Recently a friend asked me to evaluate an apparently scientifically substantial article that concluded that glyphosate—the active ingredient in Roundup, one of the world’s most widely used artificial pesticides—has caused a dramatic increase in a variety of diseases such as autism, inflammation, immune disorders, gluten intolerance, celiac disease, and cancer.

I’ve been shown various “studies” in the past making similar claims but have always seen quickly that they weren’t based on sound scientific principles.

But the very appearance of this article was different. Here’s a screen shot of its start:

Pretty impressive at first sight. I read more and thought, “This seems more credible than others I’ve seen. Maybe there’s really something to this. Maybe the Cornwall Alliance should, at last, join the chorus of glyphosate critics.”

Over the years I’ve read some 80 books and many thousands of articles, peer reviewed and otherwise, on the science, economics, and engineering of climate change and climate and energy policy, so I’m pretty confident that I can evaluate even fairly technical articles in those fields.

But I’m definitely not an expert on biochemistry, toxicology, or the biology of the human digestive tract, or even biology in general. (How could anyone be an expert in “biology in general,” by the way? Much too broad.)

Had I done some web searching, I’d have discovered quickly that the article and its authors had come under pretty serious criticism. But, frankly, doubting my ability to discern, I decided to take a shortcut.

I asked Cornwall Alliance advisory board member Dr. Daryl Sas, Professor of Biology and Chairman of the Department of Biology at Geneva College in Beaver Falls, PA, to evaluate the study for me.

Four years ago Dr. Sas wrote an excellent article for us, “What’s Incredible? Learning to Read Science Articles with a Critical Eye,” and I was confident he would be able to assess this article much more quickly than I could.

He didn’t disappoint me, but his reply justly chastened me: “Was this a trick question? If you investigate the authors of this article, you will find many sources claiming that it is not credible and is pure pseudoscience, an assessment with which I agree.”

Then he gave a few reasons. These three address the article’s reasoning:

• The numerous graphs—courtesy of Nancy Swanson—showing a correlation between glyphosate use and all kinds of diseases are mere correlations, and no proof of causation. Countless things have increased over the last 30–40 years which also correlate with increased incidence of these diseases! Let me list a few of my own: fluorescent lights, ultrasounds, TV watching, ibuprofen, birth control pills, anti-depressants, anesthetics.
• One of the hallmarks of pseudoscience is the cause-all. According to the authors, glyphosate (actually impaired sulfate) causes cancer, infertility, autism, obesity, Alzheimer’s, and depression, among others. Never believe a cause-all!
• The numerous properties of glyphosate are unbelievable: antibacterial, P450 inhibitor, metal chelator, depletor of amino acids, and herbicide.

Humbled, I realized I should have recognized those errors of reasoning myself, despite the technical, sophisticated way in which the authors wrote.

How can such bad reasoning have been done by credentialed scientists and published in a refereed journal? Aside from the fact that, as I’ve written elsewhere, even in reputable journals peer review is no guarantee of quality, Dr. Sas pointed out (and I later provided the hyperlinks) that these scientists aren’t properly credentialed for the subject, and the journals in which they’ve published on the subject aren’t even reputable:

• The authors are retired computer scientists, not biologists of any kind.
• Their primary source—Nancy Swanson [who prepared the figures and whose B.S. is in physics and mathematics and Ph.D. is in physics]—“writes and promotes GMO horror stories,” according to mrdrscienceteacher.
• This article was published in an obscure Slovakian journal, created by the Slovak Toxicology Society to publish primarily its members’ research, it would appear.
• Even the Gluten Free Club didn’t believe this article!
• [Co-author] Stephanie Seneff illustrates another characteristic of pseudoscience: the lone genius.
• A companion article was published in the journal Entropy. Why would an obscure physics journal publish an article about glyphosate? For the money. According to one of the sources I found, Entropy is a pay-to-play journal which makes money … publishing hardly-reviewed articles. [As of July 14, 2017, the current fee of 1,500 Swiss francs equals US$1,554.08.]

Following up on Dr. Sas’s comments, I found some additional relevant information.

In a review of their Entropy article, of which the Interdisciplinary Toxicology article “is essentially a clone,” Tamar Haspel wrote of the authors, Anthony Samsel and Stephanie Seneff:

Seneff is a Senior Research Scientist in the Computer Science and Artificial Intelligence Lab at MIT. Her advanced degrees are in electrical engineering. She describes herself as having “recently become interested in the effect of drugs and diet on health and nutrition.” Samsel describes himself as an “Independent Scientist and Consultant,” and, for the last 37 years, has run Anthony Samsel Environmental and Public Health Services, which does “Charitable community investigations of industrial polluters.” I think it’s fair to say they probably went into this with a point of view.

In short, they aren’t scientists trained in biochemistry, the specialty relevant to the subject on which they write. That’s not a fatal critique, but it’s significant. To override it, there should be some really solid substance to their arguments.

When I began reading rapidly through Samsel and Seneff’s article, I suspected that it committed the fallacy of confusing correlation with causation, as, for instance, in this graph:

 

Correlations can be spurious, as the website tylervigen.com humorously illustrates by showing strong correlations between such obviously no causally related pairs as

• U.S. spending on science, space, and technology, on the one hand, and suicides by hanging, strangulation, and suffocation;
• Number of people who drowned by falling into a pool and films in in which Nicolas Cage appeared;
• Per capita cheese consumption and number of people who died by becoming tangled in their bedsheets;
• Divorce rate in Maine and per capita consumption of margarine (a spectacular 99.26% correlation!); and
• My favorite, age of Miss America and murders by steam, hot vapors, and hot objects.

 

Yet Samsel and Seneff presented what appeared to me, a non-biologist, a plausible explanation of causal relationship. But Haspel showed that their causal mechanism discussion was all smoke and mirrors:

They say, “We explain the documented effects of glyphosate and its ability to induce disease, and we show that glyphosate is a ‘textbook example’ of exogenous semiotic entropy: the disruption of homeostasis by environmental toxins.” Exogenous semiotic entropy! That sounds serious. Google it, though, and you find that [as of the time Haspel wrote] those three words occur together in only place. This paper. They made it up. At first, I thought the whole thing was one of those jargon-laden academic hoaxes but, alas, it isn’t.

Slog through their argument (and, please, if you take this seriously, read the paper!), and you find it boils down to two things. Glyphosate, they claim, 1) inhibits CYP enzymes, which are active in lots of metabolic processes, and 2) disrupts gut bacteria, which are susceptible to its mechanism (disrupting the shikimate pathway), even though humans are not. Therefore, any condition that involves metabolic processes or gut bacteria must be affected by glyphosate exposure. QED!

… The evidence for these mechanisms, and their impact on human health, is all but nonexistent. The authors base their claim about CYP enzymes on two studies, one of liver cells and one of placental cells, which report endocrine disruptions when those cells are exposed to glyphosate. Neither study is CYP-specific (The effect of pesticides on CYP enzymes, by contrast, has been studied specifically.) As for the gut bacteria, there appears to be no research at all on glyphosate’s effect on them.

Samsel and Seneff didn’t conduct any studies. They don’t seem interested in the levels at which humans are actually exposed to glyphosate. They simply speculated that, if anyone, anywhere, found that glyphosate could do anything in any organism, that thing must also be happening in humans everywhere.

Paul Strode, who blogs as “Mr. Dr. Science Teacher,” wrote a critique of Seneff and Samsel’s Entropy article.

… Samsel and Seneff establish no causative link between glyphosate and human disease. Indeed, the word “associated” is used 43 times in the paper and “association” appears 24 times. Their main hypothesis is something called “exogenous semiotic entropy,” a meaningless concept that exists nowhere in the world but in this paper. The authors present their readers with an explanation—that glyphosate is behind dozens of human health problems—and then drum up all the support they can for it. All without doing any hypothesis testing or experiments themselves. This approach is the opposite of dispassionate and skeptical science.

… The sources of the data upon which the graphs in the paper are based (USDA:NASS; CDC) are not referenced in a way that they can be tracked down, and no citations for the data sources can be found in the Literature Cited section. If we look at the graph from the article that Mother Earth News celebrates, we see that the axes have been adjusted to show a “fit” that is nothing more than a spurious correlation. For this relationship to be even remotely meaningful, the exact same people who have been diagnosed with celiac disease have to be the ones eating the wheat upon which the glyphosate is applied. They’re not. The scale on the primary Y-axis has also been adjusted to make the correlation look nearly perfect.

Dr. Anna Kaplan, a physician who specializes in food intolerances, writing for the Gluten Free Club, pointed out additional errors in Seneff and Samsel’s reasoning:

There are so many faulty assumptions and outright errors in this paper that it is unnecessary to explain every one of them. Just laying out some of the facts is enough. With what you know about celiac disease or can discover by doing just a little research, you will see mistakes in the very first sentence of what is called the “abstract” or summary of the article.

The authors state, “Celiac disease, and, more generally, gluten intolerance, is a growing problem worldwide, but especially in North America and Europe, where an estimated 5% of the population now suffers from it.” They then go on to treat these celiac disease and gluten intolerance as if they are identical. They go on to use their “evidence” to explain how glyphosate causes the serious conditions associated with celiac disease.

Gluten sensitivity is not celiac disease. It is poorly understood, and occurs when people have symptoms when they eat gluten that go away when they do not. These individuals test negative for celiac disease and do not develop the complications or problems that people with celiac disease often do. It is not clear how many people have gluten sensitivity since there is no universally accepted way of diagnosing it at this time. It is also unclear where the number 5% came from, but it is not the percentage of people with celiac disease. In North America and Europe, the prevalence or rates of celiac disease are about 1%.

She then pointed out that in their articles Seneff and Samsel have blamed glyphosate for increasing rates of celiac disease, thyroid cancer, kidney failure, obesity, diabetes, heart disease, autism, Alzheimer’s disease, Parkinson’s disease, infertility, depression, and other cancers.

“All of this is based on graphs and charts showing that the increase in glyphosate parallels the increasing incidence of these diseases,” she continued.

This is not proof. These things may be associated in time, but one does not cause the other. … Just by looking at the consumption of various foods in the United States, the amounts of olive oil, poultry, palm oil, and soy bean oil have all increased in the same way as the incidence of celiac disease. It would be as correct to say that olive oil causes celiac disease as that glyphosate does.

But Seneff and Samsel do cite a study showing that fish exposed to high levels of glyphosate can suffer damage to their intestinal tracts. Does that save them from the charge of confusing correlation with causation?

No, Kaplan explained: “It almost seems unnecessary to note that fish are so different from humans that this kind of comparison is useless. Giving them our air to breathe kills them, and putting us under water to breathe kills us.”

She continued:

No tests have been done to confirm the herbicide theory.

All the authors give is lists of chemical reactions and long names of enzymes in the body that can be damaged by glyphosate in theory or in a lab. While the descriptions of the way glyphosate interrupts certain chemical pathways in bacteria seem very scientific, many of these have not been proven. Those that have been studied have been studied in bacteria, not even in any animals.

While testing this theory might be difficult, something could be done to try and see if there is any truth in it. One could draw blood levels of glyphosate from people newly diagnosed with celiac disease and compare these to similar people without celiac disease, for example. Or, people with untreated celiac disease who want to test this could be recruited and treated with either a gluten-free diet or a glyphosate-free diet to see how they do, of course without letting anyone know who got which diet (double blind study).

No one will do this study because the basic assumptions are not sound. There are strict rules governing human research. In addition to understanding the basic science, animal studies come first. None of the theories in the article have been tested on animals.

Kenrick Vezina, Senior Editor of the Genetic Literacy Project, also critiqued Seneff and Samsel’s Entropy article:

… from the first page, the report is riddled with danger signs. The two people listed as authors are an “independent scientist and consultant” and a member of MIT’s computer science and artificial intelligence laboratory—neither is a geneticist or biochemist or biologist or botanist or nutritionist or anything else that might lend them expertise in the matter of glyphosate’s effects on human health.

By the second page, despite [their] 280-odd citations, the authors have made a major claim without citing any evidence: “Research indicates that the new RNA and DNA present in genetically engineered plants … have not yet fully understood biological effects.” Given the impressive literature that exists documenting the safety of genetically engineered plants, it seems rather glib to hinge your entire argument on a premise without providing a single piece of supporting evidence. In fact, the biological effects of new RNA and DNA in genetically engineered plants has been widely studies in dozens of articles. With a few controversial exceptions, the overwhelming consensus of mainstream geneticists and major international science bodies is that the biological impacts of genetically modified crops are benign and GE foods are safe for human consumption.

The authors ignore 98% of the studies to focus on a single discredited one. The authors cite the now-infamous work of French scientist Gilles-Eric Seralini, whose deeply flawed study linking Roundup to cancer in rats has been roundly rejected by the scientific community. Not good.

The paper doesn’t get any better. The authors jump from unsupported conclusion to unsupported conclusion and from one disorder to another, daisy-chaining the supposed deleterious effects of glyphosate on our gut bacteria as a facile explanation for colitis and autism and obesity. At no point do they perform any original experiments or provide any direct evidence of the links between the various rungs in their daisy-chain of doom—it’s all couched in far-fetched plausibility-as-evidence.

In conclusion, Samsel and Seneff present no logically or scientifically valid evidence in either of their articles that glyphosate (Roundup) causes any disease whatever. The National Pesticide Information Center offers much more reliable information on glyphosate.

What first fooled me into taking Samsel and Seneff’s article seriously? I’m ashamed to admit it: its appearance. Simple as that. It’s only slightly comforting to know that I’m not alone. As Vezina pointed out:

At a quick glance, this paper gives the impression of being a proper piece of scientific writing. It’s in what sounds like a reputable journal, … and formatted like other academic papers, with an interminable list of citations. You might laugh at the idea of being tricked by formatting, but the human brain is particularly good at recognizing patterns, and conscious or not we’re influenced by how “official” a document looks. Why else would a resume in comic sans evoke laughter?

The Apostle Paul’s admonition holds true: “Test all things, hold fast what is good” (1 Thessalonians 5:21). Christians should be cautious about the risk of becoming purveyors of false fears.

For help discerning sound from unsound, scientific from “scientific,” articles, read Dr. Sas’s “What’s Incredible? Learning to Read Science Articles with a Critical Eye.”