Monday, January 30, 2012

More Analysis of Anti-GMO Claims

A simple exercise.  Open another tab in your browser and search "GMO fetus".  Go ahead, I'll wait.

Pretty scary stuff, isn't it?!!!  What you see is how our anti-GMO friends are interpreting a paper published in Reproductive Toxicology last year. By the way, according to anti-GMO interests this paper is regarded as one of the best papers in a peer reviewed journal that shows the dangers of transgenic crops.

Today I will now show you that they  1.) are either lying because they hate transgenic technologies and/or the companies that make them, or 2.) they don't know how to critically examine scientific literature.  Either way, they freely overstep the data in the paper, the limitations of the controls, and even the authors' conclusions.  For fun, let's just take a look at the first website.  Read the headline of that article carefully, along with the subhead.  Here it is:

Gaia Health - Information for the welfare of you and your children. 
Too bad they didn't read the report that they use 
for the foundation of their claim...

According to Gaia Health "toxic elements can be found in nearly all pregnant women and children." The report they refer to is by Aris and Leblanc, and it was published in Reproductive Toxicology (impact factor 3.17, pretty good).  I'll send you a copy of the original report if you'd like one. 

The authors show several apparently reasonable conclusions. However, we have to interpret them carefully within the limitations of the experiment and the absence of controls. I'll examine each main conclusion.

Conclusion 1. 
What the anti-GMO people read--  The Cry1Ab protein is present in the blood of all women and fetuses. 

What it really says. That the Cry1Ab protein is detectable in vanishingly small amounts by ELISA, a super sensitive technique, in the majority of subjects (55/69). The protein was found at 40-190 pg/ml (yes, that's picograms, 10-12 grams, that's trillionths of a gram). Their standard curve measured 0.1 ng/ml to 10 ng/ml, and when the mean is considered along with their range and standard deviation, most of the data were likely near or below the lowest point in the standard curve, especially for "fetal cord" where the mean is reported 2.5x below the lowest end of the standard curve. Basically, they are reporting detection, but they are detecting noise. 

Critical evaluation:  The commercial kit is very sensitive. Most of their data were likely at or below the lowest point on their standard curve. The authors did not measure the comparable sera from non-transgenic consuming women (e.g. organic diets), nor did they show negative controls or report background levels.  It is necessary to include these controls as the amounts shown start to approach the limit of detection, and Bt toxin is a natural part of our environment.  The Bt protein is used on organic crops and in gardens. The implication is that the Bt protein in the sera is from transgenics, but they did not include the non-transgenic control. 

Plus, the Bt protein is just a fraction of the protein in a transgenic plant. Back of the envelope math says that you'd have to eat a lot of Bt corn to have any detectable level. I'll follow up on this later in a separate blog.

But let's assume that these levels are accurate.  It is difficult to reconcile how the Bt protein could survive digestion and be localized to blood serum.  Could be!  Is that level of consequence biologically?  Probably not to an insect larvae and certainly not to a primate. I'll follow up on this one too in a future blog. That's what the authors correctly conclude as well, if one takes the time to read the research paper. 

Better controls and interpretations that do not overstep the data-- that's what this data point needs.

Conclusion 2. 
What the anti-GMO people read-  Herbicide levels from GMOs are high in all women and fetuses.

What it really says.  A breakdown product of one kind of herbicide (glufosinate) was identified in their sera using GC/MS.  It was detected in 56 of the 69 women in this study and in 24 of 30 umbilical cord blood samples. 

Sounds scary, until you look a little deeper. 

Critical Evaluation: The compound identified by the authors is 3-MPPA. It is a breakdown product of glufosinate, an herbicide used in conventional farming and at home.  It is sold as Liberty, Basta, and other names.  It is not Roundup (glyphosate), the one used on Roundup Ready transgenics.  Still, its breakdown products are there, so that's scary, right? 

It seems downright alarming until you read that the authors used a method of detection referenced as used by Motojyuku et al, (2008 ) J. Chromatography B, 875; 509-514. The authors in this paper state that you can't measure 3-MPPA using this method! Aris and Leblanc apparently are scoring an artifact. From the text on page 511, continuing to 512 in Motojyuku et al...

 "Although this method could detect 3-MPPA, there existed an interfering peak derived from the endogenous components at the elution times of the 3-MPPA. Thus, 3-MPPA was not validated in this study." 


So this "GMO-related toxin" that is first, not Roundup and then not correctly detected, and may not be present in "nearly all pregnant women and children" as the article claims. To be fair, it may be there.  It may not be there.  This method can't determine if it is or isn't. 

3. What the anti-GMO people read-  Roundup was found in women's blood in this study.

What it really says.  Glyphosate (roundup) was not found in any of the pregnant women and was found in 2 of the non-pregnant subjects.  So that's 2 out of 69.   It was not found in the umbilical cord. 

Critical Evaluation:  Glyphosate (Roundup) was detected in two subjects and not in fetal cords. The authors may be picking up legitimate herbicide residues in these few women with this sensitive technique. It is probably more likely that they may be picking up non-food exposure, like if the non-pregnant women recently sprayed some weeds with Roundup. The range of detection must be huge too, with a substantial standard deviation (based on two data points, hmmm).  It is hard to put any conclusion on this one, especially relating it directly to transgenic crops.  Glufosinate (Liberty, Basta, etc) was detected in 7 women in the study.  

Here's where it gets really fun.  Glyphosate is regarded as "mildly toxic" on its MSDS.  A quick look at that document shows acute oral toxicity (rat) at 4320 mg/kg.  That's over four grams per kilogram of body weight.  Holy cats, I don't know that I could eat that much in one sitting.  According to women's BMI charts for 33 year old white women (as in the study) the mean weight is 64 kg.  This means that they would have to ingest 275 g of Roundup to approach toxicity.  That's about three candy bars of Roundup.  What is reported to be detected is 73 ng/ml.  The average woman in the study is 64,000 ml containing 73,000 ng of Roundup (assuming levels are equal throughout the body, actually must be much less). That's 73 micrograms or about one one-millionth of one of those candy bars! 

So we looked at the first website that popped up on the screen and find that the sensational title and subsequent claims are scientifically unsupported. Even Aris and Leblanc conclude that their report simply illustrates detection and that all levels are significantly below any levels shown to have biological effects. They clearly state that it is a stepping off point for further studies and I agree with that. In my opinion, better controls would make the data much more compelling. It seems weird to detect the Bt protein in the blood, and if the test was done properly (transgenic eating vs. organic diet subjects) and the results were the same, it would have been of greater impact.

To conclude, this article is held up as some of the best evidence in anti-GMO circles as proof that the technology is dangerous. Look at some of the other websites from your original search and how they freely and incorrectly extrapolate the results in this scholarly paper. They use this flawed report to inflame their scientifically illiterate base, and it works.  Pregnant women and fetuses tug at hearts and breed fear.  Just continue down that first Google list and see how the anti-GMO interests are either lying, not critically reading, or both.  

Once again science takes a back seat to the anti-GMO agenda. It is much easier to scare you than to educate you. Especially when it is important to scare you so they can advance their agendas.


Mary said...

Thanks for the details--this is one I've seen flying around a lot.

Karl Haro von Mogel said...

Good breakdown of the non-Bt parts - that is something that I haven't focused on. There are a few other important details that should be mentioned, and are discussed in this post on Biofortified as well as in the comments:
The standard curve they used is invalid - they measured known concentrations of Bt in a completely different medium than the medium they were testing it in. In order to have a standard curve for blood sera, you have to do the curve in blood sera. The buffer they used is not blood sera.
Second, previous research, not cited by Aris and LeBlanc indicated that the particular sandwich ELISA test used in the paper cannot be used to detect levels of Bt that are below a certain amount - and it turns out that ALL of the levels reported were at or below this amount. Here is the paper:

Paul Vijay Kerstin Steink and Heinrich H.D. Meyer (2008) Development and validation of a sensitive enzyme immunoassay for surveillance of Cry1Ab toxin in bovine blood plasma of cows fed Bt-maize (MON810). Analytica Chimica Acta 607:106–113, 2008.doi:10.1016/j.aca.2007.11.022

Kevin Folta said...

Karl, excellent. I didn't know you thought about this one so much. That is grounds for retraction.

I only covered this because someone directed me to it as solid research published in a peer-reviewed journal. I told her I'd give it an evaluation. I was stunned to see how the rest of the anti-science blogosphere treated this and extrapolated the trash!

Thanks for further illumination on this one.