Sep 11 2014

19 Years of Feeding Animals GMO Shows No Harm

Often GMO critics will argue that the biotech industry is conducting a massive experiment with our food supply by introducing genetically modified organisms. The implication is that GMOs are not adequately studied, which is at best debatable, but in a way they are correct. We can look at what has happened in the 19 years of GMO use starting in 1996 to see if there have been any adverse effects.  A newly published study, Prevalence and impacts of genetically engineered feedstuffs on livestock populations, does just that. (Full study, may be behind a paywall.)

The study authors, Van Eenennaam and Young, first review the existing literature on animal feeding studies. They then review available data on livestock outcomes to see what effect feeding them mostly GMO since 1996 has had, if any.

GMO Feeding Studies

The first regulatory hurdle for safety testing of GMOs is to establish “substantial equivalence.” Researchers must show that the genetically engineered crop is essentially the same as the parent variety in all ways except for the desired introduced genetic change. The authors report:

Over the past 20 yr, the U.S. FDA found all of the 148 GE transformation events that they evaluated to be substantially equivalent to their conventional counterparts, as have Japanese regulators for 189 submissions.

The authors point out that the same testing is not required for conventional breeding or even mutation farming. It also cannot be assumed that such techniques are without risk.

There have been instances where plants bred using classical techniques have been unsuitable for human consumption. For example, the poison α-solanine, a glycoalkaloid, was unintentionally increased to unacceptable levels in certain varieties of potato through plant breeding resulting in certain cultivars being withdrawn from the U.S. and Swedish markets due to frequently exceeding the upper safe limit for total glycoalkaloid content.

They further point out that proteins, DNA, and RNA from the food we eat are digested. Whole proteins or genes from GMO have never been detected in the tissue of animals fed GMO feed. There is also no reason to suspect that transgenic genes or proteins present any more of a health risk than the countless other proteins and genes we consume. It is therefore implausible that eating an animal fed GMO poses any health risk to humans. But what about the animals themselves?

Animal feeding studies are a mainstay of testing the safety of GMO plant varieties. The authors summarize this research:

Several recent comprehensive reviews from various authors summarize the results of food-producing animal feeding studies with the current generation of GE crops (Deb et al., 2013; Flachowsky, 2013; Flachowsky et al., 2012; Tufarelli and Laudadio, 2013; Van Eenennaam, 2013). Studies have been conducted with a variety of food-producing animals including sheep, goats, pigs, chickens, quail, cattle, water buffalo, rabbits and fish fed different GE crop varieties. The results have consistently revealed that the performance and health of GE-fed animals were comparable with those fed near isogenic non-GE lines and commercial varieties.

Some of these studies were long term, conducted up to two years, and with multiple generations of animals, 2-5 generations. A recent review of these studies by Snell et. al found:

Results from all the 24 studies do not suggest any health hazards and, in general, there were no statistically significant differences within parameters observed.

Some differences were found, but they were not biologically significant and likely represent background variation. Snell et al point out that 90 day feeding trials show the same results as long term and multigenerational trials, therefore the results of the standard 90 trials are highly reliable. In other words, doing longer term trials has so far not revealed any health risks that were missed during the shorter trials.

There are only a few outliers in the feeding trials showing possible harm, such as the now infamous Seralini study. All of these studies are of dubious methodology and are not published in peer-reviewed journals (Seralini was withdrawn and then republished without peer-review).

The bottom line is that there are many short term, long term, and multigenerational feeding trials with GMO varieties and many species of test animal, many by independent scientists, all showing that current GMOs are safe and nutritionally equivalent to their non-GMO counterparts.

One Big 19-year Experiment

Van Eenennaam and Young then approach the question of GMO feeding from a different angle. Since 1996 90-95% of animal feed in the US has been GMO. Prior to 1996 0% was GMO. This offers the opportunity for a large observational study to see if the rapid and thorough introduction of GMO feed in the US resulted in any adverse health effects for the animals.

This data is observational, meaning the authors are looking at data collected out there in the world and not part of any controlled prospective experiment. Observational data is always subject to unanticipated confounding factors. However, robust observational data is still highly useful, and has the potential to detect any clear signals.

In this case the data is particularly useful for a couple of reasons. First, the number of animals for which there is data is massive – in the billions per year. Second, the industry actually carefully tracks certain outcomes, as it is necessary or critical to their business.

For example, cattle are examined both premortem and postmortem for any abnormalities, such as tumors or signs of infection or other illness. Any sign of illness and that cow is not approved for meat. The percentage of cattle that are found to have such abnormalities is called the condemnation rate, and annual condemnation rates are kept in public databases.

The authors pooled data from various such databases for various animal industries before and after the introduction of GMO into animal feed:

Livestock production statistics for the US before and after the introduction of GE feed crops in 1986 are summarized in Table 4. In all industries, there were no obvious perturbations in production parameters over time. The available health parameters, somatic cell count (SCC; an indicator of mastitis and inflammation in the udder) in the dairy data set (Figure 1), postmortem condemnation rates in cattle (Figure 1), and postmortem condemnation rates and mortality in the poultry industry (Figure 2), all decreased (i.e., improved) over time.

So, multiple health parameters for multiple animals, including billions of animals over about 15 years showed no adverse effects from the rapid introduction of GMO animal feed. If there were any significant adverse effects from GMO it seems reasonable that it would easily show up in this data.

The reason for the background improvement in health parameters is likely due to improved genetics and handling. This slow improvement over time continued without change through the introduction of GMO.


We now have a large set of data, both experimental and observational, showing that genetically modified feed is safe and nutritionally equivalent to non-GMO feed. There does not appear to be any health risk to the animals, and it is even less likely that there could be any health effect on humans who eat those animals.

In order to maintain the position that GMOs are not adequately tested, or that they are harmful or risky, you have to either highly selectively cherry pick a few outliers of low scientific quality, or you have to simply deny the science.

Here is a comprehensive list of animal feeding studies. Many of these studies are independent. The list included systematic reviews, all of which conclude that GMO feed is safe.

There is as strong a scientific consensus that GMOs do not present any novel health risk, that those in current use are safe, and that they pose no health risk to animals or humans, as there is a consensus for the safety and efficacy of vaccines or that humans are contributing significantly to global warming.

16 responses so far