When the FDA announced just before Labor Day that it found the genetically engineered AquAdvantage salmon safe, news headlines and even Alaska Senator Mark Begich called it a “frankenfish.” A closer look at AquAdvantage makes it seem unlikely that Mary Shelley could have ever dreamed up anything as wild as the fast-growing GE salmon. Even more worrisome is the science used to justify the salmon’s safety, which Consumers Union senior scientist Michael Hansen calls “sloppy,” “misleading,” and “woefully inadequate.”

If approved, AquAdvantage will be the first genetically engineered animal to directly enter the U.S. food supply — a fact that raises the stakes of the FDA’s approval process, as it sets a precedent for all future GE animals. Because of a regulatory decision in the 1980s that no new laws are needed to regulate genetically engineered foods, the FDA is actually regulating the GE salmon as a drug. The next step in the approval process will be a series of public meetings held Sept. 19-21. Already, a number of groups, including Food & Water Watch, the Center for Biological Diversity, Friends of the Earth, and Organic Consumers Association have written to President Obama, urging him to discontinue the approval process for the GE salmon. (Full disclosure: I serve on the Policy Advisory Board of the Organic Consumers Association, but I was not a part of the decision to sign onto this letter.)

The company that developed the GE salmon, AquaBounty Technologies, claims the fish grows to market weight in 16 to 18 months instead of the usual 30 required for farmed Atlantic salmon. The fish was created by inserting genetic material of both Chinook (the largest variety of Pacific salmon) and ocean pout (an eel-like fish) into the genome of Atlantic salmon. The commercialized fish will all be females, making them unable to breed. AquaBounty’s intellectual property will be further protected because the fish will be sterile, as they will all be triploids (fish with three complete sets of chromosomes instead of the usual two).

How to Make Frankenfish

To create the fish, AquaBounty begins with eggs of GE Atlantic salmon females and fertilizes them with irradiated sperm of another similar fish species, Arctic char. The eggs are then pressure-treated, causing them to produce diploid offspring (i.e. fish with two complete sets of chromosomes), with both sets of chromosomes originating from the GE female salmon. The all-female GE diploid salmon will then be treated with 17-methyltestosterone, a hormone that turns the fish into what AquaBounty calls “neomales” — genetically female fish that produce milt (sperm) instead of eggs. The milt from the GE neomales will fertilize the eggs of non-GE Atlantic salmon, and the resulting fertilized eggs will be treated with pressure to produce the final product, a line of all-female triploid GE Atlantic salmon.

According to AquaBounty’s plans, the GE salmon will begin their lives at a hatchery in Prince Edward Island, Canada and then transfer to a grow-out facility in Panama. Unlike most salmon, which begin their lives in freshwater before transferring to saltwater, the GE salmon will live their entire lives in freshwater. The good news is that currently there are no plans to raise the GE salmon in open net pens in the ocean, a method of salmon farming that has resulted in massive damage to wild salmon populations as well as frequent escapes of farmed salmon into the ocean.

Questionable Science

While farmed salmon have been an environmental catastrophe in countries like Canada and Norway, it seems that environmental concerns over AquAdvantage take a backseat to safety concerns. The science AquaBounty provided the FDA was sloppy in a number of ways, and yet the FDA accepted it and declared the fish safe. Because the approval of AquAdvantage salmon will set a precedent, it is important that the FDA set its bar for solid science high, signaling to any company that wishes to commercialize a genetically engineered animal that it must completely prove its safety if it hopes to put its product on the market. Instead, according to Hansen, “the FDA appears to have set its bar an inch from the ground.”

AquaBounty tested its GE salmon and controls for physical and behavioral problems, differences in blood test results and hormone levels, and allergenicity to humans. Although the commercialized fish will all be female triploids, they often tested both males and females and both diploids and triploids of non-GE and GE salmon to determine whether any problem that showed up was due to the genetic engineering or due to the extra set of chromosomes.

However, in many of the tests, AquaBounty used sample sizes as low as six fish, much less than the minimum of 30 needed for the results to have statistical significance. Hansen said a small sample size might make sense if the animals were elephants, but there is no reason why AquaBounty should not have tested more fish. Moreover, in one of the tests, the six fish in each study group were selected from larger groups of 100 to 200 fish, and the report did not specify that they were chosen randomly. Additionally, AquaBounty admitted to culling deformed fish prior to selecting fish for inclusion in its studies. The company justified this by saying that culling is standard practice in the industry. That may be so, but for the purpose of comparing deformities between GE and non-GE salmon, the culling and sampling practices reduce the reliability of the results.

Another alarming practice — one Hansen felt qualifies as misleading — was AquaBounty’s reliance on 2007 data (the best year for the GE fish and simultaneously the worst year for non-GE fish) and its characterization of 2005 data (the worst year for the GE fish) as an outlier to be ignored. By using 2007 data for many of its studies, AquaBounty was able to compare its best group of both diploid and triploid GE salmon against the group of non-GE salmon with the highest frequency of physical deformities (compared to each of the other years of testing, 2003-2006).

On the other hand, in 2005, the GE fish exhibited an unusually high frequency of physical deformities (only 7.9 percent of triploid GE salmon and 17.2 percent of diploid GE salmon were judged to be free of any malformations), and AquaBounty provided several justifications for ignoring this data, suggesting that perhaps the small sample size (38 fish) of GE triploids was to blame. Hansen says if that were true, we would not also see such poor results in the diploid GE fish, which had a sample size of over 1,500 salmon.

The problem could have been environmental, offered AquaBounty. Maybe the problems were caused by nutrient deficiencies, exposure to antibiotics, contaminants in feed, parasites, or water temperature. Yet, if that were the case, notes Hansen, we would also see a high rate of malformations in non-GE fish in 2005, and we do not. Both the diploid and triploid groups of non-GE fish performed well in 2005, with 98.7 percent and 89.0 percent showing no malformations, respectively. Hansen also dismissed AquaBounty’s assertion that the extra chromosomes in the triploid salmon were responsible for the 2005 data, as both the diploid and triploid GE salmon performed poorly, but the non-GE triploids performed quite well.

Despite the problems noted above, the FDA concludes from the data that, “Analyses of the behavior and gross external abnormalities of market size (1,000-1,500 g) AquAdvantage Salmon show no demonstrable differences from the comparator fish population.” One last flaw Hansen points out is the study’s examination only of adult fish, and not of fish in all life stages, beginning with the egg. The FDA, perhaps worried about this, and certainly worried about AquaBounty’s heavy culli
ng of fish in early life stages (not to mention their lack of data on fish that were culled), called for a Durability Plan that includes “monitoring, data collection, and reporting of abnormalities observed under commercial production and grow-out conditions at the Panama facility where AquAdvantage Salmon will be reared” after the fish are approved and commercialized. Hansen feels this is insufficient, comparing it to allowing the fox to guard the henhouse and report if any chickens are being eaten.

Another area where the science is flawed is in AquaBounty’s examination of hormone levels in the fish. Of 73 fish tested (30 GE and 43 control), every single fish had growth hormone levels that fell below the detection limit. Hansen criticizes AquaBounty’s conclusion that there was no detectable difference in levels of growth hormone between GE and non-GE fish, comparing it to a cop with a radar gun that cannot detect speeds below 120 mph concluding there is no evidence of exceeding the speed limit. Additionally, only six of the 73 fish had detectable levels of T4 (a thyroid hormone), and only 17 had detectable levels of insulin like growth factor 1 (IGF1), a hormone that is potentially harmful to humans. Even with the small amount of data, the GE salmon that had detectable levels of IGF1 tested nearly 40 percent higher on average than the non-GE salmon with detectable levels of IGF1.

One last area to consider is the allergenicity testing of the GE salmon, as fish allergies are one of the eight most common allergies in the United States. For this, AquaBounty used sample sizes of six, testing GE diploids and triploids against non-GE diploids. They began by sending 18 blinded salmon fillet samples to a lab that treated them with liquid nitrogen to produce “frozen salmon-fillet homogenate.” Then they unblinded the samples and tested each individual sample with sera from humans with salmon allergies and measured the magnitude of the allergic reaction to determine the “allergic potency” of the sample. AquaBounty then converted the data into an undefined estimated measure it called “relative potency,” a term the lab was unable to define when asked by the FDA.

The FDA obtained the actual data tables from the test and concluded that, “The allergic potency of triploid [AquAdvantage] salmon is not significantly different from that of [the control group of non-GE] diploid salmon.” Again, Hansen took issue with this conclusion in light of the small sample size in the study, the unblinding of the samples, and the fact that the allergic potency of all but two GE salmon were higher than the highest value of allergic potency for non-GE salmon.

Hansen felt that, while the use of actual human sera to test allergenicity was useful, it was insufficient given modern scientific techniques available to assess allergenicity. Scientists are aware of many proteins that cause salmon allergies and they could easily have analyzed the molecular structure of the fish to determine if those proteins were present. Although there was one attempt to do this for one protein, the testing technique was so crude and flawed (some of the data submitted was upside-down!) even the FDA did not accept it.

Why Consumers Should Be Concerned

Given the flawed science used to justify the safety of AquAdvantage salmon, what happens now? Currently, the FDA is preparing for its public meetings: The first meeting on Sept. 19 will review the science; the second meeting on Sept. 21 will cover labeling issues and offer an opportunity for public comment. FDA will also accept written comments until Nov. 22.

There are a few more issues for consumers to consider should the GE salmon come to market. Under current law, genetically engineered foods are not required to be labeled as such. In fact, the only labeling one can expect on a genetically engineered salmon fillet is country-of-origin labeling, which is required on most (but not all) seafood. Since all of the AquAdvantage will be produced in Panama, an uncommon location for farmed salmon, consumers can be on the lookout for — and avoid if they wish — salmon from Panama. The exceptions will be salmon sold in fish markets and processed salmon, such as smoked salmon, which do not require country-of-origin labeling.

Hansen noted two reasons why consumers may wish to avoid the GE salmon. First, he notes that reports of increased inflammation in the tissues of the GE salmon may result in increased antibiotic use. Second, consumers who care about animal welfare may wish to avoid the salmon because even in the flawed tests that were performed, the GE salmon exhibited higher rates of physical deformities than non-GE salmon.

The bigger picture, of course, is the standard AquAdvantage salmon will set for future genetically engineered animals. Even if the AquAdvantage salmon proves to be safe in the long run, if sloppy and dishonest science is all that’s required to pass a product through the U.S. regulatory system, what other disasters lie in our future?

Editor’s Note:  “The Creepy Science Behind Genetically Engineered ‘Frankenfish’ About to Enter Our Food Supply Unlabeled”  by Jill Richardson was originally published at AlterNet Sept. 13, 2010.  Republished with permission from the author.