requirements. Organizations such as the
European Green Party and Greenpeace emphasize this risk. The use of genes from known allergenic sources is discouraged in the research and development of GM foods. A 2005 review of the results from allergen testing stated that "no biotech proteins in foods have been documented to cause allergic reactions". Regulatory authorities require that new modified foods be tested for allergenicity before they are marketed. The
Codex Alimentarius recommends a
weight-of-evidence safety approach for assessing the allergenic potential of genetically modified crops, including the history of exposure and safety of the source of the inserted gene, the structure of the protein, its stability to
pepsin digestion, and the degree of exposure in the
gastrointestinal tract as a function of the abundance of the
recombinant protein in the food. In the assessment of the allergenic safety of genetically modified organisms,
bioinformatics is used to analyze whether newly expressed proteins show similarity to known human allergens. This process is based on
comparing the amino acid sequences of the introduced proteins with specialized allergen databases, allowing verification of whether the protein corresponds to a previously characterized allergen or presents potential for
cross-reactivity, i.e., whether its structure is sufficiently similar to that of a known allergen to be recognized by the same
IgE antibodies in already sensitized individuals. Dunn
et al. (2017) analyzed 83 studies that compared genetically modified foods with their conventional counterparts to determine whether they were more allergenic. No studies in humans or animals were found showing that a genetically modified food is more allergenic than its conventional equivalent. Toxicologists note that "conventional food is not risk-free; allergies occur with many known and even new conventional foods. For example, the
kiwi fruit was introduced into the U.S. and the European markets in the 1960s with no known human allergies; however, today there are people allergic to this fruit." Genetic modification can also be used to remove allergens from foods, potentially reducing the risk of food allergies. A hypo-allergenic strain of soybean was tested in 2003 and shown to lack the major allergen that is found in the beans. A similar approach has been tried in
ryegrass, which produces pollen that is a major cause of
hay fever: here a fertile GM grass was produced that lacked the main pollen allergen, demonstrating that hypoallergenic grass is also possible. The development of genetically modified products found to cause allergic reactions has been halted by the companies developing them before they were brought to market. In the early 1990s,
Pioneer Hi-Bred attempted to improve the nutrition content of soybeans intended for animal feed by adding a gene from the
Brazil nut. Because they knew that people have allergies to nuts, Pioneer ran
in vitro and skin prick allergy tests. The tests showed that the transgenic soy was allergenic. Pioneer Hi-Bred therefore discontinued further development. In 2005, a pest-resistant
field pea developed by the Australian
Commonwealth Scientific and Industrial Research Organisation for use as a pasture crop was shown to cause an allergic reaction in mice. Work on this variety was immediately halted. These cases have been used as evidence that genetic modification can produce unexpected and dangerous changes in foods, and as evidence that safety tests effectively protect the food supply. During the
Starlink corn recalls in 2000, a variety of
GM maize containing the
Bacillus thuringiensis (Bt) protein Cry9C, was found contaminating corn products in U.S. supermarkets and restaurants. It was also found in Japan and South Korea. Starlink corn had only been approved for animal feed as the Cry9C protein lasts longer in the digestive system than other Bt proteins raising concerns about its potential allergenicity. In 2000, Taco Bell-branded taco shells sold in supermarkets were found to contain Starlink, resulting in a
recall of those products, and eventually led to the recall of over 300 products. Sales of StarLink seed were discontinued and the registration for the Starlink varieties was voluntarily withdrawn by Aventis in October 2000. Aid sent by the United Nations and the United States to Central African nations was also found to be contaminated with StarLink corn and the aid was rejected. The U.S. corn supply has been monitored for Starlink Bt proteins since 2001 and no positive samples have been found since 2004. In response, GeneWatch UK and Greenpeace set up the GM Contamination Register in 2005. During the recall, the United States
Centers for Disease Control evaluated reports of allergic reactions to StarLink corn, and determined that no allergic reactions to the corn had occurred.
Horizontal gene transfer Horizontal gene transfer is the movement of genes from one organism to another in a manner other than reproduction. The risk of horizontal gene transfer between GMO plants and animals is very low and in most cases is expected to be lower than background rates. Two studies on the possible effects of feeding animals with genetically modified food found no residues of
recombinant DNA or novel proteins in any organ or tissue samples. Studies found DNA from the
M13 virus,
Green fluorescent protein and
RuBisCO genes in the blood and tissue of animals, and in 2012, a paper suggested that a specific
microRNA from rice could be found at very low quantities in human and animal
serum. Other studies however, found no or negligible transfer of plant microRNAs into the blood of humans or any of three model organisms. Another concern is that the
antibiotic resistance gene commonly used as a
genetic marker in transgenic crops could be transferred to harmful bacteria, creating resistant
superbugs. A 2004 study involving human volunteers examined whether the
transgene from modified soy would transfer to bacteria that live in the human
gut. As of 2012 it was the only human feeding study to have been conducted with GM food. The transgene was detected in three volunteers from a group of seven who had previously had their large
intestines removed for medical reasons. As this gene transfer did not increase after the consumption of the modified soy, the researchers concluded that gene transfer did not occur. In volunteers with intact digestive tracts, the transgene did not survive. The antibiotic resistance genes used in genetic engineering are naturally found in many pathogens and antibiotics these genes confer resistance to are not widely prescribed.
Animal feeding studies Reviews of animal feeding studies mostly found no effects. A 2014 review found that the performance of animals fed GM feed was similar to that of animals fed "isogenic non-GE crop lines". A 2012 review of 12 long-term studies and 12 multigenerational studies conducted by public research laboratories concluded that none had discovered any safety problems linked to consumption of GM food. A 2009 review by Magaña-Gómez found that although most studies concluded that modified foods do not differ in nutrition or cause toxic effects in animals, some did report adverse changes at a cellular level caused by specific modified foods. The review concluded that "More scientific effort and investigation is needed to ensure that consumption of GM foods is not likely to provoke any form of health problem". Dona and Arvanitoyannis' 2009 review concluded that "results of most studies with GM foods indicate that they may cause some common toxic effects such as hepatic, pancreatic, renal, or reproductive effects and may alter the hematological, biochemical, and immunologic parameters". Reactions to this review in 2009 and 2010 noted that Dona and Arvanitoyannis had concentrated on articles with an anti-modification bias that were refuted in peer-reviewed articles elsewhere. Flachowsky concluded in a 2005 review that food with a one-gene modification were similar in nutrition and safety to non-modified foods, but he noted that food with multiple gene modifications would be more difficult to test and would require further animal studies. A 2004 review of animal feeding trials by Aumaitre and others found no differences among animals eating genetically modified plants. In 2007, Domingo's search of the
PubMed database using 12 search terms indicated that the "number of references" on the safety of GM or transgenic crops was "surprisingly limited", and he questioned whether the safety of GM food had been demonstrated. The review also stated that its conclusions were in agreement with three earlier reviews. However, Vain found 692 research studies in 2007 that focused on GM crop and food safety and found increasing publication rates of such articles in recent years. Vain commented that the multidisciplinarian nature of GM research complicated the retrieval of studies based on it and required many search terms (he used more than 300) and multiple databases. Domingo and Bordonaba reviewed the literature again in 2011 and said that, although there had been a substantial increase in the number of studies since 2006, most were conducted by biotechnology companies "responsible of commercializing these GM plants." In 2016, Domingo published an updated analysis, and concluded that as of that time there were enough independent studies to establish that GM crops were not any more dangerous acutely than conventional foods, while still calling for more long-term studies. The study by Ricroch, Boisron, and Kuntz (2014) conducted a review of 90-day subchronic feeding studies using foods and feeds derived from genetically modified plants, aiming to assess the ability of these tests to detect adverse effects and their usefulness in the safety assessment process. In all cases considered valid under technical and regulatory criteria, none of the studies identified signs of toxicity, clinically relevant alterations, or unexpected effects attributable to the genetic modification. Nevertheless, the authors note a growing tendency, particularly in the European Union, to systematically require such tests even when existing analyses already demonstrate
substantial equivalence, raising concerns about the overinterpretation of minor biological variations and the unnecessary use of animal studies. Giraldo
et al. (2019) conclude that although most genetically modified crops are intended for animal feed, there is a significant lack of information and specific guidelines for the safety assessment of GM
forages used exclusively as feed. The study argues that current regulatory frameworks, which were developed primarily for foods intended for human consumption, can be adapted for the assessment of forages, provided that differences in risk profiles and levels of animal exposure are taken into account. The authors further argue that the same methodological approaches used in the assessment of GM foods can be applied to feed, with appropriate technical adjustments, and conclude that adopting a new integrated risk assessment framework would make the process more efficient, reduce unnecessary evaluations, and potentially facilitate the commercialization of GM crops with associated benefits. Sánchez and Parrott (2017) argue that the 35 studies frequently cited as evidence of adverse effects of genetically modified foods and feeds account for less than 5% of the available literature on GMO safety and exhibit methodological flaws, such as the absence of verification of the actual transgenic content of the diets, lack of appropriate
controls, and inadequate feed formulation. They also highlight that some reported negative effects were not specific to the
transgene, such as
immunogenic responses observed in both GM and non-GM varieties, and that certain studies could not be
reproduced. In addition, they identify cases in which studies were presented in a selectively incomplete manner, creating the impression of nonexistent risks, or cited as evidence of harm despite the fact that their own authors did not find adverse effects. In a review published the same year,
Panchin and Tuzhikov argue that several studies frequently cited as evidence of harm caused by GMOs exhibit major statistical flaws, particularly due to the failure to properly correct for the
problem of multiple comparisons, which leads to
false positives that cannot be distinguished from random chance. They further emphasize that many of these studies relied on small
sample sizes, conducted numerous tests without predefined hypotheses, and selectively reported results, all of which increase the risk of erroneous conclusions. The authors state that conclusions about the safety of genetically modified crops should be drawn from "the totality of the evidence … instead of far-fetched evidence from single studies." A study that analyzed the
gut microbiota and
metabolite profiles in two generations of
cynomolgus monkeys fed
genetically modified maize found no significant differences in most biological indicators, and the minor variations observed did not affect physiological functions during the feeding period.
Human studies While some groups and individuals have called for more human testing of GM food, multiple obstacles complicate such studies. The
General Accounting Office (in a review of FDA procedures requested by Congress) and a working group of the
Food and Agriculture and
World Health organizations both said that long-term human studies of the effect of GM food are not feasible. The reasons included lack of a plausible
hypothesis to test, lack of knowledge about the potential long-term effects of conventional foods, variability in the ways humans react to foods and that
epidemiological studies were unlikely to differentiate modified from conventional foods, which come with their own suite of unhealthy characteristics. Additionally, ethical concerns guide human subject research. These mandate that each tested intervention must have a potential benefit for the human subjects, such as treatment for a disease or nutritional benefit (ruling out, e.g., human toxicity testing). Kimber claimed that the "ethical and technical constraints of conducting human trials, and the necessity of doing so, is a subject that requires considerable attention." Food with nutritional benefits may escape this objection. For example,
GM rice has been tested for nutritional benefits, namely, increased levels of
Vitamin A.
Controversial studies Pusztai affair Árpád Pusztai published the first peer-reviewed paper to find negative effects from GM food consumption in 1999. Pusztai fed rats potatoes
transformed with the
Galanthus nivalis agglutinin (GNA) gene from the
Galanthus (snowdrop) plant, allowing the tuber to
synthesise the GNA
lectin protein. While some companies were considering growing GM crops expressing lectin, GNA was an unlikely candidate.
Lectin is toxic, especially to gut
epithelia. Pusztai reported significant differences in the thickness of the gut epithelium, but no differences in growth or immune system function. On June 22, 1998, an interview on
Granada Television's current affairs programme
World in Action, Pusztai said that rats fed on the potatoes had stunted growth and a repressed immune system. A
media frenzy resulted. Pusztai was suspended from the
Rowett Institute. Misconduct procedures were used to seize his data and ban him from speaking publicly. The Rowett Institute and the
Royal Society reviewed his work and concluded that the data did not support his conclusions. The work was criticized on the grounds that the unmodified potatoes were not a fair control diet and that any rat fed only potatoes would suffer from protein deficiency. Pusztai responded by stating that all diets had the same protein and energy content and that the food intake of all rats was the same.
Bt corn A 2011 study was the first to evaluate the correlation between maternal and fetal exposure to
Bt toxin produced in GM maize and to determine exposure levels of the pesticides and their
metabolites. It reported the presence of pesticides associated with the modified foods in women and in pregnant women's fetuses. The paper and related media reports were criticized for overstating the results.
Food Standards Australia New Zealand (FSANZ) posted a direct response, saying that the suitability of the
ELISA method for detecting the Cry1Ab protein was not validated and that no evidence showed that GM food was the protein's source. The organization also suggested that even had the protein been detected its source was more likely conventional or organic food.
Séralini affair In 2007, 2009, and 2011,
Gilles-Éric Séralini published re-analysis studies that used data from
Monsanto rat-feeding experiments for three modified maize varieties (insect-resistant
MON 863 and
MON 810 and
glyphosate-resistant NK603). He concluded that the data showed liver, kidney and heart damage. The
European Food Safety Authority (EFSA) then concluded that the differences were all within the normal range. EFSA also stated that Séralini's statistics were faulty. EFSA's conclusions were supported by FSANZ, a panel of expert toxicologists, and the French High Council of Biotechnologies Scientific Committee (HCB). In 2012, Séralini's lab published a paper that considered the long-term effects of feeding rats various levels of GM glyphosate-resistant maize, conventional glyphosate-treated maize, and a mixture of the two strains. The paper concluded that rats fed the modified maize had severe health problems, including liver and kidney damage and large tumors. The study provoked widespread criticism. Séralini held a press conference just before the paper was released in which he announced the release of a book and a movie. He allowed reporters to have access to the paper before his press conference only if they signed a confidentiality agreement under which they could not report other scientists' responses to the paper. The press conference resulted in media coverage emphasizing a connection between GMOs, glyphosate, and cancer. Séralini's publicity stunt yielded criticism from other scientists for prohibiting critical commentary. Criticisms included insufficient
statistical power and that Séralini's
Sprague-Dawley rats were inappropriate for a lifetime study (as opposed to a shorter toxicity study) because of their tendency to develop cancer (one study found that more than 80% normally got cancer). The
Organisation for Economic Co-operation and Development guidelines recommended using 65 rats per experiment instead of the 10 in Séralini's. Other criticisms included the lack of data regarding food amounts and specimen growth rates, the lack of a
dose–response relationship (females fed three times the standard dose showed a decreased number of tumours) and no identified mechanism for the tumour increases. Six French national academies of science issued an unprecedented joint statement condemning the study and the journal that published it.
Food and Chemical Toxicology published many critical letters, with only a few expressing support. National food safety and regulatory agencies also reviewed the paper and dismissed it. In March 2013, Séralini responded to these criticisms in the same journal that originally published his study, and a few scientists supported his work. In November 2013, the editors of
Food and Chemical Toxicology retracted the paper. The retraction was met with protests from Séralini and his supporters. In 2014, the study was republished by a different journal,
Environmental Sciences Europe, in an expanded form, including the raw data that Séralini had originally refused to reveal.
Carman et al. (2013) Australian researcher Judy Carman, known for her long-standing critical stance toward GMOs, co-authored a study in which she reported adverse health effects associated with the consumption of genetically modified feed, highlighting signs of intestinal irritation in
pigs as well as increased uterine thickness. The study was criticized by FSANZ, which pointed out serious flaws in its design, execution, and interpretation of the results. The agency emphasized that, although the authors attributed gastrointestinal and reproductive effects to the GM diet, relevant
mycotoxins commonly found in grain-based feeds, such as
trichothecenes and
zearalenone, were not analyzed, even though these could explain intestinal inflammation and
estrogenic effects independently of transgenesis. FSANZ also highlighted the lack of detailed information on the composition of the control and GM diets, which prevents the exclusion of the influence of other nutritional factors unrelated to the transgenic trait. FSANZ noted that only a single GM diet was used, making it impossible to assess a
dose–response relationship, and that there was no control of feed particle size, despite the authors themselves acknowledging the sensitivity of the porcine
gastric mucosa to this factor. Group sizes and the method used to measure feed intake introduced substantial uncertainty, and mortality rates were considered abnormally high by industry standards, suggesting the presence of uncontrolled stressors. The agency also identified major deficiencies in the
anatomical and
pathological assessments: the intestines were not weighed, the intestinal mucosa and intestinal contents were not examined, and no
histopathological analysis was performed to confirm inflammation, as the mere presence of hyperaemia is insufficient to characterize it. FSANZ further highlighted the absence of the expected changes in stomach weight relative to body weight, which contradicts the hypothesis of chronic inflammation. Marked differences in the gross appearance of stomachs from animals fed the same GM diet raised the possibility of acute stress effects, potentially associated with fasting and the slaughter process, without the study providing information to rule out this source of bias. In addition, FSANZ considered it problematic that regional
lymph nodes, production performance parameters, and
haematological tests were not evaluated, even though these could have provided objective evidence of inflammation or chronic blood loss. Finally, the agency explicitly disagreed with the authors' claim that standard haematological and
serological tests are poor measures of inflammation, emphasizing that
white blood cell counts and differentials, as well as biochemical parameters, are sensitive and widely used tools for this purpose.
Detoxification One variety of
cottonseed has been genetically modified to remove the toxin
gossypol, so that it would be safe for humans to eat. ==Environment==