Why I started Gluten Free Watchdog and how I learned what good testing takes

I still remember where I was when Gluten Free Watchdog became more than just a thought exercise. I was out for a walk, talking with Thom Grace, the CEO of Bia Diagnostics, LLC. We had worked together on the pilot study, “Gluten contamination of grains, seeds, and flours in the United States” and I very much wanted this type of information to be read by more than the scientific community. Thom was so supportive and more importantly he believed in what I was hoping to do—provide state-of-the-art testing data on the gluten content of foods to interested consumers. He was instrumental in helping Gluten Free Watchdog get off the ground. Gluten Free Watchdog was born in 2011.

How I became interested in testing food for gluten

To understand why I started GFWD, you have to go back further, to 2003. My interest in testing food for gluten actually dates back to my interest in oats and whether they could be safely included in a gluten-free diet. In 2003, I had several samples of oats tested for gluten at a top gluten testing lab with the intention of publishing the results. Unfortunately (or fortunately from the standpoint of learning a lesson), the reviewers of the manuscript had concerns (and rightly so) about my method of sampling the containers of oats and the lack of replicates—meaning I opened each container of oats, poured some of the contents into individual resealable plastic bags, coded them and sent them off to the lab. I did not ask the lab to test the samples in duplicate. This describes my bumpy initiation into the world of testing and the beginning of my learning about the importance of sampling, among other things.

Learning from the experts

I quickly found out how much I had to learn. I started reaching out to people who could help. The wonderful, late Don Kasarda was incredibly supportive and so very generous with his time. Most importantly, he never made me feel that I did not belong in the technical world. The team at R-Biopharm was also very helpful, patiently answering my many questions about their new assay—the Ridascreen Gliadin R5 ELISA (subscribers to Gluten Free Watchdog will recognize this assay). With their helpful guidance, I sent a new batch of oats to a lab for testing. Results were published in 2004 in the article, “Gluten Contamination of Commercial Oat Products in the United States.”

There have been many other teachers along the way who continue to have a huge influence on my approach to testing. I was exceedingly fortunate to work with the late, great Enrique Mendez, the developer of the R5 ELISA, on a manuscript on testing assays entitled, “Commercial Assays to Assess Gluten Content of Gluten-Free Foods: Why They Are Not Created Equal.” I met Dr. Mendez when we were both speaking on the same panel at the 2006 International Celiac Disease Symposium in NYC. His willingness to work with me on a manuscript was very generous. In more recent years, Adrian Rogers of Bio-Check (UK) is someone who I’ve relied on to bounce ideas off of and collaborate with on various testing projects. And then of course there is the entire team at Bia Diagnostics who I continue to work with.

The current process at Gluten Free Watchdog for having food tested for gluten

For those who are interested, here’s what the scientific process actually looks like in practice at Gluten Free Watchdog (the short version is that every product gets a testing plan tailored to its ingredients):

• Food requested for testing is purchased and sent to the lab unopened (unless we are testing a product due to an illness report).
• Prior to shipping, the ingredients list is recorded and scrutinized—the presence of some ingredients means that additional protocols must be requested from the lab.
  • Might hydrolyzed gluten be present? If so, the competitive R5 ELISA should be requested along with the sandwich R5 ELISA.
  • Does the product contain spice, chocolate, buckwheat or other ingredients with high tannin or polyphenol content? If so, the inclusion of a milk additive should be requested as part of testing with the sandwich R5 ELISA.
  • Does the product contain yeast extract of unknown origin? If so, the inclusion of fish gelatin should be requested as part of testing with the competitive R5 ELISA.
  • Does the product contain oats? If so, the oat testing protocol should be requested—meaning a larger representative sample is ground/homogenized by the lab and larger weight amounts are tested.
  • Is the product an enzyme? If so, it must be neutralized before testing.
  • And the list goes on.
• Sometimes the lab is asked to run a spike recovery, especially if the product isn’t food. This means a known amount of gluten is added to the sample and tested to see if the gluten can be recovered or if something about the sample is interfering with the assay.
• If test results are widely divergent—for example one test portion tests at 10 ppm gluten and the other tests at 50 ppm of gluten, this likely means the sample contains heterogeneous gluten cross-contact. Additional testing should be done.
• If for whatever reason, results simply don’t make sense, they are not accepted at face value—they are scrutinized.

Thank you to everyone who has helped make the past 15 years possible.