Gluten Free Watchdog comment on the recent gluten cross contact study “Preparation of Gluten-Free Foods Alongside Gluten-Containing Food May Not Always Be as Risky for Celiac Patients as Diet Guides Suggest”Tricia Thompson
The study, “Preparation of Gluten-Free Foods Alongside Gluten-Containing Food May Not Always Be as Risky for Celiac Patients as Diet Guides Suggest” was recently published online by the journal Gastroenterology. It is available in full at https://www.gastrojournal.org/article/S0016-5085(19)41340-1/pdf.
Personal note: I have been out of the country and a little late to the party in terms of commenting on this study. Thank you to everyone who reached out over the past week for your patience. A comment from Adrian Rogers, Senior Research Scientist at Romer Labs is also included in this post.
According to the authors, the primary aim of the study was “to quantify gluten transfer when GF foods are prepared alongside gluten-containing foods.” To this end, the gluten content of gluten-free foods prepared under three scenarios was assessed, namely:
- Gluten-free pasta cooked in shared water previously used to cook wheat pasta
- Gluten-free bread toasted in the same rolling toaster or pop-up toaster used for wheat bread with visible wheat crumbs present
- The same unwashed knife used to first cut a frosted wheat-based cupcake and then a gluten-free cupcake
- When gluten-free pasta was cooked in water previously used to cook wheat pasta, all 12 samples of pasta contained greater than 20 parts per million of gluten.
- When gluten-free bread was toasted in a shared rolling toaster (which may be used commercially in restaurants, hotels, hospitals, etc.), 16/20 slices tested contained a level of gluten below the limit of quantification of 5 parts per million of gluten and 4 contained between 5 and 10 ppm of gluten.
- When gluten-free bread was toasted in a shared pop-up toaster, all 20 slices tested contained a level of gluten below 5 ppm.
- When gluten-free cupcakes were cut with an unwashed knife previously used to cut wheat-based cupcakes, only 2/30 cupcakes tested above 20 ppm of gluten.
Praise for this study:
There is a tremendous need to increase our understanding of gluten cross contact, including when and how it occurs in both home and commercial kitchens to help inform evidence-based recommendations. The authors should be universally applauded for the study conception and the development of the various scenarios for testing. An accredited lab—Bia Diagnostics—conducted the testing using the fully scientifically validated R5 ELISA Mendez Method. There is no reason to doubt the results of the extractions tested.
Concerning aspects of this study:
- A single extraction was tested from each homogenized sample.
- The kit instructions from the assay manufacturer R-Biopharm explicitly state, “In order to ensure a high analytical performance: Each sample material should be analyzed in duplicates.” When a sample is homogenized the goal is to evenly distribute any gluten present. However, this does not always happen and in fact may not always be possible (try putting three chocolate chips into a batch of sugar cookie dough and evenly distributing them using a blender). Cross contact by nature can be very spotty, including when it comes to crumbs. The standard extraction amount is 0.25 grams. A slice of bread may weigh at least 28 grams. Even if the bread is homogenized in a blender there may be 0.25-gram portions without any wheat crumbs present. Testing at least two extractions provides some insight into whether the sample is homogeneous in terms of gluten.
- Some of the results should give everyone pause and should not necessarily be taken at face value.
- For example, does it make intuitive sense that 16 of 20 slices of gluten-free bread toasted in a rolling toaster (where the bread slice actually lands surface down on wheat crumbs) contain a level of gluten below 5 ppm? Might it be the case that what would be eaten after using such a toaster might differ from what was tested at the lab? Could the wheat crumbs have been dislodged from the surface of the gluten-free bread during collection and placement into plastic bags prior to sending to the lab? Was the lab able to remove all crumbs from the bag? Was gluten distributed evenly throughout the sample? Initial test results should have prompted additional testing of extractions from each bread slice.
- Should certain study conclusions have been made?
- For example, the study states, “Toasting in a shared toaster was not associated with gluten transfer above 20ppm.” Is this really the case? Does testing a single extraction from a slice of bread provide enough data to make this statement? If additional extractions had been tested, would the findings and conclusions have been the same?
- Cross contact must continue to be addressed and assessed.
- One study should not change recommendations.
- Please don’t use a shared toaster oven or pop up toaster without cleaning the grill or using toaster bags. And please do not use a shared rolling toaster when eating away from home.
- Please don’t use shared utensils that haven’t been washed between uses.
- More studies are needed.
- If you conduct or participate in such a study,
please make sure that more than one extraction is tested per homogenized
- If the results don’t make sense, additional testing may be warranted; Alternative reasons for your results should be considered.
- If you conduct or participate in such a study, please make sure that more than one extraction is tested per homogenized sample.
Comment from Adrian Rogers, Senior Research Scientist, Romer Labs: More testing is needed to get a picture. This could be a collaborative project with different researchers repeating the same set of experiments in duplicate or triplicate (Tricia’s note—this means testing 2 or 3 extractions from each homogenized sample). It is all about numbers–a statistician should be consulted to make sure the results are statistically relevant given the nature of the type of contamination.