Concerns About Science Justifying the Safety of AquAdvantage Salmon

Concerns About Science Justifying the Safety of AquAdvantage Salmon have arisen since the data was made public by the U.S. FDA in early September 2010. The science was submitted to the FDA by the maker of the AquAdvantage salmon, AquaBounty Technologies. AquAdvantage salmon, which - if approved - will be the first genetically engineered animal to directly enter the U.S. food supply, grow to market size in 16 to 18 months instead of the usual 30 required for Atlantic salmon.

In early September 2010, the FDA announced findings that the GE salmon is safe to eat based on AquaBounty's application. At this time, the FDA released its complete assessment of AquaBounty's safety testing of AquAdvantage salmon in a 180 page document. Outside experts have criticized the science described in the document as "sloppy" and "misleading."

The document begins by identifying the product, AquAdvantage salmon as "Triploid hemizygous, all-female Atlantic salmon (Salmo salar) bearing a single copy of the α-form of the opAFP-GHc2 rDNA construct at the α-locus in the EO-1α lineage." In other words, the salmon will each have three complete sets of chromosomes instead of two, and all of the fish will be females.

Molecular Characterization
This section of the document describes the transgenes inserted into the salmon's DNA. The GE salmon will have three sequences inserted: 1) an anti-freeze protein gene from Ocean Pout; 2) a growth hormone gene from Chinook salmon; and 3) "synthetic linkers." The document describes this by saying, "The rDNA construct is for expression of Chinook salmon growth hormone under the control of an ocean pout promoter." However, AquaBounty does not provide information on the actual sequence of the genes that were inserted.

Deformities in Fish
AquaBounty's application to the FDA included details on studies examining risks posed to the GE salmon. A total of 48 fish (six per gender in each of four categories: AquAdvantage diploid, AquAdvantage triploid, non-GE diploid, and non-GE triploid) were chosen from an initial pool of 400 to 800 fish. After the fish were selected, they were raised to maturity and then assessed for physical abnormalities as adults. Both diploids and triploids of GE and non-GE were examined to determine if any abnormalities were due to genetic engineering or if they were due to the extra set of chromosomes in the triploid fish. It appears that the data provided comes from 2007, although this is not explicitly stated in the document.

Study Flaws
Consumers' Union Senior Scientist Michael Hansen criticized this study for its small sample size (6 of each type of fish) and the failure of the study to examine the fish at all life stages from egg until adult. He also noted that the selection of 48 fish for the study from an initial pool of at least 400 may have allowed AquaBounty to select fish it felt would provide data "proving" the safety of the GE salmon.

The FDA also noted that this study may be flawed because fish with defects were culled at an early life stage, and culling may not have been equal in each of the four study groups. They say: "According to [AquaBounty Technologies], in some cases, the non-GE lower mode siblings of a cross are culled; alternatively, a predetermined number of fish are netted out and culled, or smaller or fish with irregularities are culled. For the safety study, it is not known whether culling was comparable for all four study groups, but, in general, this practice would be expected to remove those fish with moderate to severe malformations from all sample populations well before actual enrollment in the study began. This may explain why most morphological changes in the study, independent of the study group examined, were classified as “slight” in nature. Slight irregularities are less recognizable than severe ones, thus culling of these fish would have been less likely to occur. However, even if the irregularities are recognizable, culling of the fish with slight abnormalities would not be expected to be a common practice in a commercial setting because of the negative economic consequences of doing so."

Historical Data
A second set of data compares "historical data addressing the health of AquAdvantage Salmon in several consecutive year-classes." In this case, AquaBounty did not make clear how the data was compiled or whether any fish were culled before the study began. These flaws were noted by the FDA.

This study examines diploid and triploid GE and non-GE fish from the years 2003 to 2007, ranking each fish as follows: 1 = no irregularity; 2 = slight-moderate irregularity; 3 = severe irregularity. The triploid GE fish, the version of the fish that will actually be commercialized as AquAdvantage salmon, performed poorly in many years, but quite well in 2007. The percent of triploid GE fish earning a "1" rating in each year is as follows:
 * 2003: 39.1%
 * 2004: 36.0%
 * 2005: 7.9%
 * 2006: 72.3%
 * 2007: 92.4%

AquaBounty insists that 2005 was an anomaly, either because the environment the fish were raised in was flawed or because the extra chromosomes in the triploid fish caused the problems - not the inserted genes. However, Michael Hansen notes that if the former was true (the environment was to blame), then the non-GE fish would also show poor results in 2005. Yet the results do not show that. Non-GE diploid fish earned "1" ratings as follows:
 * 2003: 94.4%
 * 2004: 100%
 * 2005: 98.7%
 * 2006: 97.8%
 * 2007: 85.1%

Non-GE triploid fish earned "1" ratings as follows:
 * 2003: 80.7%
 * 2004: 96.7%
 * 2005: 89.0%
 * 2006: 66.2%
 * 2007: 28.5%

Thus, environmental factors did not contribute to unusually high amounts of malformations in non-GE fish in 2005. Hansen notes that if the extra chromosomes were to blame, then the diploid GE fish would not show unusually high numbers of malformations in 2005. And this is also not the case. The percent of GE diploid salmon receiving a "1" rating each year was as follows:
 * 2003: 42.1%
 * 2004: 91.7%
 * 2005: 17.2%
 * 2006: 61.4%
 * 2007: 95.2%

Another potential cause for high numbers of irregularities in GE triploid salmon in 2005 noted by the document is the small sample size for that year (38 fish), compared to sample sizes in the hundreds and even over 1000 in other years. However, Hansen points out that the sample size of the diploid GE salmon in 2005 was 1586 fish, and a dramatic increase in abnormalities was observed there too. Nonetheless, the FDA dismissed 2005 data as an outlier.

Also, if 2007 was the year used for the "well-controlled study" of 6 fish of each gender from each study group described above, this data implies that perhaps the fish population in that year might have provided AquaBounty with the most favorable outcome, as there were fewer abnormalities among GE fish and more abnormalities among non-GE fish in 2007 compared to other years.

FDA Conclusion
Based on these studies, the FDA concludes: "AquAdvantage Salmon show no demonstrable differences from the comparator fish population when reared under growth conditions in [AquaBounty's Prince Edward Island] facility." The FDA does note that AquaBounty's culling practices may result in problems once the GE fish are marketed commercially. However, they propose to merely monitor data of abnormalities once the GE fish are already commercialized, which Michael Hansen points out is shutting the barn door after the horse runs out.

Chemistry and Hematology Testing
Chemistry and hematology testing was performed on the GE salmon and then compared with results of tests performed on non-GE salmon. The tests looked at the following: hemoglobin, hematocrit, platelet count, neutrophils, lymphocytes, monocytes, glucose, sodium, potassium, chloride, alanine aminotransferase, aspartate aminotransferase, total bilirubin, creatine kinase, total protein, albumin, globulin, albumin/globulin ratio, calcium, inorganic phosphorus, cholesterol, and osmolality. No data tables were provided to allow any independent assessing of the results.

FDA Conclusion
The FDA concluded as follows: "We found no clinically relevant differences in the serum chemistry or hematology values for AquAdvantage Salmon compared with contemporaneous non-GE Atlantic salmon that are clearly attributable to the GE construct."

Hormone Testing
In a 2004, AquaBounty Technologies tested the muscle and skin of diploid and triploid control (non-GE) and GE salmon for the following: growth hormone (GH) and insulin like growth factor 1 (IGF1), estradiol, testosterone, T3, T4, and 11-ketotestosterone. The tests were performed on a total of 73 fish (43 control and 30 GE salmon). The test's detection limit for growth hormone was 6.24 ng/g. All of the fish tested below the detection limit for growth hormone. From this data (or lack thereof), the FDA concluded "No differences were observed in levels of growth hormone in edible tissues at the level of quantitation for the analytical method." Michael Hansen compares this to a cop using a radar gun that cannot detect speeds below 110mph to compare a car going 30mph and one going 90mph determining that there was no detectable difference in the speeds of the cars. No differences in levels of growth hormone were observed because no growth hormone was observed at all. (A similar problem occurred in tests of T4 levels; only 2 fish per group tested above the detection limit for T4.)

Growth hormone levels are linked to levels if IGF1. Only six GE salmon and 11 non-GE salmon had levels of IGF1 above the detection limit of 2.18ng/g (parts per billion). The GE salmon had a mean of 10.26ng/g with a standard deviation of 4.971ng/g, a minimum of 3.97ng/g, and a maximum of 18.43ng/g for IGF1. The non-GE salmon had a mean of 7.34ng/g, a standard deviation of 2.818ng/g, a minimum of 3.56ng/g, and a maximum of 12.24 for IGF1.

The document noted that "initial evaluation of the results suggested that there may have been an increase in the level of IGF1 in the GE fish compared to sponsor control fish" but then dismissed the differences as only existing between diploid GE and non-GE fish. All six GE salmon that tested above the detection limit for IGF1 were diploids. A comparison of the data for GE and non-GE diploids shows that "mean IGF1 level for the GE and non-GE salmon, the range of values for the GE salmon exceeded that of the non-GE salmon by more than 10%." However, as Consumers' Union points out in comments submitted to the FDA, the data on IGF1 was manipulated to reduce the difference between the average IGF1 levels in GE salmon and the control group. As the document notes "It [IGF1] has been considered as a potential hazard for human consumption following increased growth hormone levels in food producing animals."

Allergenicity
Fish allergies are one of the eight most common allergies. AquaBounty performed a test to see if its GE salmon would differ in allergenicity from non-GE salmon.

Study Design
A total of 18 fish, six from each of three groups (diploid GE, triploid GE, and diploid non-GE), were used in the study. Then, blind-coded salmon fillets were shipped to a testing laboratory, which produced a "frozen salmon-fillet homogenate (FSFH)" for use in testing and shipped it to IBT Reference Laboratory in Kansas. Then, AquaBounty Technologies "unblinded the identities of all 18 samples." Then, sera from humans with salmon allergies to test the salmon's effects on human allergies. The FDA document describes the test, saying "This assay provides a quantitative determination of inhibition of salmon-specific IgE binding which is then used to calculate the potency of salmon allergen in muscle-skin from GE salmon relative to that in a control extract, comprised of equal volumes of all six sponsor control non-GE FSFH extracts."

Study Flaws
The FDA noted two flaws in the study. First, the sample size was very small. Second, the control group of fish were not farm-raised salmon. Also, instead of noting the "percent inhibition" (a numeric, quantitative measure of allergenicity), the study determined what it called "relative potency" as its measure of allergenicity, a term that the testing laboratory was unable to define when asked by the FDA. To correct for this, the FDA requested the raw data and used it in its analysis. Additionally, Michael Hansen notes, as an additional flaw, that the study was not a blind study.

Data and Interpretation
The data from this study was analyzed by examining the mean allergenic potency of each individual fish, because salmon is typically eaten as a fillet of an individual fish and not a mix of many different fish. The data for mean allergenic potency is as follows:
 * Non-GE Diploid: 1.69, 2.04, 2.17, 2.29, 2.38, 2.65
 * GE Diploid: 2.71, 2.90, 3.36, 3.44, 3.57, 4.23
 * GE Triploid: 1.70, 2.31, 2.75, 2.88, 2.99, 3.22

The FDA interprets this data as follows: "Although confidence in the data describing the diploid GE-salmon are low, these data indicated that four diploid GE fish had mean allergenic potency greater than 3.00 U/ml, with one fish having a mean allergenic potency value of 4.23 U/ml. Only one triploid GE salmon had a mean allergenic potency value greater than 3.00 U/ml."

However, Michael Hansen points out that every single one of the GE diploid salmon and all but 2 of the GE triploid salmon showed mean allergenic potencies higher than the highest value for non-GE diploid salmon. Coupling this point along with the small sample size of the study, Hansen feels there is good cause to perform another, more comprehensive study. In comments submitted to the FDA, Consumers' Union stated: "The data presented, although woefully incomplete, do raise a potential serious human health issue—that of increased allergenicity. If this product (GE animal) does increase the allergenic risk (e.g. an increase in allergenic potency), it should not be approved.  Data from a mere 6 salmon (e.g. GE triploids) is not sufficient nor rigorous enough to conclude that no problem exists. Because FDA’s assessment is inadequate, we are particularly concerned that this salmon may pose an increased risk of sever, even life-threatening allergic reactions to sensitive individuals.  Instead of approving this product, FDA should be requiring studies with data from many more engineered fish, not the tiny sample of six fish on which it currently bases its conclusions.  Unfortunately, even the data from those six fish raises concerns, especially given the data on six GE diploid fish that were ignored."

FDA Conclusion
The FDA concluded that "The allergenic potency of triploid ABT salmon is not significantly different from that of sponsor control diploid salmon. There is insufficient data and information to draw a conclusion on the allergenic potency of diploid ABT salmon."

Related SourceWatch articles

 * Biotechnology
 * Veterinary Medicine Advisory Committee
 * AquAdvantage Salmon
 * AquAdvantage Salmon Approval Process

External resources

 * Briefing Packet for AquAdvantage Salmon Veterinary Medicine Advisory Committee (PDF - 2506KB), U.S. FDA, September 20, 2010.
 * Background Document: The VMAC Meeting on Science-Based Issues Associated with AquAdvantage Salmon, U.S. FDA.
 * Environmental Assessment for AquAdvantage Salmon (Aqua Bounty Technologies, Inc.) (PDF - 1095KB), U.S. FDA.
 * Consumers Union’s Senior Scientist to Appear at Veterinary Medicine Advisory Committee Meeting on Genetically Engineered Salmon, September 19-20: Data Suggests Increased Allergy Risk; FDA Needs Much More Data Before Approval, Consumers' Union, September 17, 2010.
 * Letter from Consumers Union to FDA About Concerns On VMAC Committee Members for GE Salmon Hearing, Consumers' Union, September 15, 2010.
 * Peter Bridson, Re: FDA Veterinary Medicine Advisory Committee (VMAC) hearing September 19-20, 2010. AquAdvantage genetically engineered salmon, Seafood Watch Program, Center for the Future of the Oceans, Monterey Bay Aquarium, September 14, 2010.
 * "FDA Isn’t Fishing for Feedback on GE Salmon", Food and Water Watch, September 7, 2010.
 * FDA Regulatory Fish Encyclopedia (RFE), U.S. FDA.

External articles

 * Susan Heavey, Biotech Salmon Leaves Many Questions, Planet Ark World Environmental News, September 22, 2010.
 * Andrew Zajac, No agreement imminent on salmon labeling, Los Angeles Times, September 22, 2010.
 * Paul Voosen, Panel Advises More Aggressive FDA Analysis of Engineered Salmon, New York Times, September 21, 2010.
 * Bryan Walsh, Food: Why the Debate Over GM Salmon Misses the Point, Time, September 21, 2010
 * Lyndsey Layton, Fears over modified salmon voiced, Washington Post, September 21, 2010.
 * Jill Richardson, Why is the FDA About to Rubber Stamp GE Salmon?, Grist, September 20, 2010.
 * Andrew Pollack, Panel Leans in Favor of Engineered Salmon, New York Times, September 20, 2010.
 * Meredith Melnick, 'Frankenfish' May Soon Be Spawning: Is Genetically Modified Salmon Safe?, Time, September 19, 2010.
 * Jill Richardson, "The Creepy Science Behind Genetically Engineered "Frankenfish" About to Enter Our Food Supply Unlabeled", AlterNet, September 13, 2010.
 * "Will FDA Approve Genetically Modified Salmon?", All Things Considered, NPR, September 7, 2010.
 * Dan Kennedy, "Who dares question the industrial food system over GM salmon?", The Guardian, September 7, 2010.
 * Andrew Pollack, "Modified Salmon Is Safe, F.D.A. Says", New York Times, September 3, 2010.
 * FrankenFish: How Genetically Engineered Salmon Could Hurt Our Health and Environment, Food and Water Watch, September 2010.