There have been quite a few analyses of the Aris and Leblanc paper Maternal and fetal exposure to pesticides associated to genetically modiﬁed foods in Eastern Townships of Quebec, Canada, but since the myth is so pervasive, I thought I’d throw in my thoughts as well. Perhaps because this paper deals with blood or perhaps it’s the link to pregnant women and fetuses, it keeps circulating as a prime example of how Bt is harmful (even more so than the recent Seralini paper or the older Seralini paper).
There are perfectly valid reasons to be concerned about Bt and about genetic engineering in general, but using bad science as a reason doesn’t help anyone. I don’t know if it’s possible to combat the misinformation with careful analysis of the science but here goes, just in case it will help.
As with so many things, the details of the “Bt in human blood” saga are a bit different from what you can find in the news and on most blogs. Long story short, the science of the study is less than sound, and their conclusions do not follow from the results. Aris and Leblanc found levels of Bt below the lower detection limit of the method they used and they didn’t determine what type of Bt was present. Even if they did find Bt protein (whole or in pieces) in human blood, that does not indicate that it is a “threat to human health”. Let’s look a bit deeper.
What is Bt and is it safe?
For starters, what is Bt? It is a protein produced by common soil bacteria that has some insecticidal properties. It has been and continues to be used as a pesticide spray for decades, including in organic farming. In the past 15 years or so, the gene for Bt has been engineered into corn and cotton (and less common, potato) such that the plants produce the protein themselves.
There has been a lot of research on the safety of this particular protein, with little indication that it causes harm to mammals, birds, or fish, and are even safe for most non-target invertebrates. This is thanks to the unique chemistry of the protein. Much of the research is independently funded and spans many disciplines and experiment types. Some of this research is summarized in the following papers that each analyzed many previously published papers:
- Bt crop effects on functional guilds of non-target arthropods: a meta-analysis by scientists from the University of Nebraska in 2008.
- A meta-analysis of effects of Bt cotton and maize on nontarget invertebrates by environmental scientists from Santa Clara University in 2007.
- Environmental fate and effects of Bacillus thuringiensis (Bt) proteins from transgenic crops: a review by entomologists from Iowa State in 2005.
- The mammalian safety of Bacillus thuringiensis-based insecticides by a USDA ARS entomologist in 2001.
- Assessment of the health impact of GM plant diets in long-term and multigenerational animal feeding trials: A literature review by a group of publicly funded European scientists in 2012 (not all of the papers analyzed here were about Bt, but many were, and there are many other feeding studies in various mammals that also show no problem).
For more general background, please visit my post on Bt.
Bt in blood?
While studies like this Aris and LeBlanc paper do a really good job of sounding sciency, unfortunately they just don’t hold up to scrutiny. Did this paper actually find Bt protein in human blood? Maybe, but probably not. They used the wrong test and they used the wrong standard. Let me explain.
They used the wrong kit and probably the wrong test
The ELISA kit used by Aris and LeBlanc to detect Bt was made by a company called Agdia (as described in section 2.4. of their paper). The kit was created and tested to detect Bt in plant tissues (Agdia doesn’t make any kits for animal tissues). There’s a lot of reasons why such a kit might not work on mammalian tissues. For example, the antibodies might cross-react with proteins found in mammals that aren’t found in plants.
Their use of the plant kit is really strange, because there have been researchers that used ELISA methods to detect Bt in mammalian blood – although with varying success.
For example, German researchers developed (PubMed ID 18155416) an ELISA method for cows’ blood. They did not find any significant difference between cows that had been fed Bt and conventional maize for a two month period, and all values detected in all cows’ blood were less than 1.5 nanograms per microliter. One hypothesis for why they would not have found a difference is that the ELISA test doesn’t properly detect Bt in blood (another hypothesis is that there just isn’t any Bt in the blood to detect).
Aris and Leblanc did cite a paper (PubMed ID 15740023) that found an ELISA method could detect fragments of Bt in addition to intact Bt protein in the gastrointestinal tract (not in blood), but they neglected to mention the actual findings of the paper. Those researchers found that Bt was probably digested in cattle, but suggested that a different method besides the ELISA should be used.
Aris and Leblanc also cited a paper (PubMed ID 14552382) that used ELISA to detect Bt in the gastrointestinal tract and blood of pigs. Again, they neglected to mention the findings. Bt protein fragments and gene fragments were found in the gastrointestinal tract, showing again that Bt is digested in cattle. They were unable to detect Bt protein in blood using ELISA, immunochromatography, or immunoblot – so they weren’t able to make any conclusions about that.
They used the wrong standards
Standard solutions are often used to calibrate a test, to create what is called a standard curve (but it’s really a straight line if done right). Here is an example of a standard curve.
Absorbance measurements of known concentrations of protein were measured in duplicate to create this graph. For example, for the known protein concentration of 10 milligrams per milliliter (mg/mL), the measured absorbance is about 1.9 absorbance units. When an unknown sample is measured, the value of the absorbance units can be compared to the standard curve to tell us what is the corresponding protein concentration. Here, the unknown reads at about 0.4 absorbance units, which corresponds to 30 mg/mL of protein.
Aris and LeBlanc created Bt standard solutions of 0.1 to 10 nanograms per microliter (ng/mL). Presumably, they used these standards to create a standard curve. They didn’t show their standard curve in the paper, but it’s typical to not show this relatively basic information. Scientists are just expected to know how to make a standard curve that is appropriate for their experiment.
In Table 2, they report that a a range of 0 to 1.50 ng/mL was detected in maternal blood and 0 to 0.14 was detected in fetal cord blood. They also report that a mean of 0.19 ng/mL was detected for maternal and that a mean of 0.04 ng/mL was detected for fetal. They also report the standard deviations for these means, so we have 0.19 ± 0.30 (-0.11 to 0.49) ng/mL for maternal and 0.04 ± 0.04 (0 to 0.08) ng/mL for fetal. The lowest value on the standard curve was 0.1 ng/mL.
Ideally, your test values will be right in the middle of your standard curve – which is obviously not the case here. For any values outside or at the edges of of the standard curve, we can not be sure that the value of ng/mg is correct. These values may be correct, but they may be false positives. They could have confirmed their results with a Western blot, but for whatever reason, they did not.
After really looking at this paper, I have to wonder how it got through peer review just on the methods issue. The conclusions they make are not supported by the data because the tests they used are faulty. Anyone using this paper as an example of how “Bt is dangerous” or how “GMOs are bad” should go back and read the paper.
Maybe there is is Bt protein in blood and the current tests (ELISA, immunochromatography, immunoblot) just don’t show it. But, even if there is, that doesn’t mean there is a problem. All of the evidence at this time shows that Bt is a safe insecticide (or at least as safe as an insecticide can ever be expected to be). The amount of Bt present in any food that we eat is far, far lower that we would ever expect to cause a problem (experiments with humans showed 1 gram can be eaten and 100 milligrams inhaled for multiple days with no problems).
Could there be something we don’t yet know? Sure. There always could be something else, but the evidence we have is pretty solid. However, one thing is clear. If one is to be concerned about Bt expressed by transgenic crops, then one must also be concerned about Bt residues from organic sprays (which include the Cry1Ab form of Bt) and from soil microorganisms as well, which could be sources of Bt in human blood if it was indeed present.
Note: The paper also examines glyphosate and metabolites but this post is long enough
I may edit this post to add analysis of the glyphosate and metabolites but I do have a lot on my plate.