Using Nima to test mustard for gluten contaminationTricia Thompson
The video is embedded below. First a recap of Nima testing to date …
Sampling methodology for Nima: Gluten Free Watchdog has been assessing Nima’s sampling methodology. Users are advised to place a small sample of food inside a disposable capsule. A small sample is defined as “pea sized.”
In the opinion of Gluten Free Watchdog, this sampling method is not well suited for testing packaged foods that may have heterogeneous (unevenly distributed) gluten contamination or restaurant meals with spotty cross contact from gluten-containing foods. For an explanation please see the following two videos
How well does the Nima detect gluten? Putting sampling method aside, does Nima detect gluten if it is actually present in the sample of food placed in the testing capsule?
Testing barley: Nima was able to detect gluten when Gluten Free Watchdog tested single grains of barley. These grains came from bags of lentils and were identified as barley by a grain farmer via photographs and physical characteristics. The Nima test results for the two grains (tested separately) were “low gluten.”
The “low gluten” results were a bit surprising. We thought this reading might have something to do with the low weight of the sample so we tested a similar amount of wheat-based breadcrumbs. The result for the breadcrumbs was “high gluten.”
These test results taken as a whole make us very curious about the cross-reactivity of the antibody used by Nima with the various gluten-containing grain. Might this antibody have a “low” cross-reactivity to barley?
A video describing the testing of barley grain is available at https://www.glutenfreewatchdog.org/news/mystery-grain-in-edison-grainery-lentils-tests-low-gluten-using-nima/
How does the Nima team define low gluten? The grain test results also caused us to question how the term “low gluten” is defined by the Nima team. The Nima website provides the following information:
Smiley Face: 0 to 20 ppm gluten
Low Gluten: 5 ppm to 15,000 ppm gluten
High Gluten: At least 100 ppm gluten
What is the explanation for this large range and overlap?
- For the purposes of gluten-free labeling, gluten levels are stated in the unit ppm. This stands for parts per million. It is the same as milligrams per kilogram. Why? Because 1 kilogram of food equals 1 million milligrams of food. So 1 ppm of gluten means 1 milligram of gluten in 1 kilogram of food.
- When testing food for gluten, a lab tests a precise weighed amount of food for gluten. One potential flaw with the Nima (in the opinion of Gluten Free Watchdog) is that the sample placed into the testing capsule is not weighed. Consumers instead are told to add a pea-size amount.
- If the Nima does not know how much a sample weighs then how can it know the part per million value of the sample? Simple answer: It can’t. The Nima teams states this on their website writing, “The chemistry is based on weight of the sample, and currently Nima cannot measure your sample weight and therefore statistically provides a range.”
- Based on the numbers provided by the Nima team, a sample with the result “low gluten” may contain less gluten than a sample with the result “smiley face.” In other words a product labeled gluten-free or certified gluten-free and actually containing less than 20 ppm of gluten or less than 10 ppm of gluten, respectively may nonetheless receive a “low gluten” result.
Now onto mustard (please watch the video and read the text discussion below)
The Nima accurately tested the mustard without gluten ingredients to be “gluten-free.” It also accurately detected gluten in the mustard containing wheat flour. The issue that came to light when testing mustard pertains to sample dilution.
The Nima website states that certain samples should be diluted with water before testing https://nimasensor.com/nima-testing/. These samples include dry or powdery foods, intensely colored foods, and vitamins and pills. Instructions for testing these sample types state to “use only a small amount and dilute with water.” Small amount is not defined and the amount of water is not defined. Keep in mind that when you add water to a sample you are decreasing the ppm level of gluten in that sample. This is likely why straight mustard tested high gluten and diluted mustard tested low gluten.
Test results were shared with testing expert Adrian Rogers, Senior Research Scientist at Romer Labs. He provided the following statement for publication:
“My main concern from the mustard testing would be that you have to dilute some samples to make them suitable for use in the sensor, but there is no way to take this dilution into consideration with the final result. If in the lab we dilute a sample we know this would have the knock on effect of increasing the LOD of the test. So if the mustard was low gluten due to cross contamination this may be missed when you dilute the sample out with water. (You are eating the mustard not the watered down mustard!).”
Gluten Free Watchdog is in the process of purchasing foods to have tested using the Nima, R5 ELISA and possibly R-biopharm’s lateral flow device. This testing will be done by Bia Diagnostics.
The devil is in the details! Anybody who tried to do what you are doing, Tricia, without your precision and attention to detail, would make a mess of the whole affair. But you do it so well! Can’t wait to hear the final verdict on the NIMA vs. ELISA testing.
Wow, Kevin. Thank you so much.
Very helpful information about Nima, but video also caused alarm.
We use Coleman’s Mustard all the time.
While the jar you showed briefly in the video, was not the same as the tin that we use, the picture above the video showed the tin.
Checking the ingredients on my tin, it said mustard flour. was a comma missing and it should indicate mustard and flour.
Fortunately, I think the flour applies to the jar, although you do mention having to dilute the powder with water.
I think I am not alone in my concern. Is the powdered Coleman’s Mustard safe?
Hi Doris, Please click on the photo links to see the specific Colman’s mustard that was tested. It is the jar not the tin. Looking at the dry mustard powder online the only ingredient listed is mustard flour. The tin of dry mustard does not contain any gluten ingredients.
Thank you Tricia.
Thank you for your research into the pros and cons of Nima. I subscribed to your site hoping I would find this exact information. I have been using Nima for the last several weeks (my six year-old son has celiac). One of my main concerns is understanding the significance of the low gluten result. I have seen the low gluten result most recently when testing ground beef in a Mexican restaurant we frequent. We bring our own GF tortilla chips and make nachos. I was surprised that the meat tested positive for low gluten because it just contains ground beef and spices. I tested beef samples both from the serving area (possibly affected by cross contamination) and from the sautée pan in the kitchen (which was a clean surface only used for meat, with dedicated utensils). Both tested low gluten. I concluded that the source of low gluten may be the spices, many of which I know might be contaminated. However, if that is the case, I would expect the amount of gluten to be way below 20 ppm and perhaps unlikely to make my son ill. Unfortunately, the low gluten result provided by Nima offers no useful guidance because it can range from 5ppm to much higher as you know. I’m at a loss how to use this result. I hope your future research can allow the manufacturers and users of Nima to better understand the benefits and limitations of testing. Sorry for the long comment, but I hope your research and reports will enhance the utility of this device. It has such great potential, especially for naturally somewhat homogenized foods like soups, stews, sautéed foods. Thanks for your work.
Hi Daniel, If you have a good relationship with this restaurant you might want to consider asking them if the hamburger is 100% beef with no fillers. It also would be helpful to look at all the spices used in this dish to make sure they contain only spices. The result you are seeing may be accurate (meaning a gluten ppm level between 5 and 15,000) or this result could be a false positive. Bia Diagnostics, the lab used by Gluten Free Watchdog started assessing the Nima this week. As soon as I know more, the information will be shared with subscribers to Gluten Free Watchdog.
I have been using the Nima since January. Since my friend is a chemist we have had a similar discussion.
The overlap in the range on the website between low gluten and a smiley face is because of a firmware update. When the Nima was shipped out it was only able to show a yes or no response. Smiley face meant safe, wheat sheaf meant not safe.
The firmware update can only be obtained using the IOS app. Once you have the update, you get Smiley Face- no gluten; low gluten meaning 5-15ppm; and the wheat sheaf meaning over 20ppm. I believe they should make this more clear on the website.
I also think each tester should come with a scoop to standardize the amount. For now, any positive test means I don’t eat it. It’s better to miss out on a food than to ingest gluten. It’s a lot easier than depending on people who don’t know about gluten to advocate properly.
Hi Elaine, This is not the information provided by Nima on their website or in a graphic released yesterday. According to the website https://support.nimasensor.com/customer/en/portal/articles/2581724-what%E2%80%99s-the-difference-between-gluten-levels- Smiley face means no gluten detected; low gluten means at least 5 ppm detected up to as much as 15,000; and high gluten means at least 100 ppm. The graphic released yesterday pic.twitter.com/vmTTm9inr1 “corrects” the low gluten reading to mean 20 to 15,000 ppm with a statement that in rare occasions the Nima detects gluten below 20 ppm. Yes, a scoop should be provided along with mini pipets to help standardize the amount tested.
Hi Tricia, In referring to point #4 in your explanation about the gap range, conversely, does the logic hold true that a “low gluten” result could actually indicate a higher gluten level in the sample tested, than a “high gluten” result in another sample test? In using the Nima, what triggers a “low gluten” test result of 10,000 versus a “high gluten” test result of 6,250? You raised a similar question when you tested a crushed grain of barley found in lentils. Under what other circumstances are the low/high testing results misleading and/or inaccurate, or just plain not helpful? The good news is that GFWD will find out! I hope the manufacturers of the LFDs are paying attention. Thank you Tricia!
It appears so, Bonnie (that a low gluten result could indicate a higher gluten level than a high gluten result). As for your other questions, I will get back to you after we have done more testing…
People should also be aware that mustard flour usually means plain ground mustard seeds not mustard and wheat flour. Same as ground almonds in Europe are called almond flour in US. Word useage standardization would help.