Have GMOs been “proven safe,” and are long-term human studies needed to confirm their safety?

The following blog post was written by Carole Bartolotto, RD, and published on the Huffington Post. It claimed that genetically engineered crops (aka GMOs, GE) are risky and that current evidence is not enough to conclude that they are safe to eat. Dr. Kevin Folta (KF), and Dr. Karl Haro von Mogel (KJHvM) respond to these claims below.

Genetically-Modified Organisms (GMOs) Have NOT Been Proven Safe

By Carole Bartolotto, RD

The resounding claim of GMO proponents is that GMOs have been proven safe. Some scientists are quite emphatic about this, such as Dr. Pamela Ronald from UC Davis, who says:

"Genetically engineered crops currently on the market are as safe to eat and safe for the environment as organic or conventional foods."

KJHvM: In science, we know that nothing is ever "proven," so to demand that science "prove" something safe is a standard that can never be met. It cannot even be met for conventional foods. That being said, the statement that GE foods are "as safe to eat and as safe for the environment" by Dr. Pamela Ronald is a conclusion based on the weight of evidence that supports that conclusion.

KF: The first paragraph shows two different uses of the word "safe". As scientists, we speak softly and never say anything can be "proven safe". that is just not the way a scientist thinks or speaks. All we can demonstrate is evidence of harm.

Dr. Ronald implements a familiar use of the word "safe". When we talk to public audiences we can say GM foods are safe, which they are. At least in 17 years there is absolutely no evidence of harm. They have an excellent safety record.

It is a dichotomy of uses where the first use is "proven safe", meaning a certainty of no ill effects. Science does not work that way. The second use is "safe to eat", which reflects our knowledge based on all data.

Dr. Roger Clemens, from the USC Department of Pharmacology, also weighs in, saying:

"They're tested and evaluated in voluminous documentation that would fill this backyard. We don't know of any health risk at this particular time."

Dr. Clemens also defends food additives, sugar, and processed foods, but I digress…

KJHvM: This is an ad hominem argument, and misrepresents what Dr. Clemens has said in the linked documents. Saying that people are hyper-focused on sugar and not on the combination of food and activity is not "defending sugar." This is an attempt to sway readers to disagree with Dr. Clemens' statement with buzzwords rather than consider the evidence that supports it.

The problem with concluding that GMOs are safe is that the argument for their safety rests solely on animal studies. These studies are offered as evidence that the debate over GMOs is over. Nothing could be farther from the truth.

KJHvM: The safety of genetically engineered crops does not rest solely on animal feeding studies, although they are a big part of it. Primarily, the safety assessment of genetically engineered foods rests on detailed compositional analyses that show that the food is nutritionally equivalent to the isoline - a variety that is genetically identical except for the absence of the transgene. Nutrients, antinutrients, and toxicants are measured and compared to the non-genetically engineered isoline, and to other varieties of the same species to make sure that the composition was not significantly altered. Some of these studies are quite sophisticated, and look at thousands of metabolites at once, which is known as 'metabolomics.' Animal feeding studies add further weight to this, by showing that the food is digested the same, has the same impact on growth and physiology as the isoline, and even impacts on the gut microbiota. But this is not all that goes into safety assessment. New proteins are assessed both in silico - which means to compare their composition and structure to a database of known allergenic proteins, and in skin pinprick tests to make sure that they do not cause allergic reactions.

Animal studies have value in that if something demonstrates harm in animals, it will also likely cause harm in humans. Although some animal studies have found harm from a GMO diet, these hotly debated studies are not the point of this article. The point is, if an animal study does not find harm with a particular substance, it could still cause harm in humans.

KF: The problem here is that this is a classical blunder, the "argument from the unknown". Sure, it could cause harm in humans, but equally, it could cause great benefit, maybe allowing us to live forever or grow a coffee/beer arm. Neither are likely, there is no evidence for them, nor plausible mechanisms that could cause them.

To a scientist, such claims would begin with plausible and testable hypotheses. Others that don't understand science and the scientific method propose non-scientific "fishing trips" for problems unknown. This is a great strategy to argue against a given product or treatment, because when nothing is found, they can easily move the goal post and say we need to look deeper.

Again, this is not how science works. We understand the chemistry and biology of the agent in question, and then develop specific end points to analyze in the test subjects.

The easy question for trasngenic crop opponents demanding might be, "What would be sufficient tests that would convince you of no harm?"

Outside of a 500 year, million person study, it may not be possible. Most of the time when I ask someone about the specific tests and end points that would satisfy them, they either say "none- Monsanto is evil" or "We should leave that to the experts." Actually, that's what we do.

A good example of this is what's happened with artificial sweeteners. The U.S. Food and Drug Administration (FDA) approved artificial sweeteners for use using animal toxicology studies. Once these sweeteners were added to the food supply -- and labeled as such -- scientists were able to do epidemiological studies (also called observational studies) in humans. Several of these studies found that artificial sweeteners are linked with negative health effects.

KF: Actually, studies in cyclamates were all done in rats back in the 1960s and were the basis for discontinued use. Because of their ubiquitous use in foods, sweeteners are scrutinized very carefully.

The introduction of aspartame was accompanied by careful consideration of the phenylalanine component. Some people have a genetic predisposition to dietary phenylalanine (an amino acid in all our bodies). Scientists knew this, and the health hazards were clearly described before the product hit the market. Because there is a clear, demonstrated mechanism of harm, products are appropriately labeled.

The Framingham Observational Study found that both diet and regular sodas are associated with metabolic syndrome (a constellation of symptoms such as abdominal obesity, high blood sugar, elevated triglycerides, and high blood pressure that are linked to an increased risk of heart disease). Yet another study revealed that diet sodas may increase the risk of diabetes. The Nurses' Health Study found that two or more diet sodas a day were associated with a 30 percent decrease in kidney function over time. Yet none of these results were found in animal studies. Clearly, there are still many unknowns about the impact of artificial sweeteners on human health.

KF: If it is both diet and regular soda, then the control has an effect like the treatment. Moreover, if it was not seen in animal studies, but both diet and regular colas "were associated with" (causality implied) metabolic syndrome in humans, it suggests that the syndrome may be due to secondary-effects of obesity (along with insulin resistance, etc), not the soda per se.

This again implies that there is a causal link between diet soda and their resident sweeteners and metabolic syndrome.

However, my feeling is that the relationship is likely less onerous. Those affected with metabolic syndrome are typically placed on diets restricting calories and sugar. Perhaps that is the real causal relationship.

KJHVM: There are some published studies where genetically engineered foods have been eaten by humans and they have been examined afterward, but these are conducted for crops that are also designed to interact with human physiology, such as nutritionally-enhanced crops. Because dosage and efficacy and side effects are an issue with these, human studies are justified, and are done as minimally invasive as possible to gather the needed evidence.

There are always unknowns, but let's not let that distract us from the things we do know.

Dr. Walter Willet, from the Harvard School of Public Health, sums things up nicely by saying:

It's difficult to make blanket statements about the safety or risks of low-calorie sweeteners because all are very different in their structure and how they work in the body. The reality is, some studies have been done in animals, but we really don't have good long-term data on humans with any of these.

And the same is true for GMOs.

Considering that biology, gene regulation and expression, and the impact of a substance on a particular gene can vary so much, it makes perfect sense that animal research is not the best model to determine the long-term health effects of GMOs in humans.

KF: Actually animal systems are excellent ways to examine how well mammalian physiology tolerates and responds to compounds. Even minor changes in physiology are good indicators of a need for further analysis either in animals or careful evaluation in human clinical trials. Depending on the process studied, various animals can provide increased sensitivity to symptom presentation because of their shorter lives and compressed developmental timelines. It is easy to observe changes that happen during transitions to adulthood or old age in a timeline of months rather than years.

In fact, Dr. Ralph Heywood, past scientific director of the Huntington Research Centre (U.K.), found that the agreement between animal and human toxicology tests is below 25 percent. He has determined that there is no way of knowing what kind of toxic effect will show up in animals versus humans.

Instead of animal studies, epidemiological studies have been identified as the best way to verify the effects of a substance and its risk to humans.

KF: After 17 years no epidemiological trends have been established between GM and health concerns... and we don't "verify the effects" in science. We test for them. Many in the anti-GM world seek to "verify the effects" which is why they gravitate to data that reinforce their opinions. The use of "verify" speaks loudly to someone's level of scientific training and desire for confirmation bias.

As an author, editorditor and reviewer of scholarly literature, such words are a major red flag. Science is a question of gathering data to test a hypothesis, not "verify" a hunch.

Ultimately, we need GMO labeling so we can do the epidemiological studies that are essential to determine their risk. Without long-term data -- in humans -- no one can make the claim that GMOs are proven safe.

KJHvM: There is no single "GMO" substance that an epidemiological study can track in the manner that she assumes.

The labeling proposals being considered will not be able to assist with epidemiological studies. There are different genetically engineered crops (maize, soy, canola, sugar beets, etc), and different kinds of traits and gene constructs inserted into each variety of the same crop - they are not all identical. The kinds of labeling proposals on the table do not label which component of the food is genetically engineered, nor do they differentiate between the different GE traits present in each crop. They just say that something in the package may be genetically engineered. The main crops that have been genetically engineered are part of a commodity system where one week your corn chips are made from GE maize, the next week, they are not. An epidemiological study needs data on what people are actually eating in order to link it to any malady.

Bartolotto spent a considerable amount of effort trying to get California proposition 37 passed, however, this proposition would have provided no more information to epidemiologists than the current situation, where foods containing these commodity crops that are not labeled “non-GMO” are assumed to possibly contain them. If the argument for labeling is based on making it possible to do an epidemiological study or two, then I recommend trying to design a labeling scheme that would actually assist with such a study rather than arguing for one that would not. Better yet, efforts to make it easier for epidemiologists to test foods for their genetic identity would be much more useful, because that would provide actual information about the status of the food than a nebulous “may contain” label.

KF: You don't need labels to develop epidemiological data. In fact, they stand to bias data collection. We can tell which children have not been vaccinated for measles. Not because they have a "no vax" tattoo or t-shirt, but because they have the disease or antibody profiles indicating prior infection.

Epidemiology keys off of health trends in a population and works back to causes. Sometimes it does anticipate and model future trends. There are no health trends related to the Bt or EPSPS traits in GM crops, and no plausible mechanism by which there could be.

Just about every disorder in human history started with specific patterns in a population that could be traced back to a given cause. Other mysterious ones, like autism, are a collection of disorders with multifactorial roots.