Comments We Submitted Regarding Proposed FDA Labeling for Gluten-Free Foods
By: Peter Olins, PhD, and Gillian Olins, PhD, on November 3, 2011.
The U.S. Food & Drug Administration (FDA) has developed a proposed regulation for the labeling of gluten-free foods, based on a threshold of 20 ppm (parts per million), and opened up the topic for public comment on August 3, 2011. We wrote a blog about this on August 18, 2011, entitled “Should FDA Set 20 ppm Limit for Gluten-Free Food for Celiacs?”. Based on the feedback we received on this blog, we refined our thinking on this issue and submitted our comments to the FDA.
Drawing on the feedback we received, we refined our position on this topic, and submitted our letter to the FDA on October 2, 2011. http://www.regulations.gov/#!documentDetail;D=FDA-2005-N-0404-0461 which is shown in full below.
The main theme of our argument is that there is a remarkable lack of definitive clinical data on the safety threshold for individuals who are most sensitive gluten, and therefore a “gluten-free” regulation needs to err on the side of caution.
The FDA open comment period ended on October 3, 2011, and so far the texts of approximately 400 new comments on the proposed FDA regulation have been released (screening and release of these comments is ongoing). The contents of FDA Docket # FDA-2005-N-0404 can be reviewed at: http://www.regulations.gov/#!docketDetail;rpp=100;so=DESC;sb=postedDate;po=0;D=FDA-2005-N-0404
While we have not read all of the submissions, a few general themes have emerged:
- The majority of comments were simple statements urging the FDA to reach closure and issue a final regulation.
- Many people supported the proposed 20 ppm regulation (sometimes with slight refinements).
- A few people (including the Celiac Sprue Association), urged for a more stringent criterion for a designation of “gluten-free”.
- Other comments related to the specific details requested by the FDA, and to business and social implications.
Letter submitted to the FDA by Peter & Gillian Olins:
Regarding: “Food Labeling; Gluten-Free Labeling of Foods; Reopening of the Comment Period”. Docket No. FDA-2005-N-0404-0135
Thank you for opening this proposed regulation to public comment. After reviewing the Safety Assessment document we sympathize with the FDA and suspect that the delay in finalizing this regulation was in part due to the surprising scarcity of definitive, controlled studies to address the safety of dietary gluten, particularly for those people who have the highest degree of gluten-sensitivity.
On August 18, 2011, we presented an analysis of some of the issues regarding gluten-free labeling, together with supporting literature (Ref. 1). Below we offer our recommendations to the FDA, plus a rationale which incorporates information and feedback that we received from our earlier analysis.
Ref. 1: http://ultimateglutenfree.com/2011/08/fda-20-ppm-regulation-gluten-free-food-celiac-disease/
Based on the considerations discussed below, we respectfully offer the following recommended modification to the labeling process, which we believe will be of greatest benefit to the largest number of celiacs, without being unduly onerous to manufacturers.
- “Gluten-free” should be defined based on the limits of measurement obtainable with easy to use and reliable assays. Currently, it appears that a level of 5 ppm gluten can be reliably measured, and we see no benefit in setting a higher threshold.
- In light of the substantial uncertainty regarding the actual safe level of gluten consumption for those who are most sensitive, we recommend that all “gluten-free” labeling be accompanied by a colored text box warning, to the effect of: “This product may contain traces of gluten, and its suitability for people with a high degree of gluten sensitivity has not been determined”. This approach would provide the best “truth-in-labeling”,and would satisfy the needs of the consumer/physician/dietitian in selecting appropriate foods.
- A label such as 20 ppm gluten is really a measure of concentration; however, it seems most likely that the the toxicity of gluten is related to total daily quantity ingested, not concentration. Consumers are already very comfortable with calculating calories or daily values, so we recommend that the labeling be adjusted in some way to reflect amount per serving. This would have the added benefit of allowing more products to be labeled “gluten-free”, by virtue of their small serving size (compare, for example, the amount of gluten in a serving of hamburger bun versus the ketchup used as condiment).
CONCERNS REGARDING USE OF AN ANALYTIC METHODS-BASED APPROACH FOR GLUTEN-FREE LABELING
- While we understand that the scarcity of relevant safety data has led the FDA to choose a methods-based labeling rather than a safety-based regulation, we strongly believe that a simple “gluten-free” designation will be interpreted by the typical consumer as an implied declaration of safety, i.e. zero gluten. We therefore cannot support a simple measurement-limit criterion of 20 ppm, especially since there is ample evidence that the current detection limit is close to the level of gluten that can cause a cellular response in some celiac patients (see below).
- If the FDA nevertheless proceeds with its planned approach, we recommend that the most sensitive available assay be adopted, within the contraints of feasibility, reliability and cost. We do not understand the FDA’s choice of a 20 ppm gluten cutoff when at least one commercially available assay is claimed to have a quantitation limit of 5 ppm.
- The use of an ELISA directed to a short peptide, which recognizes multiple epitopes within the family of gluten proteins, is really a surrogate for an actual protein measurement. Indeed, wheat gluten is not a single protein, but instead a family of proteins, which are further diversified because of the complex genetics of different varieties of wheat. This is further complicated by the variable interaction of antibodies with epitopes present in non-wheat grains. While we do not wish to be pedantic, and appreciate that the consumer and industry would prefer a simple ppm number, it should be noted that the biological activity of a given (ppm) concentration of wheat from a given strain (or from a non-wheat grain) has not been validated.
GLUTEN-FREE FOODS ARE USED THERAPEUTICALLY, NOT JUST AS A DIETARY CHOICE, SO A SAFETY-BASED APPROACH SHOULD BE USED, IF POSSIBLE
It is obvious that the safety of gluten-free products is fundamentally different from the familiar “fat-free” designation: clearly “fat-free” does not imply a lack of detectable fat, but is simply a piece of information that allows the consumer to make dietary choices which may affect their overall health. In contrast, gluten has a safety profile that spans several orders of magnitude, depending on the individual, ranging from complete tolerance to extreme sensitivity. While a “gluten-free” designation may be informative to the large number of non-celiacs wishing to make dietary choices, the “gluten-free” designation is critical for the safety of individual celiac patients who are highly sensitive. In contrast to “fat-free” foods, in which the inevitable presence of residual traces of fat in no way affects their utility, gluten-free foods may be thought of almost as “drugs”, in that they are consumed in order to treat a disease and reverse the mucosal damage induced by celiac disease. Using this “drug” analogy, gluten-free foods are efficaceous only when a contaminant (gluten) is minimized. In effect, GF foods are used as a chronic, life-long treatment. Active celiac disease can take years to develop; the damaged intestine often heals slowly over a period years; healing is not always complete; and incidental exposure to small amounts of gluten can cause a relapse.
IMPORTANCE OF A LARGE MARGIN OF SAFETY
Given the limited available clinical data on the threshold of toxicity of gluten, we believe that the current goal should be to achieve a wide margin of safety. (It is unclear to us why the FDA has chosen to disregard the 1 ppm gluten recommendation generated by its own safety assessment team).
There are multiple sources of variability and uncertainty in defining a threshold of toxicity of ingested gluten:
- Significant inter-patient differences can be seen in the magnitude of response to gluten.
- Individual patients have different immunological reactions to different peptide sequences present in the gluten protein mixture.
- In addition to the requirement for a specific set of HLA genes, there are several other genes and alleles that affect the sensitivity to gluten.
- There are wide differences in the prevalence of celiac disease between national/ethnic groups, suggesting that it may be hard to extrapolate from clinical data obtained in one country to the genetically diverse population of the US.
- Available clinical data, by necessity, involves relatively short time-frames of administration; this may not extrapolate to a lifelong exposure.
- People who are most sensitive are also those who are least likely to be represented in a clinical trial, partly because they may not meet the enrollment criteria for an effective study, and partly because of ethical considerations. (We are not aware of any data on the number of highly-sensitive patients, as has been requested by the FDA. Anecdotally, we have received a number of responses to our original article (Ref. 1) from individuals who believe themselves to be in the highly-sensitive category. Unfortunately, we cannot share these reports, since they are obviously not accompanied by any accurate measure of actual gluten consumption.)
- While we do not claim to be professional toxicologists, we are particularly interested in the interpretation of the data from one of the few well-controlled studies, which has been widely cited as providing critical information on gluten safety (Ref. 2). For a number of reasons, we caution against over-interpretation of the data from this excellent study.
- Patients with the most severe mucosal damage were excluded prior to the trial, and 4 out of 49 patients were excluded during clinical trial because they still exhibited severe enteropathy while ostensibly on a gluten-free diet. While the need to select a narrow cohort of patients is understandable from a statistical perspective, it also means that it is hard to extrapolate the safety results of these studies to the patients who may be most likely to be sensitive to gluten traces to begin with.
- In one arm of the study, the 13 patients consuming an extra 50 mg gluten per day exhibited a statistically significant degree of mucosal damage (but without overt clinical symptoms). If this degree of physiological change is considered to be undesirable, what margin of safety should be employed in labeling a gluten-free product? Surely 10-fold; perhaps substantially more.
- At the lower dose of 10 mg/day, morphological changes were observed in a few of the patients, but the response of the group did not reach statistical significance. While clear, statistical, endpoints are obviously most desirable, it seems important to also take into consideration individual reponses, since these may reflect real differences in gluten-sensitivity.
- Patients enrolled in this study were already consuming substantial quantities of commercially-available “gluten-free” products, so it is possible that the low-level gluten consumption may have masked the threshold of the precise amounts of extra gluten administered in this study.
In conclusion, given the very small sample size of patients in clinical trials performed so far, it is very hard to extrapolate from available safety data to the estimated 2 million individuals in the US with celiac disease; therefore it seems imperative to apply a wide margin of safety, until further information is available.
Ref. 2: A prospective, double-blind, placebo-controlled trial to establisha safe gluten threshold for patients with celiac disease. Am J Clin Nutr 2007; 85:160-166.
SOCIAL AND BUSINESS IMPACT
An ideal labeling regulation should balance the needs of patients and physicians for accurate information with the legitimate needs of food manufacturers and restaurants for labeling that is practically feasible and commercially viable.
The use of the suffix “-free” already has huge commercial marketing value, both in terms of product differentiation and in the ability to demand a price premium for gluten-free products. We think that a label including the term “-free” is desirable, providing that it consistent with the safety of the consumer, and believe that it is possible to reach a compromise labeling which balances the needs of the most sensitive consumer with the desires and needs of the food industry.
We thank you for your attention, and look forward to the implementation of a final labeling standard in the US.
Peter Olins, Ph.D., Gillian Olins, Ph. D.
Ultimate Gluten Free