Flawed Study on CBD Mislabeling is Making Headlines: CLR Article November 10, 2017

First published in Cannabis Law Report; 11/10/2017
By Rod Kight, 11/2017

A study conducted in late 2016 and recently published in JAMA found that almost 70% of CBD products sold online contain either higher or lower concentrations of CBD than indicated on the label. The study was led by Marcel Bonn-Miller, of the University of Pennsylvania, and included researchers from the Veterans Affairs San Diego Health Care System, RTI International, Americans for Safe Access, and Johns Hopkins University School of Medicine.

The study was flawed. However, it does bring to light some troubling issues in the CBD market. In the immediate wake of warning letters sent by the FDA to four CBD companies, plus the DEA continuing to assert that CBD is an illegal substance, the fact that a study asserting that two-thirds of CBD products sampled were mislabeled is a troubling development.

There were at least two problems with the study. First of all, there were a number of conflicts of interest. For instance, the authors received fees from the Realm of Caring Foundation (ROCF) and Insys Therapeutics (Insys). ROCF is a nonprofit that funds cannabis research. Insys markets a sublingual formulation of Fentanyl. It also markets Syndros, a synthetic THC product.

Second, the vetting method for choosing products was random and haphazard. Researchers only acquired eighty-four (84) products from thirty-one (31) CBD companies. They identified the products by conducting internet searches for the terms “CBD”, “cannabidiol”, “oil”, “tincture”, and “vape” and purchased products available online that included CBD content on the packaging. The researchers asserted that “all unique CBD extracts” that met their criteria were purchased. This is difficult to believe. At best, it was arbitrary. For example, a general Google search I performed for “CBD” yielded 116,000,000 results with no products appearing on the first page and only two (both of which were ads) appearing on the second. The “Shopping” filter provided more results, but actual CBD products (versus merchandise and unrelated hits) were scattered over hundreds of pages. An Amazon search of “CBD” yielded 8065 results, most of which were products that likely met the researchers’ criteria. The researchers did not indicate what search engine(s) they used, nor how they narrowed down the products (other than excluding the same formulation in other products by the same brand). All of my information was based on only one search of the most prominent of the keywords used by the researchers.

The researchers used high performance liquid chromatography, which is probably better than gas chromatography for this particular study since it does not use heat, which can convert the compounds during the testing process. Therefore the acid forms of the cannabinoids remain intact for measuring.

All in all, and despite its flaws, this study emphasizes the fact that accurate labeling is crucial. If consumers cannot count on accurate and reliable labeling then the industry stands to lose its hard won trust. This will almost certainly invite more government intervention.

The Cannabis Law Report is an international report of cannabis related legal matters.  Rod is featured as a weekly columnist in Cannabis Law Report.  This is the first article in his series.  You can find this article on the Cannabis Law Report blog here: https://cannabislaw.report/flawed-study-on-cbd-mislabeling-is-making-headlines/

2 thoughts on “Flawed Study on CBD Mislabeling is Making Headlines: CLR Article November 10, 2017

  1. Michael Morgenstern MD

    The study is pretty damning in my opinion for several reasons. Overall 84 products from 31 suppliers is actually a lot. Try searching cbd oil on google and sort by sellers. There arent 31 sellers listed. There is around 20. On Amazon, you can sort by brand and only around 15 products are listed (dont know if you can sort by product).

    If they had selected things “randomly” as you suggest that would actually strengthen their research as it would be a statistically representative sample. Regarding the conflicts of interests, they are typical in research and they were disclosed. It is not insignificant, though its hard to see why it would make a difference. Quite frankly it is JAMA which is a big name but there are many reports of this. As you know, FDA sampled and sent warning letters to some distributors for the same problems with labeling.

    1. Rod Kight

      I appreciate your comments, Dr. Morgenstern. You may be correct about the (non)effect of the conflicts of interest and the likely outcome of a more random selection; however, these issues with the study are significant and would likely prevent it from being admitted as evidence in a court of law. That being said, I agree that the reports of mislabeling are troubling and need to be addressed.


Leave a Reply

Your email address will not be published. Required fields are marked *