Daniel Bier, the editor of this blog, brought to my attention a news story by Reuters regarding a report published on the active ingredient in Roundup® called glyphosate. The conclusions and the mode of publication immediately raised some skeptical red-flags. So I started reading the study, and all I can do is shake my head at how poor the “science” really is in this paper.
The first red-flag comes right away in the article’s title: Glyphosate’s Suppression of Cytochrome P450 Enzymes and Amino Acid Biosynthesis by the Gut Microbiome: Pathways to Modern Diseases. The terms “modern disease” and “gut bacteria/microbes/flora/etc” have been used and abused by pseudoscientific practitioners of fad diets, anti-vaccination people, anti-genetically modified food people, and other groups which employ the naturalistic fallacy. The term “modern disease” is a loaded term, with the claim that all diseases and conditions (cancer, autism, MS, Parkinson’s, etc) are caused by our modern lifestyle and medicine. It is also popular technique to tie all diseases to problems in the intestine. While there is a kernel of truth in a few such claims, too often the terms get pseudoscientific treatment. Now my skeptical senses were tingling.
The next set of red-flags come from the abstract. I will post the entire thing here. I think those familiar with skepticism can find several overly bold statements here. I will highlight a few below.
Glyphosate, the active ingredient in Roundup®, is the most popular herbicide used worldwide. The industry asserts it is minimally toxic to humans, but here we argue otherwise. Residues are found in the main foods of the Western diet, comprised primarily of sugar, corn, soy and wheat. Glyphosate’s inhibition of cytochrome P450 (CYP) enzymes is an overlooked component of its toxicity to mammals. CYP enzymes play crucial roles in biology, one of which is to detoxify xenobiotics. Thus, glyphosate enhances the damaging effects of other food borne chemical residues and environmental toxins.
Negative impact on the body is insidious and manifests slowly over time as inflammation damages cellular systems throughout the body. Here, we show how interference with CYP enzymes acts synergistically with disruption of the biosynthesis of aromatic amino acids by gut bacteria, as well as impairment in serum sulfate transport.
Consequences are most of the diseases and conditions associated with a Western diet, which include gastrointestinal disorders, obesity, diabetes, heart disease, depression, autism, infertility, cancer and Alzheimer’s disease. We explain the documented effects of glyphosate and its ability to induce disease, and we show that glyphosate is the “textbook example” of exogenous semiotic entropy: the disruption of homeostasis by environmental toxins.
The last two sentences stand out the most. The human body is a very complex system. The implication that any one chemical is the cause of every health disorder is a sign of either intellectual dishonesty or cognitive dissonance. A similar tactic is used in opposite way for cannabinoids. Proponents of marijuana legalization claim cannabis will “cure” a host of diseases it could theoretically effect, but the research indicates it is much more complex, perhaps treating certain conditions while exacerbating others. While it is technically possible glyphosate plays a role in all of the diseases listed, it is highly unlikely. To implicate glyphosate as the number one example of an “environmental toxin” indicates a misunderstanding of the complexity of biology.
Glyphosate and Cancer
One of the first claims in the paper is as follows:
Thus, while short-term studies in rodents have shown no apparent toxicity, studies involving life-long exposure in rodents have demonstrated liver and kidney dysfunction and a greatly increased risk of cancer, with shortened lifespan.
The cited study on the long-term effects glyphosate is rife with inadequacies. If you look at the link, several follow-ups are posted regarding these methodological problems. It is a failure of review not to acknowledge these follow-ups, and a failure not to point out that one study with rats bred to have a predisposition to tumors is not enough to make a strong conclusion. As David Gorski at Science-Based Medicine points out:
The investigators measured numerous parameters in each group, some of them at multiple different time points. An experiment with this many groups and this many parameters measured this many times is virtually guaranteed to generate multiple “positive” results. How did they control for all these multiple comparisons? I’ve read the study a few times now, and I still can’t figure it out. An experiment with this many groups in which this many parameters are measured is guaranteed to produce “statistically significant” differences in a number of variables by random chance alone. Heck, in Figure 5, I counted 47 different parameters measured, and in some tables thirteen different parameters recorded curiously as percentage changes. Even worse, for the mortality data (arguably the most critical data), no confidence intervals are reported, and there appears to be no discussion of how the mortality data were analyzed…
I am skeptical that rats in a control group bred to develop tumors 72% of the time are the best method for finding a substance that’s a strong contributor to cancer.
Glyphosate and Gut Bacteria
Their next conclusion is as follows:
The incidence of inflammatory bowel diseases such as juvenile onset Crohn’s disease has increased substantially in the last decade in Western Europe and the United States. It is reasonable to suspect that glyphosate’s impact on gut bacteria may be contributing to these diseases and conditions.
This is a common correlation/causation error, which is an error demonstrated in this viral post from a couple months back:
While research on gut bacteria and the role it plays on human biology, and there is intriguing research on gut flora and fecal transplants in treating specific digestive diseases, it hasn’t shown effects outside of that to this point. As Mark Crislip concludes in his article on fecal transplants:
The practice of medicine is always in flux* and off the wall ideas today are tomorrows standard of care. For fecal transplant and some colonic diseases, it is an intervention that, outside of C. difficile, is still unproven, although a promising idea. For diseases outside the colon, biologic plausibility makes stool transplant unlikely to have any benefit with real potential downsides. Stool transplants are unlikely to be of widespread to benefit, but when all you have to offer is crap, everything is a toilet.
Glyphosate and Autism
The authors try connecting glyphosate to autism, but do so through the gastrointestinal hypothesis. Their statement is as follows:
It is now well established that autism spectrum disorder (ASD) is associated with dysbiosis in the gut, and, indeed, this is viewed by many as an important contributor to ASD.
The study cited is by Karoly Horvath, who has pushed the connection between “leaky gut” and autism, similar to Andrew Wakefield. Karoly Horvath is listed among “Autism’s False Prophets” in Paul Offit’s book of the same title. The leaky gut hypothesis is centered around the idea that something triggers the intestine to become inflamed (vaccines, foods, “modern” foods, etc) and allows “macronutrients” in the blood and damages the brain. While it is true that autistic kids have a higher incidence of gastrointestinal issues, it has been hard to find a single cause. There are errors in reporting, since many studies relied on parental reporting of issues. There are other possible causes, such as one recent study that showed those with autism that reported gastrointestinal issues also reported higher rates of anxiety and sensory over-responsivity. The possible pathways of inflammation cited by Wakefield and Horvath have all since been discredited in follow-up studies. SInce the original study by Horvath was published in 2003, I would have to assume Anthony Samsel and Stephanie Seneff willfully ignored the followup in order to prove their point.
As a side note, it is important to note the genetic component of autism. Families with an autistic child have an increased chance of another child having ASD. Other brain disorders share a genetic link with autism. A small study showed the brains of autistic boys were dramatically different in their structure than a control group. Because of a genetic link, it is possible the intestines of autistic kids are more susceptible to inflammation, but it is much less likely the inflammation is what causes autism.
Glyphosate and Obesity
The authors of the paper also make an auspicious connection to obesity. They state:
The obesity epidemic began in the United States in 1975, simultaneous with the introduction of glyphosate into the food chain, and it has steadily escalated in step with increased usage of glyphosate in agriculture.
What’s interesting is the studies they link to all have to do with dietary changes, and not with glyphosate. The studies correlate sugar consumption, portion size, and similar issues as leading to obesity, not glyphosate. It is also of note that obesity started rising the year glyphosate was introduced, yet elsewhere the authors claim it is chronic exposure over the long-term that causes problems in the body. If it indeed is long-term exposure, then there should be a lag in the rise of obesity, not an immediate correlation.
Other Opposing Views to the Claims of the Paper
Kevin Kloor writes about the poor reporting done on this study (and poor science reporting as a whole). For example, Reuters called the paper a “study,” which it clearly is not – it is a review – but really should be treated more as an editorial commentary since the literature review is very biased and incomplete. Hank Campbell at Science 2.0 talks a little about this paper and the dangers of an open-access journal such as Entropy. Even though it carries a label of “peer-review,” it is merely a small editorial board looking for certain gross errors, but not truly peers in the field looking at the science itself. Hank Campbell does offer hope that some science reporters did actually look at the study, and showed it to be nonsense…but the fact the Reuters article still exists without so much as an update is concerning.
Monsanto makes a few valid points within its response on its blog. Similar to Hank Campbell’s response, they state:
Publication in a physics journal would not have subjected the paper to appropriate peer review by experts in relevant fields of biology and medicine.
A few of the other points in the blog, such as the affiliations of the authors, are not proof by themselves that the study is flawed. But taken as a whole, along with the content (and missing content) of the paper, it is apparent the authors of the review were biased in their conclusions. Glyphosate has been the source of controversy for quite some time. While concern over Monsanto’s influence on agriculture is a valid discussion, it doesn’t invalidate the science behind it (what I nickname the Capitalistic Fallacy).
Dave Mance III takes a bit more of a common sense, middle-ground approach to glyphosate in an article for Northern Woodlands magazine. While pointing out that the precautionary principle would dictate finding alternatives to any herbicide, there are certain good uses for them. Mr. Mance also looks at how science is misreported, because good scientists are honest in their assessment that there is no 100% certainty. The article further points out the importance of looking at the source of studies and their funding sources. When looking at independently funded studies (such as by universities), he found:
Researchers at Cornell, among other major universities, analyzed toxicology tests in 1996 and concluded that glyphosate is “practically nontoxic” by ingestion, and it’s unlikely that the chemical would produce reproductive, teratogenic (birth defects), mutagenic, or carcinogenic effects in humans (read the results yourself). But because science is never 100 percent certain about anything, you can see where phrases like “practically nontoxic” leave an opening for attack. The web is awash with antiglyphosate stories with headlines like: 50 percent of rats given this died – Why is it on your dinner plate? And, indeed, if you fed rats more than 5,600 mg/kg of the chemical they would probably die. (To put this number into perspective, a proportional dose for a 150-pound human would be 840 g, almost two pounds; table salt would kill you at 350 g.) But by making an extreme example the rule, and overlooking the fact that rats fed 400 mg/kg a day for most of their lives showed no adverse effects, such stories give readers a skewed picture of the real-world health risks.
This is a nice summary of how scientific conclusions can be misinterpreted, and how dose is an important factor in many things biological and chemical.
Conclusion
The bold conclusions reached by the review are not supported by the evidence. The authors did include many citations, but did not look at followup to many of the citations or willfully ignored them. There are many leaps of correlation to causation that do not fit their own hypothesis, but they force fit them anyway. The connections to a popular topic among pseudoscientists, gut bacteria, are very shaky at best. The review is easily dismissed, and already has been by science writers with expertise and good science sense.
Reuters would be well served to update their article to point out these errors. In fact, this should be the policy of any science writer – to welcome discussion and a willingness to provide updates as new information comes to light. While I feel I do sufficient research in my own writing, if solid evidence comes to light which is in opposition to my conclusion, I am comfortable with changing any of my conclusions. I believe that is the mark of a good scientist, a good science writer, and a good skeptic.